Application Development Using C# and

Application Development Using C# and .
NET
By Michael Stiefel, Robert J. Oberg...............................................
Publisher: Prentice Hall PTR
Pub Date: December 21, 2001
Print ISBN-10: 0-13-093383-X
Print ISBN-13: 978-0-13-093383-6
Pages: 656
Slots: 1.0
Contents
The Integrated .NET Series From Object Innovations
Preface
Organization
Sample Programs
Caveat
Web Sites
Acknowledgments
The Integrated .NET Series from Object Innovations and Prentice Hall PTR
Introduction
Introductory .NET Language Books
Introduction to C# Using .NET
Introduction to Programming Visual Basic Using .NET
Programming Perl in the .NET Environment
Intermediate .NET Framework Books
Application Development Using C# and .NET
Application Development Using Visual Basic .NET
.NET Architecture and Programming Using Visual C++
Fundamentals of Web Applications Using .NET and XML
Chapter 1. What Is Microsoft .NET?
Microsoft and the Web
Windows on the Desktop
A New Programming Platform
The Role of XML
Summary
Chapter 2. .NET Fundamentals
Problems of Windows Development
Applications of the Future
.NET Overview
Summary
Chapter 3. C# Overview for Sophisticated Programmers
Hello World in C#
Performing Calculations in C#
Classes
C# Type System
Strings
Arrays and Indexers
More about Methods
Exceptions
Unsafe Code
Summary
Chapter 4. Object-Oriented Programming in C #
Review of Object-Oriented Concepts
Acme Travel Agency Case Study: Design
Inheritance in C#
Access Control
Acme Travel Agency Case Study: Implementation
More about Inheritance
Summary
Chapter 5. C# in the .NET Framework
System.Object
Collections
Interfaces
Acme Travel Agency Case Study: Step 2
Generic Interfaces in .NET
Delegates
Attributes
Summary
Chapter 6. User Interface Programming
Windows Forms Hierarchy
Simple Forms Using .NET SDK
Menus
Controls
Visual Studio.NET and Forms
Dialog Boxes
ListBox Control
Acme Travel Agency Case Study—Step 3
Summary
Chapter 7. Assemblies and Deployment
Assemblies
Private Assembly Deployment
Shared Assembly Deployment
Assembly Configuration
Multimodule Assemblies
Setup and Deployment Projects
Summary
Chapter 8. .NET Framework Classes
Metadata and Reflection
Input and Output in .NET
Serialization
.NET Application Model
Context
Application Isolation
Asynchronous Programming
Remoting
Custom Attributes
Garbage Collection and Finalization
Summary
Chapter 9. Programming with ADO.NET
.NET Data Providers
The Visual Studio.NET Server Explorer
Data Readers
Parameters Collection
SqlDataAdapter and the DataSet Class
DataSet Collections
DataSet Fundamentals
Database Transactions and Updates
Optimistic vs. Pessimistic Locking and the DataSet
Working with DataSets
Acme Travel Agency Case Study
XML Data Access
AirlineBrokers Database
Schema with Relationships
Typed DataSet
Summary
Chapter 10. ASP.NET and Web Forms
What Is ASP.NET?
Web Forms Architecture
Request/Response Programming
Web Applications Using Visual Studio.NET
ASP.NET Applications
State in ASP.NET Applications
ASP.NET Configuration
Server Controls
HTML Server Controls
Database Access in ASP.NET
Summary
Chapter 11. Web Services
Protocols
Web Service Architecture
SOAP Differences
Web Service Class
Hotel Broker Web Service
Summary
Chapter 12. Security
User-Based Security
Code Access Security
Internet Security
Role-Based Security in .NET
Forms-Based Authentication
Code Access Permissions
Code Identity
Security Policy
Summary
Chapter 13. Tracing and Debugging in .NET
The TraceDemo Example
Enabling Debug and Trace Output
Using the Debug and Trace Classes
Using Switches to Enable Diagnostics
Enabling or Disabling Switches
TraceListener
Listeners Collection
Summary
Chapter 14. Interoperability
Calling COM Components from Managed Code
Calling Managed Components from COM Client
Platform Invocation Services (PInvoke)
Summary
Appendix A. Visual Studio.NET
Overview of Visual Studio.NET
Creating a Console Application
Project Configurations
Debugging
Summary
The Integrated .NET Series From Object Innovations Top
C#
Oberg
*
Stiefel/Oberg
Visual Basic
*
Wyatt/Oberg
*
Application Development Using Visual Basic and .NET
Oberg/Thorsteinson/Wyatt
Visual C++
Thorsteinson/Oberg
Web Applications
Bell/Feng/Soong/Zhang/Zhu
PERL
Programming PERL in the .NET Environment
Saltzman/Oberg
Preface
Microsoft .NET is an advance in programming technology that greatly simplifies

application development both for traditional, proprietary applications, and for the
emerging paradigm of Web-based services. .NET is a complete restructuring of
Microsoft's whole system infrastructure and represents a major learning challenge for
programmers developing applications on Microsoft platforms. The new platform includes
a new programming language C# and a major class library, the .NET Framework.
This book covers important topics in the .NET Framework for experienced programmers.
You do not need prior experience in C#, because there is a self-contained treatment, but
you should have experience in some object-oriented language such as C++ or Java. The
book could also be read by a seasoned Visual Basic programmer who has experience
working with objects and components in VB.
If you already understand C#, you may safely skip or skim Chapters 3 and 4. Chapter 5
contains important information about the interactions of C# and the .NET Framework.
You may then proceed with a detailed study of the .NET Framework in Chapters 6 and
beyond. For a thorough introduction to the C# language you may read the book
Introduction to C# Using .NET.
The book is practical, with many examples and a major case study. The goal is to equip
you to begin building significant applications using the .NET Framework. The book is
part of The Integrated .NET Series from Object Innovations and Prentice Hall PTR.
Organization
The book is organized into five major parts, and is structured to make it easy for you to
navigate to what you most need to know. The first part, consisting of Chapters 1 and 2,
should be read by everyone. It answers the question "What is Microsoft .NET?" and
outlines the programming model of the .NET Framework.
The second part, consisting of Chapters 3-5, covers the C# programming language. If you
are already familiar with C# you can skim these chapters, paying the most attention to
Chapter 5, which covers topics such as interfaces, delegates, and events. This chapter also
describes important interactions between C# and the .NET Framework. The case study,
which is elaborated throughout the entire book, is introduced in Chapter 4.
The third part, Chapters 6-9, covers important fundamental topics in the .NET
Framework. Chapter 6 covers user interface programming using the Windows Forms
classes. Chapter 7 discusses assemblies and deployment, which constitute a major
advance in the simplicity and robustness of deploying Windows applications, ending the
notorious "DLL hell." Chapter 8 delves into important .NET Framework classes,
including the topics of metadata, serialization, threading, attributes, application domains,
asynchronous programming, remoting, and memory management. Chapter 9 covers
ADO.NET, which provides a consistent set of classes for accessing both relational and
XML Data.
The fourth part of the book provides an in-depth introduction to Web programming using
ASP.NET and SOAP. Chapter 10 introduces the fundamentals of ASP.NET, including the
use of Web Forms, which greatly simplifies the development of sophisticated Web sites.
Chapter 11 covers SOAP and Web Services, which provide an easy-to-use and robust
mechanism for heterogeneous systems to interoperate.
The final part of the book covers additional important topics in the .NET Framework.
Chapter 12 covers the topic of security in detail, including code access security,
declarative security, and the securing of Web applications and services. Chapter 13
introduces the debug and trace classes provided by .NET. Chapter 14 covers
interoperability of .NET with COM and with Win32 applications.
Sample Programs
The only way to really learn a major framework is to read and write many, many
programs, including some of reasonable size. This book provides many small programs
that illustrate pertinent features of .NET in isolation, which makes them easy to
understand. The programs are clearly labeled in the text, and they can all be found in the
software distribution that accompanies this book.
A major case study, the Acme Travel Agency, is progressively developed in Chapters 4
through 12. It illustrates many features of C# and .NET working in combination, as they
would in a practical application.
The sample programs are provided in a self-extracting file on the book's Web site. When
expanded, a directory structure is created, whose default root is c:\OI\NetCs. The sample
programs, which begin with the second chapter, are in directories Chap02, Chap03, and
so on. All the samples for a given chapter are in individual folders within the chapter
directories. The names of the folders are clearly identified in the text. Each chapter that
contains a step of the case study has a folder called CaseStudy, containing that step. If
necessary, there is a readme.txt file in each chapter directory to explain any instructions
necessary for getting the examples to work.
This book is part of The Integrated .NET Series. The sample programs for other books in
the series are located in their own directories underneath \OI, so all the .NET examples
from all books in the series will be located in a common area as you install them.
These programs are furnished solely for instructional purposes and should not be
embedded in any software product. The software (including instructions for use) is
provided "as is" without warranty of any kind.
Caveat
The book and the associated code were developed with Beta 2 of the .NET Framework.
Microsoft has indicated that this version of .NET is close to what will be the final
version. Nonetheless, changes will be made before .NET is released. The code in the
examples has been verified to work only with Windows 2000. Database code has been
verified with SQL Server 2000. Several examples in the database and security chapters
have machine names embedded in connection strings or role names. When trying to run
these examples, you will have to replace those names with the appropriate name for your
machine. To make installation easy, the database examples run with user name "sa" and
without a password. Needless to say, in a real system you should NEVER have any login
id without a password or have a database application use sa to log into a database. [1]
[1] That is just one of several steps necessary to avoid an SQL Injection attack.
Web Sites
The Web site for the book series is:
www.objectinnovations.com/dotnet.htm
A link is provided at that Web site for downloading the sample programs for this book.
Additional information about .NET technology is available at:
www.reliablesoftware.com
The book sample programs are available at this Web site as well.
The Web site for the book will also have a list of .NET learning resources that will be
kept up-to-date.
Acknowledgments
We are indebted to Mike Meehan for helping to get this project off the ground, starting at
a meeting at the PDC when Microsoft announced .NET. That conversation put into
motion what has become a substantial series of books on .NET technology, in which this
volume is the second. We would also like to thank Jill Harry at Prentice Hall for her
ongoing support with this ambitious book project.
Several people at Microsoft reviewed parts of the book: Steven Pratschner, Jim Hogg,
Michael Pizzo, Michael Day, Krzysztof Cwalina, Keith Ballinger, and Eric Olsen. We
thank them for taking time out from their very tight schedules to correct our manuscript.
Connie Sullivan and Stacey Giard coordinated technical sessions and helped assure our
access to resources at Microsoft.
Moshe Raab took precious time off from his consulting work and provided many helpful
suggestions. Peter Thorsteinson, an author of another book in our series, was a valuable
resource for understanding the deployment of .NET applications. Will Provost helped
clarify several issues related to XML. We also want to thank all the other authors in the
.NET series, because there is much synergy in a group working on parallel books, even if
in the heat of writing we did not always collaborate as closely as we might have. These
hardworking people include Eric Bell, Howard Feng, Michael Saltzman, Ed Soong, Dana
Wyatt, David Zhang, and Sam Zhu.
As always, reviewers should get credit for improving the quality of the work; any
remaining errors are the responsibility of the authors.
Robert always has a hard time writing acknowledgments, because there are so many
people to thank on such a major project. I (Robert) usually thank Michael Stiefel, but this
time he is my co-author, and so we are on the same side of the fence, thanking others. My
wife, Marianne, has provided enormous support and encouragement for all my writing
efforts. This project was especially demanding, and so her support is all the more
appreciated. Thank you all, and the other colleagues, friends, and students—too
numerous to mention individually—who have helped me over the years.
Michael would like to thank his wife not only for her understanding of his intellectual
lack of presence while writing the book (even if he was physically present), but also for
the associated behaviors, not the least of which was the repeated playing of music that
one social critic referred to as "Das Lied von der Erde and other light classics." Of course
I did not follow his other advice about how to write a book.
The Integrated .NET Series from Object Innovations

and Prentice Hall PTR Top
About this Series
Robert J. Oberg, Series Editor
Introduction
The Integrated .NET Book Series from Object Innovations and Prentice Hall PTR is a
unique series of introductory and intermediate books on Microsoft's important .NET
technology. These books are based on proven industrial-strength course development
experience. The authors are expert practitioners, teachers, and writers who combine
subject-matter expertise with years of experience in presenting complex programming
technologies such as C++, MFC, OLE, and COM/COM+. These books teach in a
systematic, step-by-step manner and are not merely summaries of the documentation. All
the books come with a rich set of programming examples, and a thematic case study is
woven through several of the books.
From the beginning, these books have been conceived as an integrated whole, and not as
independent efforts by a diverse group of authors.. The initial set of books consists of
three introductory books on .NET languages and four intermediate books on the .NET
Framework. Each book in the series is targeted at a specific part of the important .NET
technology, as illustrated by the diagram below.
Introductory .NET Language Books
The first set of books teaches several of the important .NET languages. These books
cover their language from the ground up and have no prerequisite other than
programming experience in some language. Unlike many .NET language books, which
are a mixture of the language and topics in the .NET Framework, these books are focused
on the languages, with attention to important interactions between the language and the
framework. By concentrating on the languages, these books have much more detail and
many more practical examples than similar books.
The languages selected are the new language C#, the greatly changed VB.NET, and
Perl.NET, the open source language ported to the .NET environment. Visual C++ .NET is
covered in a targeted, intermediate book, and JScript.NET is covered in the intermediate
level .NET Web-programming book.
This book provides thorough coverage of the C# language from the ground up. It is
organized with a specific section covering the parts of C# common to other C-like
languages. This section can be cleanly skipped by programmers with C experience or the
equivalent, making for a good reading path for a diverse group of readers. The book gives
thorough attention to the object-oriented aspects of C# and thus serves as an excellent
book for programmers migrating to C# from Visual Basic or COBOL. Its gradual pace
and many examples make the book an excellent candidate as a college textbook for
adventurous professors looking to teach C# early in the language's life-cycle.
Learn the VB.NET language from the ground up. Like the companion book on C#, this
book gives thorough attention to the object-oriented aspects of VB.NET. Thus the book is
excellent for VB programmers migrating to the more sophisticated VB.NET, as well as
for programmers experienced in languages such as COBOL. This book would also be
suitable as a college textbook.
Programming Perl in the .NET Environment
A very important part of the vision behind Microsoft® .NET is that the platform is
designed from the ground up to support multiple programming languages from many
sources, and not just Microsoft languages. This book, like other books in the series, is
rooted in long experience in industrial teaching. It covers the Perl language from the
ground up. Although oriented toward the ActiveState Perl.NET compiler, the book also
provides excellent coverage of the Perl language suitable for other versions as well.
Intermediate .NET Framework Books
The second set of books is focused on topics in the .NET Framework, rather than on
programming languages. Three parallel books cover the .NET Framework using the
important languages C#, VB.NET, and Visual C++. The C# and VB.NET books include
self-contained introductions to the languages suitable for experienced programmers,
allowing them to rapidly come up to speed on these languages without having to plow
through the introductory books. The fourth book covers the important topic of web
programming in .NET, with substantial coverage of XML, which is so important in the
.NET Framework.
The design of the series makes these intermediate books much more suitable to a wider
audience than many similar books. The introductory books focus on languages frees up
the intermediate books to cover the important topics of the .NET Framework in greater
depth. The series design also makes for flexible reading paths. Less experienced readers
can read the introductory language books followed by the intermediate framework books,
while more experienced readers can go directly to the intermediate framework books.
This book does not require prior experience in C#. However, the reader should have
experience in some object-oriented language such as C++ or Java™. The book could also
be read by seasoned Visual Basic programmers who have experience working with
objects and components in VB. Seasoned programmers and also a less experienced reader
coming from the introductory C# book can skip the first few chapters on C# and proceed
directly to a study of the Framework. The book is practical, with many examples and a
major case study. The goal is to equip the reader with the knowledge necessary to begin
building significant applications using the .NET Framework.
Application Development Using Visual Basic .NET
This book is for the experienced VB programmer who wishes to learn the new VB.NET
version of VB quickly and then move on to learning the .NET Framework. It is also
suitable for experienced enterprise programmers in other languages who wish to learn the
powerful RAD-oriented Visual Basic language in its .NET incarnation and go on to
building applications. Like the companion C# book, this book is very practical, with
many examples, and includes the same case study implemented in VB.NET.
This parallel book is for the experienced Visual C++ programmer who wishes to learn the
.NET Framework to build high-performing applications. Unlike the C# and VB.NET
book, there is no coverage of the C++ language itself, because C++ is too complex to
cover in a brief space. This book is specifically for experienced C++ programmers. Like
the companion C# and VB.NET books, this book is very practical, with many examples,
and includes the same case study implemented in Visual C++.
The final book in the series provides thorough coverage of building Web applications
using .NET. Unlike other books about ASP.NET, this book gives attention to the whole
process of Web application development. The book incorporates a review tutorial on
classical Web programming, making the book accessible to the experienced programmer
new to the Web world. The book contains significant coverage on ASP.NET, Web Forms,
Web Services, SOAP, and XML.
Chapter 1. What Is Microsoft .NET? Top
.NET is Microsoft's vision of applications in the Internet age. .NET provides enhanced
interoperability features based upon open Internet standards. .NET improves the
robustness of classic Windows applications. .NET offers developers a new programming
platform and superb tools, with XML playing a fundamental role.
Microsoft .NET is a platform built on top of the operating system. Three years in the
making before the public announcement, .NET represents a major investment by
Microsoft. .NET has been influenced by other technological advances such as XML,
Java™, and COM.
Microsoft .NET provides:
• The Common Language Runtime, a robust runtime platform.

• Multiple language development, with no language being more preferred over
any other.
• The .NET Framework, an extensible programming model, which provides a very
large class library of reusable code available to any .NET language.
• Support for a networking infrastructure built on top of Internet standards that
allows a high level of communication among applications.
• Support for the new industry standard of Web Services. Web Services represent a
new mechanism of application delivery that extends the idea of component-based
development to the Internet.
• ASP.NET, which allows you to use standard programming practices to develop
Web applications.
• A Deployment model that allows for versioning and the end of "DLL Hell."
• A Security model that is easy for programmers to use in their programs.
• An interoperability mechanism that enables .NET programs to access legacy code,
including COM components.
• Powerful development tools.
Microsoft and the Web
The World Wide Web has been a big catch-up challenge to Microsoft. Actually the Web
coexists quite well with Microsoft's traditional strength, the PC. Through a PC
application, the browser, a user gains access to a whole world of information.
The Web relies on standards such as HTML, HTTP, and XML, which are essential for
communication among diverse users on a wide variety of computer systems and devices.
While complex, the Windows PC is quite standardized. While the Web is based on
standard protocols, there is a Tower of Babel of multiple languages, databases,
development environments, and devices running on top of those protocols. This
exploding complexity of technology exacerbates a growing shortage of knowledge
workers who can build the needed systems using the new technologies. .NET provides
the infrastructure so that programmers can concentrate on adding value in their
applications without having to reinvent solutions to common programming problems.
Applications in the Internet Age
Originally the Web was a vast information repository. Browsers would make requests for
pages of existing information, and Web servers would deliver this information as static
HTML pages. Even when interactive Web applications were introduced, HTML, which
combines information with the details of how it is formatted for viewing, was still used.
XML provides a standard way of transmitting data independent of its formatting. XML
can thus provide ways for companies to agree on standards for documents and
information flows, such as purchase orders and invoices. E-commerce can then be
automated among cooperating companies (B-to-B). XML, however, only describes the
data; it does not supply the actions to be performed on that data. For that we need Web
Services.
Web Services
One of the most important aspects of .NET is the support for Web Services. Based on the
industry standard SOAP protocol, Web Services allow you to expose your applications'
functionality across the Internet. From the perspective of a .NET programmer, a Web
Service is no different from any other kind of service implemented by a class in a .NET
language. The programming model is the same for calling a function within an
application, in a separate component on the same machine, or as a Web Service on a
different machine.
This inherent simplicity will make it very easy for companies to create and host
applications. If desired, a whole application could be completely outsourced, removing
issues of development, deployment, and maintenance. Or you could use third-party Web
Services that did not exist when you designed your application.
ASP.NET
.NET includes a totally redone version of the popular Active Server Pages technology,
known as ASP.NET. Whereas ASP relied on interpreted script code in languages with
limited capabilities interspersed with page-formatting commands, ASP.NET code can be
written in any NET language, including C#, VB.NET, JScript, and C++ with managed
extensions. Since this is compiled code, you can separate your interface code from your
business logic in a separate "code behind" file. Although C#, VB.NET, and JScript may
be left as embedded script within the Web page, managed C++ must be placed in a code
behind file.
ASP.NET provides Web forms, which vastly simplifies creating Web user interfaces.
Drag-and-drop in Visual Studio.NET makes it very easy to lay out forms. You can add
code to form events such as a button click.
ASP.NET will automatically detect browser capability. For high-end browsers code
processing can be performed on the client. For low-end browsers the server does the
processing and generates standard HTML. All this is done transparently to the developer
by ASP.NET.
The combination of Web Services and compiled full-blown languages such as C#,
VB.NET, and managed C++, allows Web programming to follow an object-oriented
programming model, which had not been possible with ASP scripting languages and
COM components.
Open Standards and Interoperability
The modern computing environment contains a vast variety of hardware and software
systems. Computers range from mainframes and high-end servers to workstations and
PCs and to small mobile devices such as PDAs and cell phones. Operating systems
include traditional mainframe systems, many flavors of UNIX, Linux, several versions of
Windows, real-time systems, and special systems such as PalmOs for mobile devices.
Many different languages, databases, application development tools, and middleware
products are used.
In the modern environment, few applications are an island unto themselves. Even shrink-
wrapped applications deployed on a single PC may use the Internet for registration and
updates. The key to interoperability among applications is the use of standards. Since
applications typically run over a network, a key standard is the communications protocol
used.
Communications Protocols
TCP/IP sockets is highly standard and widely available. Too much detail, however, has to
be mastered, for programmers to be productive in writing robust distributed applications.
Somewhat higher is the remote procedure call (RPC), but RPC is still very complex, and
there are many flavors of RPC. Popular are higher level protocols, such as CORBA, RMI,
and DCOM. These are still complex, and require special environments at both ends.
These protocols suffer other disadvantages, such as difficulty in going across firewalls.
One communication protocol has become ubiquitous: HTTP. For this reason, Microsoft,
IBM, and other vendors have introduced a new protocol called SOAP (Simple Object
Access Protocol). SOAP uses text-based XML to encode object method requests and the
accompanying data. The great virtue of SOAP is its simplicity, leading to ease of
implementation on multiple devices. While SOAP can run on top of any protocol, its
ability to run on top of standard Internet protocols, such as HTTP, allows it to pass
through firewalls without any connectivity problems.
Windows on the Desktop
Microsoft began with the desktop. The modern Windows environment has become
ubiquitous. Countless applications are available, and most computer users are at least
somewhat at home with Windows. While Microsoft has made much progress in
modernizing Windows, there are still significant problems.
Problems with Windows
Maintaining a Windows PC is a chore, because applications are quite complex. They

consist of many files, registry entries, shortcuts, and so on. Different applications can
share certain DLLs, and installing a new application can overwrite a DLL an existing
application depends on, possibly breaking an old application ("DLL Hell"). Removing an
application is a complex operation and is often imperfectly done.
A PC can gradually become less stable, sometimes requiring the drastic cure of
reformatting the hard disk and starting from scratch. While there is tremendous economic
benefit to using PCs, because standard applications are inexpensive and powerful and the
hardware is cheap, the savings are reduced by the cost of maintenance.
Windows was originally developed when personal computers were not connected over a
network and security was not an issue. While security was built into Windows NT and
Windows 2000, the programming model is difficult to use. (Pop quiz: Did you ever pass
anything but NULL to a Win32 LPSECURITY_ATTRIBUTES argument?)
The Glass House and Thin Clients
The old "glass house" model of a central computer that controls all applications has had
an appeal, and there has been a desire to move toward "thin clients" of some sort. But the
much heralded "network PC" never really caught on. There is too much of value in
standard PC applications. Users like the idea of their "own" PC, with their data stored
safely and conveniently on their local computer. Without broadband connectivity a
server-based application such as word processing would not perform very well. Security
is also a very difficult issue to solve with thin clients. The personal computer is
undoubtedly here to stay.
A Robust Windows
With all the hype about .NET and the Internet, it is important to realize that .NET has
changed the programming model to allow the creation of much more robust Windows
applications. Applications no longer rely on storing extensive configuration data in the
fragile Windows Registry. .NET applications are self-describing, containing metadata
within the program executable files themselves. Different versions of an application or
component can be deployed side-by-side. Applications can share components through the
Global Assembly Cache. Versioning is built into the deployment model. A
straightforward security model is part of .NET. Windows Forms technology is a new
paradigm for building Windows GUI applications.
A New Programming Platform
Let us look at what we have just discussed from the point of view of .NET as a new
programming platform:
• Code can be validated to prevent unauthorized actions.

• It is much easier to program than the Win32 API or COM.
• All or parts of the platform can be implemented on many different kinds of
computers (as has been done with Java).
• All the languages use one class library.
• Languages can interoperate with each other.
There are several important features to the .NET platform:
• .NET Framework
• Common Language Runtime
• Multiple language development
• Development tools
.NET Framework
Modern programming relies heavily on reusable code provided in libraries. Object-

oriented languages facilitate the creation of class libraries, which are flexible, have a
good degree of abstraction, and are extensible by adding new classes and basing new
classes on existing ones, "inheriting" existing functionality.
The .NET Framework provides over 2500 classes of reusable code, which can be called
by all the .NET languages. The .NET Framework is extensible, and new classes can
inherit from existing classes, even those implemented in a different language.
Examples of classes in the .NET Framework include Windows programming, Web

programming, database programming, XML, and interoperability with COM and Win32.
The .NET Framework is discussed in the next chapter and throughout the rest of the
book.
Common Language Runtime
A runtime provides services to executing programs. Traditionally there are different

runtimes for different programming environments. Examples of runtimes include the
standard C library, MFC, the Visual Basic runtime, and the Java Virtual Machine. The
runtime environment provided by .NET is called the Common Language Runtime or
CLR.
Managed Code and Data
The CLR provides a set of services to .NET code (including the .NET Framework, which
sits on top of the CLR). In order to make use of these services, .NET code has to behave
in a predictable fashion, and the CLR has to understand the .NET code. For example, to
do runtime checking of array boundaries, all .NET arrays have identical layout. NET
code can also be restricted by type safety requirements.
As we will discuss in the next chapter, the restrictions on .NET code are defined in the
Common Type System (CTS) and its implementation in the Microsoft Intermediate
Language (MSIL or IL). The Common Type System defines the types and operations that
are allowed in code running under the CLR. For example, it is the CTS that restricts types
to using single implementation inheritance. MSIL code is compiled into the native code
of the platform.
.NET applications contain metadata, or descriptions of the code and data in the
application. Metadata allows the CLR, for example, to automatically serialize data into a
storage.
Code that can use the services of the Common Language Runtime is called managed
code.
Managed data is allocated and deallocated automatically. This automatic deallocation is

called garbage collection. Garbage collection reduces memory leaks and similar
problems.
Microsoft and ECMA

Microsoft has submitted specifications for the C# programming language and
core parts of the .NET Framework to the European Computer Manufacturers
Association (ECMA) for standardization. The ECMA specification defines the
platform-independent Common Language Infrastructure (CLI). The CLR can be
thought of as the CLI plus the Base Class Libraries (BCL). The BCL has support
for the fundamental types of the CTS such as file I/O, strings, and formatting.
Since the CLR is platform dependent, it makes use of the process and memory
management models of the underlying operating system.
The ECMA specification defines the Common Intermediate Language (CIL). The
ECMA specification allows for CIL to be compiled into native code or
interpreted.
Verifiable Code
Managed code can be checked for type safety. Type safe code cannot be subverted. For
example, a buffer overwrite cannot corrupt other data structures or programs. You can
only enter and leave methods at fixed points, you cannot calculate a memory address and
start executing code at an arbitrary point. Security policy can be applied to type safe
code. For example, access to certain files or user interface features can be allowed or
denied. You can prevent the execution of code from unknown sources.
Not all code that makes use of the facilities of the CLR is necessarily type safe. The
canonical example is managed C++. Managed C++ code can make use of CLR facilities
such as garbage collection, but cannot be guaranteed to be type safe.
Multiple Language Development
As its name suggests, the CLR supports many programming languages. A "managed
code" compiler must be implemented for each language. Microsoft itself has
implemented compilers for managed C++, Visual Basic.NET, Jscript, and the new
language C#. Well over a dozen other languages are being implemented by third parties,
among them COBOL by Fujitsu and Perl by ActiveState. To accommodate the use or
creation of .NET data types, however, new syntax often has to be introduced.
Nonetheless, programmers do not need to be retrained in a completely new language in
order to gain the benefits of .NET. Legacy code can be accessed through the
interoperability mechanism.
Development Tools
A practical key to success in software development is a set of effective tools. Microsoft

has long provided great tools, including Visual C++ and Visual Basic. With .NET they
have combined their development tools into a single integrated environment called Visual
Studio.NET.
• VS.NET provides a very high degree of functionality for creating applications in

all the languages supported by .NET.
• You can do multiple language programming, debugging, and so on.
• VS.NET has many kinds of designers for forms, databases, and other software
elements.
As with the languages themselves, third parties can provide extensions to Visual
Studio.NET, creating a seamless development environment for their language that
interoperates with the other .NET language. The tool set includes extensive support for
building Web applications and Web Services. There is also great support for database
application development.
The Importance of Tools
The importance of tools should not be underestimated. Ada, a very powerful

programming language, never achieved widespread use. While part of the initial vision
was to create a standard Ada Programming Support Environment (APSE), most of the
attention was paid to specifying the language, not the APSE. Consequently, Ada never did
develop any development environment comparable to that of Visual Studio, Smalltalk, or
some of the Java IDEs.
Visual Studio.NET will be highly tuned for productivity, and much training will be
available. Microsoft has far more resources to throw at Visual Studio.NET than do
smaller vendors in the highly fragmented tools market. Java is highly standardized in the
language and API, but tools, which are required for productivity, are not standard.
The Role of XML
XML is ubiquitous in .NET and is highly important in Microsoft's overall vision. Some
uses of XML in .NET include:
• XML can be used to model data in coordination with ADO.NET datasets.

• XML is used in configuration files.
• XML documentation can be automatically generated by some .NET languages.
• XML is used for encoding requests and responses in Web Services.
• XML is used to describe and transmit data in Web Services.
Success Factors for Web Services
The ultimate success of Microsoft's Internet vision depends on two external factors: the
infrastructure of the Internet and the success of the proposed Web Services business
model. The widespread use of Web Services depends on having high bandwidth widely
available. This capability will probably indeed materialize within the next several years.
The prospect for the business model remains to be seen.
It is important to understand that the overall .NET technology includes far more than the
widely hyped Internet part. The more robust Windows platform and the very powerful
.NET Framework and tools will be enduring features.
Summary
Microsoft .NET is a new platform built on top of the operating system. It provides many
capabilities for building and deploying both standard applications and new Web-based
ones. Web Services allow applications to expose functionality across the Internet,
typically using the SOAP protocol. SOAP supports a high degree of interoperability,
since it is based on widely adopted standards such as HTTP and XML.
NET uses managed code running on the Common Language Runtime that employs the
Common Type System. The .NET Framework is a very large class library available
consistently across many languages. XML plays a fundamental role in .NET. All this
functionality can be used to build more robust Windows applications as well as Internet
applications.
Chapter 2. .NET Fundamentals Top
What kind of problems is .NET designed to solve? .NET solves problems that have
plagued programmers in the past. .NET helps programmers develop the applications of
the future. This chapter is designed to present an overview of Microsoft .NET by looking
at a simple program rather than talking in vague generalities. While we will start
discussing Microsoft .NET in detail in Chapter 6, this chapter will enable you to get a feel
for the big picture right away.
Problems of Windows Development
Imagine a symphony orchestra where the violins and the percussion sections had different
versions of the score. It would require a heroic effort to play the simplest musical
composition. This is the life of the Windows developer. Do I use MFC? Visual Basic or
C++? ODBC or OLEDB? COM interface or C style API? Even within COM: do I use
IDispatch, dual, or pure vtable interfaces? Where does the Internet fit into all of this?
Either the design had to be contorted by the implementation technologies that the
developers understood, or the developers had to learn yet another technological approach
that was bound to change in about two years.
Deployment of applications can be a chore. Critical entries have to be made in a Registry

that is fragile and difficult to back up. There is no good versioning strategy for
components. New releases can break existing programs often with no information about
what went wrong. Given the problems with the Registry, other technologies used other
configuration stores such as a metabase or SQL Server.
Security in Win32 is another problem. It is difficult to understand and difficult to use.
Many developers ignored it. Developers who needed to apply security often did the best
they could with a difficult programming model. The rise of Internet-based security threats
transforms a bad situation into a potential nightmare.
Despite Microsoft's efforts to make development easier problems remained. Many system
services had to be written from scratch, essentially providing the plumbing code that had
nothing to do with your business logic. MTS/COM+ was a giant step in the direction of
providing higher level services, but it required yet another development paradigm. COM
made real component programming possible. Nonetheless, you either did it simply, but
inflexibly in Visual Basic, or powerfully, but with great difficulty in C++, because of all
the repetitive plumbing code you had to write in C++.
Applications of the Future
Even if .NET fixed all the problems of the past, it would not be enough. One of the
unchanging facts of programming life is that the boundaries of customer demand are
always being expanded.
The growth of the Internet has made it imperative that applications work seamlessly
across network connections. Components have to be able to expose their functionality to
other machines. Programmers do not want to write the underlying plumbing code, they
want to solve their customers' problems.
.NET Overview
The Magic of Metadata
To solve all these problems .NET must provide an underlying set of services that is
available to all languages at all times. It also has to understand enough about an
application to be able to provide these services.
Serialization provides a simple example. Every programmer at some time or another has
to write code to save data. Why should every programmer have to reinvent the wheel of
how to persist nested objects and complicated data structures? Why should every
programmer have to figure out how to do this for a variety of data stores? .NET can do
this for the programmer. Programmers can also decide to do it themselves if required.
To see how this is done, look at the Serialize sample associated with this chapter. For the
moment ignore the programming details of C# which will be covered in the next three
chapters, and focus on the concepts.
[Serializable] class Customer

{
public string name;
public long id;
}
class Test
{
static void Main(string[] args)
{
ArrayList list = new ArrayList();
Customer cust = new Customer();

cust.name = "Charles Darwin";
cust.id = 10;
list.Add(cust);
cust = new Customer();

cust.name = "Isaac Newton";
cust.id = 20;
list.Add(cust);
foreach (Customer x in list)

Console.WriteLine(x.name + ": " + x.id);
Console.WriteLine("Saving Customer List");

FileStream s = new FileStream("cust.txt",
FileMode.Create);
SoapFormatter f = new SoapFormatter();
f.Serialize(s, list);
s.Close();
Console.WriteLine("Restoring to New List");

s = new FileStream("cust.txt", FileMode.Open);
f = new SoapFormatter();
ArrayList list2 = (ArrayList)f.Deserialize(s);
s.Close();
foreach (Customer y in list2)

Console.WriteLine(y.name + ": " + y.id);
}
}
We have defined a Customer class with two fields: a name and an id. The program first
creates an instance of a collection class that will be used to hold instances of the
Customer class. We add two Customer objects to the collection and then print out the
contents of the collection. The collection is then saved to disk. It is restored to a new
collection instance and printed out. The results printed out will be identical to those
printed out before the collection was saved. [1]
[1]
The sample installation should have already built an instance that you can run. If not,
double-click on the Visual Studio.NET solution file that has the .sln suffix. When Visual
Studio comes up, hit Control-F5 to build and run the sample.
We wrote no code to indicate how the fields of the customer object are saved or restored.
We did have to specify the format (SOAP) and create the medium to which the data was
saved. The .NET Framework classes are partitioned so that where you load/save, the
format you use to load/save, and how you load/save can be chosen independently. This
kind of partitioning exists throughout the .NET Framework.
The Customer class was annotated with the Serializable attribute in the same way the
public attribute annotates the name field. If you do not want your objects to be
serializable, do not apply the attribute to your class. If an attempt is then made to save
your object, an exception will be thrown and the program will fail. [2]
[2]
Comment out the Serializable attribute in the program (you can use the C/C++/* */
comment syntax) and see what happens.
Attribute-based programming is used extensively throughout .NET to describe how the

Framework should treat code and data. With attributes you do not have to write any code;
the Framework takes the appropriate action based on the attribute. Security can be set
through attributes. You can use attributes to have the Framework handle multithreading
synchronization. Remoting of objects becomes straightforward through the use of
attributes.
The compiler adds this Serializable attribute to the metadata of the Customer class to
indicate that the Framework should save and restore the object. Metadata is additional
information about the code and data within a .NET application. Metadata, a feature of the
Common Language Runtime, provides such information about the code as:
• Version and locale information

• All the types
• Details about each type, including name, visibility, and so on
• Details about the members of each type, such as methods, the signatures of
methods, and the like
• Attributes
Since metadata is stored in a programming-language-independent fashion with the code,

not in a central store such as the Windows Registry, it makes .NET applications self-
describing. The metadata can be queried at runtime to get information about the code
(such as the presence or absence of the Serializable attribute). You can extend the
metadata by providing your own custom attributes.
In our example, the Framework can query the metadata to discover the structure of the
Customer object in order to be able to save and restore it.
Types
Types are at the heart of the programming model for the CLR. A type is
analogous to a class in most object-oriented programming languages, providing
an abstraction of data and behavior, grouped together. A type in the CLR
contains:
Fields (data members)
Methods
Properties
Events
There are also built-in primitive types, such as integer and floating point numeric
types, string, etc. We will discuss types under the guise of classes and value types
when we cover C#.
.NET Framework Class Library
The Formatter and FileStream classes are just two of more than 2500 classes in the .NET
Framework that provide plumbing and system services for .NET applications. Some of
the functionality provided by the .NET Framework includes:
• Base class library (basic functionality such as strings, arrays, and formatting)
• Networking
• Security
• Remoting
• Diagnostics
• I/O
• Database
• XML
• Web services that allow us to expose component interfaces over the Internet
• Web programming
• Windows User Interface
Interface-Based Programming
Suppose you want to encrypt your data and therefore do not want to rely on the
Framework's serialization. Your class can inherit from the ISerializable interface and
provide the appropriate implementation. (We will discuss how to do this in a later
chapter.) The Framework will then use your methods to save and restore the data.
How does the Framework know that you implemented the ISerializable interface? It can
query the metadata related to the class to see if it implements the interface! The
Framework can then use either its own algorithm or the class's code to serialize or
deserialize the object.
Interface-based programming is used in .NET to allow your objects to provide
implementations to standard functionality that can be used by the Framework. Interfaces
also allow you to program using methods on the interface rather than methods on the
objects. You can program without having to know the exact type of the object. For
example, the formatters (such as the SOAP formatter used here) implement the
IFormatter interface. Programs can be written using the IFormatter interface and thus are
independent of any particular current (binary, SOAP) or future formatter and still work
properly.
Everything Is an Object
So if a type has metadata, the runtime can do all kinds of wonderful things. But does
everything in .NET have metadata? Yes! Every type, whether it is user defined (such as
Customer) or part of the Framework (such as FileStream), is a .NET object. All .NET
objects have the same base class, the system's Object class. Hence everything that runs in
.NET has a type and therefore has metadata.
In our example, the serialization code can walk through the ArrayList of customer objects
and save each one as well as the array it belongs to, because the metadata allows it to
understand the object's type and its logical structure.
Common Type System
The .NET Framework has to make some assumptions about the nature of the types that
will be passed to it. These assumptions are the Common Type System (CTS). The CTS
defines the rules for the types and operations that the Common Language Runtime will
support. It is the CTS that limits .NET classes to single implementation inheritance. Since
the CTS is defined for a wide range of languages, not all languages need to support all
features of the CTS.
The CTS makes it possible to guarantee type safety, which is critical for writing reliable
and secure code. As we noted in the previous section, every object has a type and
therefore every reference to an object points to a defined memory layout. If arbitrary
pointer operations are not allowed, the only way to access an object is through its public
methods and fields. Hence it's possible to verify an object's safety by analyzing the
object. There is no need to know or analyze all the users of a class.
How are the rules of the CTS enforced? The Microsoft Intermediate Language (MSIL or
IL) defines an instruction set that is used by all .NET compilers. This intermediate
language is platform independent. The MSIL code can later be converted to a platform's
native code. Verification for type safety can be done once based on the MSIL; it need not
be done for every platform. Since everything is defined in terms of MSIL, we can be sure
that the .NET Framework classes will work with all .NET languages. Design no longer
dictates language choice; language choice no longer constrains design.
MSIL and the CTS make it possible for multiple languages to use the .NET Framework
since their compilers produce MSIL. This one of the most visible differences between
.NET and Java, which in fact share a great deal in philosophy.
ILDASM
The Microsoft Intermediate Language Disassembler (ILDASM) can display the metadata
and MSIL instructions associated with .NET code. It is a very useful tool both for
debugging and for increasing your understanding of the .NET infrastructure. You can use
ILDASM to examine the .NET Framework code itself. [3] Figure 2-1 shows a fragment of
the MSIL code from the Serialize example, where we create two new customer objects
and add them to the list. [4] The newobj instruction creates a new object reference using
the constructor parameter. [5] Stloc stores the value in a local variable. Ldloc loads a local
variable. [6] It is strongly recommended that you play with ILDASM and learn its features.
[3]
ILDASM is installed on the Tools menu in Visual Studio.NET. It is also found in the
Microsoft.NET\FrameworkSDK\Bin subdirectory. You can invoke it by double-clicking
on its Explorer entry or from the command line. If you invoke it from the command line
(or from VS.NET) you can use the /ADV switch to get some advanced options.
[4]
Open Serialize.exe and Click on the plus (+) sign next to Test. Double-click on Main to
bring up the MSIL for the Main routine.
[5]
Technically it is not a parameter. IL is a stack-based language, and the constructor is a
metadata token previously pushed on the stack.
[6]
You can read all about MSIL in the ECMA documents, specifically the Partition III CIL
Instruction Set.
Figure 2-1. Code fragment from Serialize example.

Language Interoperability
Having all language compilers use a common intermediate language and common base
class make it possible for languages to interoperate. But since all languages need not
implement all parts of the CTS, it is certainly possible for one language to have a feature
that another does not.
The Common Language Specification (CLS) defines a subset of the CTS representing the
basic functionality that all .NET languages should implement if they are to interoperate
with each other. This specification enables a class written in Visual Basic.NET to inherit
from a class written in COBOL.NET or C#, or to make interlanguage debugging possible.
An example of a CLS rule is that method calls need not support a variable number of
arguments, even though such a construct can be expressed in MSIL.
CLS compliance applies only to publicly visible features. A class, for example, can have
a private member that is non-CLS compliant and still be a base class for a class in another
.NET language. For example, C# code should not define public and protected class names
that differ only by case-sensitivity, since languages such as VB.NET are not case-
sensitive. Private fields could have case-sensitive names.
Microsoft itself is providing several CLS-compliant languages: C#, Visual Basic.NET,

and C++ with Managed Extensions. Third parties are providing additional languages
(there are over a dozen so far). ActiveState is implementing Perl and Python. Fujitsu is
implementing COBOL.
Managed Code
In the serialization example a second instance of the Customer object was assigned to the
same variable (cust) as the first instance without freeing it. None of the allocated storage
in the example was ever deallocated. .NET uses automatic garbage collection to reclaim
memory. When memory allocated on the heap becomes orphaned, or passes out of scope,
it is placed on a list of memory locations to be freed. Periodically, the system runs a
garbage collection thread that returns the memory to the heap.
By having automatic memory management the system has eliminated memory leakage,
which is one of the most common programming errors. In most cases, memory allocation
is much faster with garbage collection than with classic heap allocation schemes. Note
that variables such as cust and list are object references, not the objects themselves. This
makes the garbage collection possible.
Garbage collection is one of several services provided by the Common Language

Runtime (CLR) to .NET programs. [7] Data that is under the control of the CLR garbage
collection process is called managed data. Managed code is code that can use the services
of the CLR. .NET compilers that produce MSIL can produce managed code.
[7]
Technically, metadata, the CTS, the CLS, and the Virtual Execution System (VES) are
also part of the CLR. We are using CLR here in the sense that it is commonly used. The
VES loads and runs .NET programs and supports late binding. For more details refer to
the Common Language Infrastructure (CLI) Partition I: Concepts and Architecture
document submitted to ECMA. This document is loaded with the .NET Framework SDK.
Managed code is not automatically type safe. C++ provides the classic example. You can
use the __gc attribute to make a class garbage collected. The C++ compiler will prevent
such classes from using pointer arithmetic. Nonetheless, C++ cannot be reliably verified.
[8]
[8]
The most immediate reason for this is that the C Runtime Library (CRT) that is the
start-up code for C++ programs was not converted to run under .NET because of time
constraints. Even if this were to be done, however, there are two other obstacles to
verifying C++ code. First, to ensure that the verification process can complete in a
reasonable amount of time, the CLR language specifications require certain IL language
patterns to be used and the managed C++ compiler would have to be changed to
accommodate this. Second, after disallowing the C++ constructs that inhibit verification
(like taking the address of a variable on the stack, or pointer arithmetic), you would wind
up with a close approximation to the C# language.
Code is typically verified for type safety before compilation. This step is optional and can
be skipped for trusted code. One of the most significant differences between verified and
unverified code is that verified code cannot use pointers. [9] Code that used pointers could
subvert the Common Type System and access any memory location.
[9]
It would not be correct to say that code written in MSIL is managed code. The CTS
permits MSIL to have unmanaged pointers in order to work with unmanaged data in
legacy code. The reverse is not true; unmanaged code cannot access managed data. The
CLS prohibits unmanaged pointers.
Type safe code cannot be subverted. A buffer overwrite is not able to corrupt other data
structures or programs. Methods can only start and end at well-defined entry and exit
points. Security policy can be applied to type safe code. [10] For example, access to certain
files or user interface features can be allowed or denied. You can prevent the execution of
code from unknown sources. You can prevent access to unmanaged code to prevent
subversion of .NET security. Type safety also allows paths of execution of .NET code to
be isolated from one another. [11]
[10]
This is discussed in more detail in Chapter 12.
[11]
See the discussion of Application Domains in Chapter 8.
Assemblies
Another function of the CLR is to load and run .NET programs.
.NET programs are deployed as assemblies. An assembly is one or more EXEs or DLLs
with associated metadata information. The metadata about the entire assembly is stored in
the assembly's manifest. The manifest contains, for example, a list of the assemblies upon
which this assembly is dependent.
In our Serialize example there is only file in the assembly, serialize.exe. That file contains
the metadata as well as the code. Since the manifest is stored in the assembly and not in a
separate file (like a type library or registry), the manifest cannot get out of sync with the
assembly. Figure 2-2 shows the metadata in the manifest for this example. [12] Note the
assembly extern statements that indicate the dependencies on the Framework assemblies
mscorlib and System.Runtime.Formatters.SOAP. These statements also indicate the
version of those assemblies that serialize.exe depends on.
[12]
Open serialize.exe in ILDASM and double-click on the MANIFEST item.
Figure 2-2. Manifest for the Serialize assembly.

Assemblies can be versioned, and the version is part of the name for the assembly. To
version an assembly it needs a unique name. Public/private encryption keys are used to
generate a unique (or strong) name.
Assemblies can be deployed either privately or publicly. For private deployment all the
assemblies that an application needs are copied to the same directory as the application. If
an assembly is to be publicly shared, an entry is made in the Global Assembly Cache
(GAC) so that other assemblies can locate it. For assemblies put in the GAC a strong
name is required. Since the version is part of the assembly name, multiple versions can be
deployed side by side on the same machine without interfering with each other. Whether
you use public or private deployment there is no more "DLL Hell." [13]
[13]
This is discussed in much more detail in Chapter 7.
Assembly deployment with language interoperability makes component development

almost effortless.
JIT Compilation
Before executing on the target machine, MSIL has to be translated into the machine's
native code. This can either be done before the application is called, or at runtime. At
runtime, the translation is done by a just-in-time (JIT) compiler. The Native Image
Generator (Ngen.exe) translates MSIL into native code so that it is already translated
when the program is started.
The advantage of pretranslation is that optimizations can be performed. Optimizations are
generally impractical with JIT because the time it takes to do the optimization can be
longer than it takes to compile the code. Start-up time is also faster with pretranslation
because no translation has to be done when the application starts.
The advantage of JIT is that it knows what the execution environment is when the
program is run and can make better assumptions, such as register assignments, when it
generates the code. Only the code that is actually executed is translated, code that never
gets executed is never translated.
In the first release of .NET, the Native Image Generator and the JIT compiler use the
same compiler. No optimizations are done for Ngen, its only current advantage is faster
start-up. For this reason we do not discuss Ngen in this book.
Performance
You may like the safety and ease-of-use features of managed code but you might be
concerned about performance. Early assembly language programmers had similar
concerns when high-level languages came out.
The CLR is designed with high performance in mind. With JIT compilation, the first time
a method is encountered, the CLR performs verifications and then compiles the method
into native code (which will contain safety features, such as array bounds checking). The
next time the method is encountered, the native code executes directly. Memory
management is designed for high performance. Allocation is almost instantaneous, just
taking the next available storage from the managed heap. Deallocation is done by the
garbage collector, which has an efficient multiple-generation algorithm.
You do pay a penalty when security checks have to be made that require a stack walk as
we will explain in the Security chapter.
Web pages use compiled code, not interpreted code. As a result ASP.NET is much faster
than ASP.
For 98% of the code that programmers write, any small loss in performance is far
outweighed by the gains in reliability and ease of development. High performance server
applications might have to use technologies such as ATL Server and C++.
Summary
.NET solves the problems that have plagued Windows development in the past. There is
one development paradigm for all languages. Design and programming language choices
are no longer in conflict. Deployment is more rational and includes a versioning strategy.
While we will talk more about it in later chapters, metadata, attribute-based security, code
verification, and type-safe assembly isolation make developing secure applications much
easier. The plumbing code for fundamental system services is provided, yet you can
extend or replace it if you must.
The Common Language Runtime provides a solid base for developing applications of the
future. The CLR is the foundation whose elements are the Common Type System,
metadata, the Common Language Specification, and the Virtual Execution System (VES)
that executes managed code. [14] As we shall see in future chapters, .NET makes it easier
to develop Internet applications for both service providers and customer-based solutions.
With the unified development platform .NET provides, it will be much easier than in the
past for Microsoft or others to provide extensions.
[14]
The Base Class Libraries classes (BCL) are also part of the CLR.
All this is made possible by putting old technologies together in the CLR creatively:
intermediate languages, type-safe verification, and of course, metadata. As you will see,
metadata is used in many features in .NET.
We shall expand on these topics in the course of the book. We next cover the C#
language. Depending on your knowledge of C#, you might be able to skim Chapters 3, 4,
and 5. Chapter 4 introduces the Acme Travel Agency case study, which is used
throughout the book. Chapter 5 covers important topics about the interaction of C# and
the .NET Framework.
Chapter 3. C# Overview for Sophisticated Programmers Top
In this chapter we quickly cover the essentials of the C# language, which should be quite
easy for you to learn if you have experience with C++ or Java. A "hello, world" program
introduces the basic structure of C# programs. We then cover variables, operators, control
structures, formatting, methods, and input/output. Classes are fundamental in C#, and we
examine them in some detail. Besides the standard features, C# adds some convenience
features, such as properties. We cover the essentials of data types in C#, which
correspond to types in the Common Type System. We discuss the fundamental distinction
between value and reference types and see how to convert between them using boxing
and unboxing operations.
C# has a string type, and the StringBuilder class can be used for dynamically changing
strings. We examine arrays in C# and some operations provided by the System.Array
class. We then cover some additional topics concerning methods, including parameter
passing, variable length parameter lists, method overloading, and operator overloading.
We discuss exception handling in C# in some detail, including the use of user-defined
exception classes and structured exception handling.
We conclude the chapter by looking at how you can have "unsafe" sections of C# code,
which can be used to work with pointers for efficiency or for interoperating with legacy
code.
Hello World in C#
Whenever learning a new programming language, a good first step is to write and run a
simple program that will display a single line of text. Such a program demonstrates the
basic structure of the language, including output.
Here is "Hello, World" in C#. (See the Hello directory for this chapter.)
// Hello.cs
using System;
class Hello
{
public static int Main(string[] args)
{
Console.WriteLine("Hello, World");
return 0;
}
}
Compiling and Running (Command Line)
You can learn how to use the Microsoft Visual Studio.NET IDE (integrated development
environment) in Appendix A. You can also use the command-line tools of the .NET
Framework SDK. Be sure to get the environment variables set up properly, as described
in the sidebar. To compile this program at the command line, enter the command
>csc Hello.cs
An executable file Hello.exe will be generated. To execute your program, type at the
command line:
>Hello
The program will now execute, and you should see displayed the greeting:
Hello, World
Setting Environment Variables

In order to run command-line tools such as the C# compiler using a simple name
such as csc rather than a complete path, we must set certain environment
variables. To do so we can use a batch file, corvars.bat, which can be found in the
bin directory of the Framework SDK.
I experienced different behavior in different beta versions of the .NET

Framework SDK. In one version the environment variables were set up
automatically as part of the install, and in another version I had to use the
corvars.bat file.
If you have Visual Studio.NET installed, you can ensure that the environment
variables are set up by starting your command prompt session from Start |
Programs | Microsoft Visual Studio.NET 7.0 | Microsoft Visual Studio Tools |
Microsoft Visual Studio.NET Command Prompt.
Program Structure
// Hello.cs
class Hello
{
...
}
Every C# program has at least one class. A class is the foundation of C#'s support of
object-oriented programming. A class encapsulates data (represented by variables) and
behavior (represented by methods). All of the code defining the class (its variables and
methods) will be contained between the curly braces. We will discuss classes in detail
later.
Note the comment at the beginning of the program. A line beginning with a double slash
is present only for documentation purposes and is ignored by the compiler. C# files have
the extension .cs.
// Hello.cs
...
An alternate form of comment is to use an opening /* and a closing */.
/* This is a comment
that may be continued over
several lines */
There is a distinguished class, which has a method whose name must be Main. The
method should be public and static. An int exit code can be returned to the operating
system. Note that in C# the file name need not be the same as the name of the class
containing the Main method.
// Hello.cs
using System;
class Hello
{
{
...
return 0;
}
}
Use void if you do not return an exit code.
public static void Main(string[] args)
Command-line arguments are passed as an array of strings. The runtime will call this
Main method—it is the entry point for the program. All the code of the Main method will
be between the curly braces.
// Hello.cs
using System;
class Hello
{
{
return 0;
}
}
Every method in C# has one or more statements. A statement is terminated by a

semicolon. A statement may be spread out over several lines.
The Console class provides support for standard output and standard input. The method
WriteLine displays a string, followed by a new line.
Namespaces
Much standard functionality in C# is provided through many classes in the .NET

Framework. Related classes are grouped into namespaces. Many useful classes, such as
Console, are in the System namespace. The fully qualified name of a class is specified by
the namespace followed by a dot followed by a class name.
System.Console
A using statement allows a class to be referred to by its class name alone.
// Hello.cs
using System;
class Hello
{
{
return 0;
}
}
Performing Calculations in C#
Our "Hello, World" program illustrated the basic structure of a C# program, but we will
need a slightly more elaborate example to show the use of other basic programming
constructs, such as variables, expressions, and control structures. Our next example is a
simple calculator for an IRA account. We calculate the accumulation of deposits to an
IRA of $2000.00 a year at 6% interest for 10 years, assuming that each deposit is made at
the end of the year. Our calculation is performed two ways:
• In a loop, year by year, accumulating a total as we go

• Using a formula
The example program is in the folder Ira\Step1.
// Ira.cs - Step 1
using System;
class Ira
{
{
int years = 10;
decimal rate = 0.06m;
decimal amount = 2000M;
decimal interest;
decimal total = 0m;
Console.WriteLine("{0,4} {1,12} {2,12} {3,12}",
"Year", "Amount", "Interest", "Total");
for (int i = 1; i <= years; i++)
{
interest = total * rate;
total += amount + interest;
Console.WriteLine(
"{0, -4} {1, 12:C} {2, 12:C} {3, 12:C}",
i, amount, interest, total);
}
Console.WriteLine("\nTotal using formula = {0}",
IraTotal(years, (double) rate, (double) amount));
return 0;
}
public static double IraTotal(int years, double rate,
double amount)
{
double total =
amount * (Math.Pow(1 + rate, years) - 1) / rate;
long total_in_cents = (long) Math.Round(total * 100);
total = total_in_cents /100.0;
return total;
}
}
If you compile and run it, you will see this output:
Year Amount Interest Total

1 $2,000.00 $0.00 $2,000.00
2 $2,000.00 $120.00 $4,120.00
3 $2,000.00 $247.20 $6,367.20
4 $2,000.00 $382.03 $8,749.23
5 $2,000.00 $524.95 $11,274.19
6 $2,000.00 $676.45 $13,950.64
7 $2,000.00 $837.04 $16,787.68
8 $2,000.00 $1,007.26 $19,794.94
9 $2,000.00 $1,187.70 $22,982.63
10 $2,000.00 $1,378.96 $26,361.59
Total using formula = 26361.59
Variables
In C# variables are of a specific data type. Some common types are int for integers and
double for floating-point numbers. C# has the decimal data type, which has a high degree
of precision, suitable for monetary calculations.
You must declare and initialize variables before you can use them.
int years = 10; // reserves space and assigns

// an initial value
decimal interest; // reserves space but does
// not initialize it to any value
If an initial value is not specified in the declaration, the variable must be initialized in
code before it can be used. We will discuss initialization later in the chapter.
Variables must be either local within a method or members of a class. There are no global
variables in C#.
Literals
A literal is used when you explicitly write a value for a variable in a program. An integer
literal is represented by either an ordinary decimal integer or a hexadecimal integer. A
floating-point or decimal literal is represented by a number with a decimal point or by
exponential notation. You may influence the type [1] that is used for storing a literal by a
suffix. The suffix f or F indicates single precision floating point. The suffix d or D
indicates double precision floating point. The suffix m or M indicates decimal (think
money).
[1]
We discuss C# types, such as float, double, and decimal, later in the chapter.
decimal rate = 0.06m;

decimal amount = 2000M;
There are two forms for string literals. Escape sequences are not processed for string
literals that are prefixed with @.
string file1 ="c:\\test1.txt";

string file2 = @"c:\test2.txt";
C# Operators and Expressions
You can combine variables and literals via operators to form expressions. The C#
operators are similar to those in C and C++, having similar precedence and associativity
rules. There are three kinds of operators,
• Unary operators take one operand and use prefix notation (e.g., --a) or postfix
notation (e.g., a++).
• Binary operators take two operands and use infix notation (e.g., a + b).
• The one ternary operator ?: takes three operands and uses infix notation (e.g., expr
? x : y).
Operators are applied in the precedence order shown in Table 3-1. For operators of the
same precedence, order is determined by associativity.
• The assignment operator is right-associative (operations are performed from right

to left).
• All other binary operators are left-associative (operations are performed from left
to right).
Precedence and associativity can be controlled by parentheses; what is done first is

shown as the primary operator (x) in the precedence table. C# has checked and
unchecked operators, which will be discussed later.
Table 3-1. Operator Precedence in C#

Category Operators
Primary (x) x.y f(x) a[x] x++ x- new typeof sizeof checked
unchecked
Unary + - ! ~ ++x --x (T)x
Multiplicative * / %
Additive + -
Shift << >>
Relational < > <= >= is as
Equality == !=
Logical AND &
Logical XOR ^
Logical OR |
Conditional &&
AND
Conditional OR ||
Conditional ?:
Assignment = *= /= %= += -= <<= >>= &= ^= |=
Output and Formatting
The Console class in the System namespace supports two simple methods for performing
output:
• WriteLine writes out a string followed by a new line.

• Write writes out just the string without the new line.
You can write out other data types by relying on the ToString method of System.Object,
which will provide a string representation of any data type. We will discuss the root class
System.Object in Chapter 5, where you will also see how to override ToString for your
own custom data type. You can use the string concatenation operator + to build up an
output string.
int x = 24;
int y = 5;
int z = x * y;
Console.Write("Product of " + x + " and " + y);
Console.WriteLine(" is " + z);
The output is all on one line:
Product of 24 and 5 is 120
Placeholders
A more convenient way to build up an output string is to use placeholders {0}, {1}, and
so on. An equivalent way to do the output shown above is
Console.WriteLine("Product of {0} and {1} is {2}", x,y,z);
The program OutputDemo illustrates the output operations just discussed.
We will generally use placeholders for our output from now on. Placeholders can be
combined with formatting characters to control output format.
Format Strings
C# has extensive formatting capabilities, which you can control through placeholders and
format strings.
• Simple placeholders: {n}, where n is 0, 1, 2, . . . , indicating which variable to

insert
• Control width: {n,w}, where w is width (positive for right-justified and negative
for left-justified) of the inserted variable
• Format string: {n:S}, where S is a format string indicating how to display the
variable
• Width and format string: {n,w:S}
A format string consists of a format character followed by an optional precision specifier.

Table 3-2 shows the available format characters.
Table 3-2. C# Format Characters
Format Character Meaning
C Currency (locale specific)
D Decimal integer
E Exponential (scientific)
F Fixed point
G General (E or F)
N Number with embedded commas
X Hexadecimal
Sample Formatting Code
The program FormatDemo illustrates formatting. Our sample program Ira\Step1 provides
another example. The header uses width specifiers, and the output inside the loop uses
width specifiers and the currency format character.
...
{
Console.WriteLine(
"{0, -4} {1, 12:C} {2, 12:C} {3, 12:C}",
}
...
Control Structures
The preceding code fragment illustrates a for loop. The C# control structures include the
familiar control structures of the C family of languages,
• if
• while
• do
• for
• switch
• break
• continue
• return
• goto
These all have standard semantics, except for switch, which is less error-prone in C#.
There are several other control statements in C#:
• There is a foreach loop, which we will discuss later in connection with arrays and
collections.
• The throw statement is used with exceptions. We will discuss exceptions later in
this chapter.
• The lock statement can be used to enforce synchronization in multithreading
situations. We will discuss multithreading in Chapter 8.
Switch Statement
In C#, after a particular case statement is executed, control does not automatically
continue to the next statement. You must explicitly specify the next statement, typically
by a break or goto label. (As in C and C++, you may call for identical handling of several
cases by having empty statements for all the case labels except the last one.) In C# you
may also switch on a string data type. The program SwitchDemo illustrates use of the
switch statement in C#.
...
switch(scores[i])
{
case 1:
Console.Write("Very ");
goto case 2; // cannot fall through
case 2:
Console.WriteLine("Low");
break;
case 3:
Console.WriteLine("Medium");
break;
case 4:
case 5:
Console.WriteLine("High");
break;
default:
Console.WriteLine("Special Case");
break;
}
...
Methods
Our Ira\Step1 example program has a method IraTotal for computing the total IRA
accumulation by use of a formula. In C# every function is a method of some class; there
are no freestanding functions. If the method does not refer to any instance variables of the
class, the method can be static. We will discuss instance data of a class later in this
chapter. Since the method is accessed only from within the class, it is designated as
private.
Note the use of the Pow and Round methods of the Math class, which is another class in
the System namespace. These methods are static methods. To call a static method from
outside the class in which it is defined, place the name of the class followed by a period
before the method name. In C# you cannot employ the alternative C++ style of using an
instance name to qualify a static method.
...
private static double IraTotal(int years, double rate,
double amount)
{
double total =
return total;
}
...
Console Input in C#
Our first Ira program is not too useful, because the data are hardcoded. To perform the
calculation for different data, you would have to edit the source file and recompile. What
we really want to do is allow the user of the program to enter the data at runtime.
An easy, uniform way to do input for various data types is to read the data as a string and
then convert to the desired data type. Use the ReadLine method of the System.Console
class to read in a string. Use the ToXxxx methods of the System.Convert class to convert
the data to the type you need.
Console.Write("amount: ");
string data = Console.ReadLine();
amount = Convert.ToDecimal(data);
Although console input in C# is fairly simple, we can make it even easier using object-
oriented programming. We can encapsulate the details of input in an easy-to-use wrapper
class, InputWrapper (which is not part of the .NET Framework class library).
Using the InputWrapper Class
In C# you instantiate a class by using the new keyword.
InputWrapper iw = new InputWrapper();

This code creates the object instance iw of the InputWrapper class.
The InputWrapper class wraps interactive input for several basic data types. The
supported data types are int, double, decimal, and string. Methods getInt, getDouble,
getDecimal, and getString are provided to read those types from the command line. A
prompt string is passed as an input parameter. The directory InputWrapper contains the
files InputWrapper.cs, which implements the class, and TestInputWrapper.cs, which tests
the class. (For convenience, we provide the file InputWrapper.cs in each project where we
use it.)
You can use the InputWrapper class without knowing its implementation. With such
encapsulation, complex functionality can be hidden by an easy-to-use interface. (A listing
of the InputWrapper class is in the next section.)
Here is the code for Ira\Step2. We read in the deposit amount, the interest rate, and the
number of years, and we compute the IRA accumulation year by year. The first input is
done directly, and then we use the InputWrapper class. The bolded code illustrates how to
use the InputWrapper class. Instantiate an InputWrapper object iw by using new. Prompt
for and obtain input data by calling the appropriate getXXX method.
// Ira.cs - Step 2
using System;
class Ira
{
{
decimal amount; // annual deposit amount
decimal rate; // interest rate
int years; // number of years
decimal total; // total accumulation
decimal interest; // interest in a year
Console.Write("amount: ");
string data = Console.ReadLine();
amount = Convert.ToDecimal(data);
rate = iw.getDecimal("rate: ");
years = iw.getInt("years: ");
total = 0m;
{
Console.WriteLine(
"{0, -4} {1, 12:C} {2, 12:C} {3, 12:C}",
}
Console.WriteLine("\nTotal using formula = {0}",
IraTotal(years, (double) rate, (double) amount));
return 0;
}
private static double IraTotal(int years, double rate,
double amount)
{
double total =
return total;
}
}
Compiling Multiple Files
The program in Ira\Step2 is our first example of the common situation of a program with
multiple files (in this case, just two: Ira.cs and InputWrapper.cs). It is easy to compile
multiple files at the command line.
> csc /out:Ira.exe *.cs
This will compile all the files in the current directory. You should use the /out option to
specify the name of the output file.
If multiple classes contain a Main method, you can use the /main command-line option to
specify which class contains the Main method that you want to use as the entry point into
the program.
>csc /main:Ira /out:Ira.exe *.cs
InputWrapper Class Implementation
The InputWrapper class is implemented in the file InputWrapper.cs. You should find the
code reasonably intuitive, given what you already know about classes.
// InputWrapper.cs
//
// Class to wrap simple stream input
// Datatype supported:
// int
// double
// decimal
// string
using System;
class InputWrapper
{
public int getInt(string prompt)
{
Console.Write(prompt);
string buf = Console.ReadLine();
return Convert.ToInt32(buf);
}
public double getDouble(string prompt)
{
return Convert.ToDouble(buf);
}
public decimal getDecimal(string prompt)
{
return Convert.ToDecimal(buf);
}
public string getString(string prompt)
{
return buf;
}
}
Note that, unlike the method IraTotal, the methods of the InputWrapper class are used
outside of the class so they are marked as public.
If bad input data is presented, an exception will be thrown. Exceptions are discussed later
in this chapter.
Classes
In this section we carefully examine the C# class, which is fundamental to programming

in C#. For illustration we introduce two classes, Customer and Hotel, which will be
elaborated in a case study that is used throughout the book. We will introduce the case
study itself in Chapter 4.
If you are a Java programmer, you will find the C# class to be quite familiar, and you
should be able to skim this section. C++ programmers must read much more carefully. C#
differs from C++ with respect to object instantiation, assignment, and destruction. Our
pace is somewhat more leisurely in this section, because classes are so fundamental to
programming in C#.
Classes as Structured Data
C# defines primitive data types that are built into the language. Data types, such as int
and decimal, can be used to represent simple data. C# provides the class mechanism to
represent more complex forms of data. Through a class, you can build up structured data
out of simpler elements, which are called data members, or fields. (See
TestCustomer\Step1.)
// Customer.cs - Step 1
public class Customer

{
public int CustomerId;
public string FirstName;
public string LastName;
public string EmailAddress;
public Customer(string first, string last, string email)
{
FirstName = first;
LastName = last;
EmailAddress = email;
}
}
Customer is now a new data type. A customer has a CustomerId, a FirstName, a

LastName, and an EmailAddress.
Classes and Objects
A class represents a "kind of," or type of, data. It is analogous to the built-in types like int
and decimal. A class can be thought of as a template from which individual instances can
be created. An instance of a class is called an object. Just as you can have several
individual integers that are instances of int, you can have several customers that are
instances of Customer. The fields, such as CustomerId and FirstName in our example, are
sometimes also called instance variables.
References
There is a fundamental distinction between the primitive data types and the extended data
types that can be created using classes. When you declare a variable of a primitive data,
you are allocating memory and creating the instance.
int x; // 4 bytes of memory have been allocated
When you declare a variable of a class type (an object reference), you are only obtaining
memory for a reference to an object of the class type. No memory is allocated for the
object itself, which may be quite large. This behavior is very different from that of C++,
where declaring an object in this way causes an instance to be created, using the default
constructor. The behavior is identical to what happens in Java.
Customer cust; // cust is a reference to a Customer object

// The object itself does not yet exist
Constructors
Through a constructor, you can initialize individual objects in any way you wish. Besides
initializing instance data, you can perform other appropriate initializations (e.g., open a
file).
A constructor is like a special method that is automatically called when an object is

created via the new keyword. A constructor
• has no return type

• has the same name as the class
• should usually have public access
• may take parameters, which are passed when invoking new
In the calling program, you use new to instantiate object instances, and you pass desired
values as parameters.
Default Constructor
If you do not define a constructor in your class, C# will implicitly create one for you. It is
called the default constructor and takes no arguments. The default constructor will assign
instance data, using any assignments in the class definition. Fields without an initializer
are assigned default values (0 for numerical data types, empty string for string, and so
on). The default constructor is called when an object instance is created with new and no
parameters. If you provide code for any constructor in your class, you must explicitly
define a default constructor with no arguments, if you want one.
Instantiating and Using an Object
You instantiate an object by the new operator, which will cause a constructor to be
invoked.
cust = new Customer("Rocket",

"Squirrel",
"[email protected]");
// Customer object now exists and cust is a reference to it
Once an object exists, you work with it, including accessing its fields and methods. Our
simple Customer class at this point has no methods, only four fields. You access fields
and methods using a dot.
cust.CustomerId = 1; // all fields have now been assigned

TestCustomer\Step0 provides a simple test program to exercise the Customer class. Note
that an unassigned field of a class receives a default value, such as 0, when an object is
instantiated.
Assigning Object References
TestCustomer\Step1 provides a more complete test program to exercise the Customer

class. Two object instances are created, an assignment is made of one object reference to
another, and a field is assigned a value.
// TestCustomer.cs
using System;
public class TestCustomer

{
public static void Main()
{
Customer cust1, cust2;
cust1 = new Customer("Rocket",
"Squirrel",
cust1.CustomerId = 1;
cust2 = new Customer("Bullwinkle",
"Moose´´,
cust2.CustomerId = 2;
ShowCustomer("cust1", cust1);
cust1 = cust2;
cust1.EmailAddress = "[email protected]";
}
private static void ShowCustomer(string label,
Customer cust)
{
Console.WriteLine("- - - - {0} - - - -", label);
Console.WriteLine("CustomerId = {0}",
cust.CustomerId);
Console.WriteLine("FirstName = {0}", cust.FirstName);
Console.WriteLine("LastName = {0}", cust.LastName);
Console.WriteLine("EmailAddress = {0}",
cust.EmailAddress);
}
}
Figure 3-1 shows the object references cust1 and cust2 and the data they refer to after the
objects have been instantiated and the CustomerId field has been assigned.
Figure 3-1. Two object references and the data they refer to.
When you assign an object variable, you are assigning only the reference; there is no
copying of data. [2] Figure 3-2 shows both object references and their data after the
assignment:
[2]
C and C++ programmers will recognize assignment of references as similar to
assignment of pointers.
Figure 3-2. Two references refer to the same data.
cust1 = cust2; // cust1 and cust2 now refer to same object
Now when you assign a new value a field of one object,
cust1.EmailAddress = "[email protected]";
you will see the same data through both object references. Here is the output from
running TestCustomer\Step1.
- - - - cust1 - - - -
CustomerId = 1
FirstName = Rocket
LastName = Squirrel
EmailAddress = [email protected]
- - - - cust2 - - - -
CustomerId = 2
FirstName = Bullwinkle
LastName = Moose
- - - - cust1 - - - -
CustomerId = 2
LastName = Moose
cust1 1
"Rocket"
"Squirrel"
"[email protected]"
cust2 2
"Bullwinkle"
"Moose"
"[email protected]"
- - - - cust2 - - - -
CustomerId = 2
LastName = Moose
Garbage Collection
Through the assignment of a reference, an object may become orphaned. Objects may
also be orphaned when they pass out of scope. Such an orphan object (or "garbage") takes
up memory in the computer, which can now never be referenced. In Figure 3-2 the
customer with CustomerId of 1 is now garbage.
The Common Language Runtime automatically reclaims the memory of unreferenced

objects. This process is known as garbage collection. Garbage collection takes up some
execution time, but it is a great convenience for programmers, helping to avoid a
common program error known as a memory leak. Garbage collection is discussed in more
detail in Chapter 8.
Methods
Typically, a class will specify behavior as well as data. A class encapsulates data and
behavior in a single entity. A method specifies the behavior and consists of
• An access specifier, typically public or private
• A return type (can be void if the method does not return data)
• A method name, which can be any legal C# identifier
• A parameter list, enclosed by parentheses, which specifies data that is passed to
the method (can be empty if no data is passed)
• A method body, enclosed by curly braces, which contains the C# code that the
method will execute
public void RaisePrice(decimal amount)

{
rate += amount;
}
In this example the return type is void (no data is passed back), the method name is
RaisePrice, the parameter list consists of a single parameter of type decimal, and the body
contains one line of code that increments the member variable rate by the value that is
passed in.
RaisePrice is a method in the Hotel class. The initial version of the Hotel class with a
simple test program is in the folder TestHotel\Step1.
Public and Private
Fields and methods of a C# class can be specified as public or private. Normally, you
declare fields as private. A private field can be accessed only from within the class, not
from outside.
public class Hotel

{
private string city;
private string name;
private int number = 50; // legal in C#
private decimal rate;
...
Note that in C# you can initialize fields when they are declared. This kind of initialization
is not legal in C++.
Methods may be declared as either public or private. Public methods are called from
outside the class and are used to perform calculations and to manipulate the private data.
You may also provide public "accessor" methods to provide access to private fields.
...
public decimal GetRate()
{
return rate;
}
public void SetRate(decimal val)
{
rate = val;
}
...
You may also have private methods, which can be thought of as "helper functions" for
use within the class. Rather than duplicating code in several places, you may create a
private method, which will be called wherever it is needed. An example is the ShowHotel
method in TestHotel.cs.
This
Sometimes it is convenient within code for a method to be able to access the current
object reference. C#, like C++, defines a keyword this, which is a special variable that
always refers to the current object instance. With this you can then refer to the instance
variables. The Hotel class has a constructor to initialize its instance data with values
passed as parameters. We can make use of the same names for parameters and fields and
remove ambiguity by using the this variable. Here is the code for the constructor:
public Hotel(string city, string name, int number,

decimal rate)
{
this.city = city;
this.name = name;
this.number = number;
this.rate = rate;
}
Sample Program
The program TestHotel\Step1 illustrates all the features we have discussed so far. Here is
the class definition:
// Hotel.cs - Step 1
public class Hotel

{
private int number = 50; // legal in C#
public Hotel(string city, string name, int number,
decimal rate)
{
this.city = city;
this.name = name;
this.number = number;
this.rate = rate;
}
public Hotel()
{
}
public string GetCity()
{
return city;
}
public string GetName()
{
return name;
}
public int GetNumber()
{
return number;
}
public void SetNumber(int val)
{
number = val;
}
public decimal GetRate()
{
return rate;
}
public void SetRate(decimal val)
{
rate = val;
}
public void RaisePrice(decimal amount)
{
rate += amount;
}
}
Here is the test program:
// Test.cs - Step 1
using System;
public class TestHotel

{
{
Hotel generic = new Hotel();
ShowHotel(generic);
Hotel ritz = new Hotel("Atlanta", "Ritz", 100, 95m);
ShowHotel(ritz);
ritz.RaisePrice(50m);
ritz.SetNumber(125);
ShowHotel(ritz);
}
private static void ShowHotel(Hotel hotel)
{
Console.WriteLine(
"{0} {1}: number = {2}, rate = {3:C}",
hotel.GetCity(), hotel.GetName(),
hotel.GetNumber(), hotel.GetRate());
}
}
Here is the output:
: number = 50, rate = $0.00

Atlanta Ritz: number = 100, rate = $95.00
Atlanta Ritz: number = 125, rate = $145.00
Properties
The encapsulation principle leads us to typically store data in private fields and to provide
access to this data through public accessor methods that allow us to set and get values.
For example, in the Hotel class we provided a method GetCity to access the private field
city. You don't need any special syntax; you can simply provide methods and call these
methods what you want, typically GetXXX and SetXXX.
C# provides a special property syntax that simplifies user code. You can access a private
field as if it were a public member. Here is an example of using a Number property of our
Hotel class.
ritz.Number = 125;
Console.WriteLine("There are now {0} rooms", ritz.Number);
As you can see, the syntax using the property is a little more concise. Properties were
popularized in Visual Basic and are now part of .NET and available in selected other
.NET languages, such as C#. The program TestHotel\Step2, illustrates implementing and
using several properties, City, Name, Number, and Rate. The first two properties are
read-only (only get defined), and the other properties are read/write (both get and set). It
is also possible to have a write-only property (only set defined). Here is the code for the
properties Name (read-only) and Number (read-write) in the second version of the Hotel
class. Notice the syntax and the C# keyword value to indicate the new value of the field.
public class Hotel

{
private int number;
...
public string Name
{
get
{
return name;
}
}
public int Number
{
get
{
return number;
}
set
{
number = value;
}
}
...
Static Fields and Methods
In C# a field normally is assigned on a per-instance basis, with a unique value for each
object instance of the class. Sometimes it is useful to have a single value associated with
the entire class. This type of field is called a static field. Like instance data members,
static data members can be either public or private. To access a public static member, you
use the dot notation, but in place of an object reference before the dot you use the name
of the class.
Static Methods
A method may also be declared static. A static method can be called without instantiating
the class. An example we have already seen is the Main method in a class, which the
runtime system is able to call without instantiating an object. The Main method must
always be static.
You call a static method by using the dot notation, with the class name in front of the dot.
Because you must call a static method without an instance, a static method can use only
static data members and not instance data members.
Static methods may be declared public or private. A private static method, like other
private methods, may be used as a helper function within a class, but not called from
outside.
Sample Program
Our previous Customer class relied on the user of the class to assign a CustomerId for the
customer. A better approach is to encapsulate assigning an id within the class itself, so
that a unique id will be automatically generated every time a new Customer object is
created. It is easy to implement such a scheme by using a static field nextCustId, which is
used to assign an id. Every time the id is assigned, nextCustId is incremented.
TestCustomer\Step2 demonstrates this solution and also illustrates the use of a static
method. Here is the code defining the Customer class:
// Customer.cs - Step 2

{
static private int nextCustId = 1;
public Customer(string first, string last, string email)
{
CustomerId = nextCustId++;
FirstName = first;
LastName = last;
}
public static int GetNextId()
{
return nextCustId;
}
}
// TestCustomer.cs - Step 2
using System;

{
{
Console.WriteLine("next id = {0}",
Customer.GetNextId());
cust1 = new Customer("John", "Doe",
cust2 = new Customer("Mary", "Smith",
}
private static void ShowCustomer(string label,
Customer cust)
...
Note that the static method GetNextId is accessed through the class Customer and not
through an object reference such as cust1. This program also illustrates the fact that Main
is a static method and is invoked by the runtime without an instance of the TestCustomer
class being created. Since there is no instance, any method of TestCustomer called from
within Main must also be declared static, as illustrated by ShowCustomer.
Static Constructor
Besides having static fields and static methods, a class may also have a static constructor.
A static constructor is called only once, before any object instances have been created. A
static constructor is defined by prefixing the constructor with static. A static constructor
mut take no parameters and has no access modifier (such as public or private).
In a language such as C++, where there can be global variables not attached to any class,
you may initialize a library through the constructor for a global object. In C# there are no
such freestanding global objects, but you can achieve similar initialization through use of
a static constructor. As a somewhat whimsical example of a static constructor, consider
the StaticWorld program, which provides an alternative implementation of "Hello,
World."
// StaticWorld.cs
public class Hello

{
static Hello()
{
System.Console.Write("Hello, ");
}
public static void World()
{
System.Console.WriteLine("World");
}
}
public class World

{
{
Hello.World();
}
}
Constant and Readonly Fields
If you want to make sure that a variable always has the same value, you can assign the
value via an initializer and use the const modifier. Such a constant is automatically static,
and you will access it from outside the class through the class name.
Another situation may call for a one-time initialization at runtime, and after that the value
cannot be changed. You can achieve this effect through a readonly field. Such a field may
be either an instance member or a static member. In the case of an instance member, it
will be assigned in an ordinary constructor. In the case of a static member, it will be
assigned in a static constructor.
The program ConstantHotel illustrates the use of both const and readonly. In both cases,
you will get a compiler error if you try to modify the value.
// ConstantHotel.cs
public class Hotel

{
public const decimal rate = 100m;
public readonly string name;
public Hotel(string name)
{
this.name = name;
}
}
// TestHotel.cs
using System;

{
{
Console.WriteLine("rate = {0:C}", Hotel.rate);
//Hotel.rate = 150m; // illegal
Hotel hotel = new Hotel("Ritz");
Console.WriteLine("hotel name = {0}", hotel.name);
//hotel.name = "Sheraton"; // illegal
}
}
Here is the output:
rate = $100.00
hotel name = Ritz
C# Type System
In C# there is a fundamental distinction between value types and reference types. Value
types have storage allocated immediately on the stack when the variable is declared.
Reference types have storage allocated on the heap, and the variable is only a reference to
the actual data, which can be allocated later.
We have been looking at classes in some detail. A class defines a reference type. In this
section we survey the entire C# type system, including simple types such as int and
decimal. In C# a struct has many similarities to a class but is a value type. Another
important kind of value type in C# is an enum.
We examine later several other important types, including string, array, interface, and
delegate. We will discuss the default values that get assigned to variables when there is
not an explicit initialization. We will see that all types in C# are rooted in a fundamental
base class called object. In C# "everything is an object," and value types are transparently
converted to object references as needed through a process known as boxing. The inverse
process, unboxing, returns an object to the value type from which it came.
Overview of Types in C#
In C# there are three kinds of types:
• Value types
• Reference types
• Pointer types
Value Types
Value types directly contain their data. Each variable of a value type has its own copy of
the data. Value types typically are allocated on the stack and are automatically destroyed
when the variable goes out of scope. Value types include the simple types like int and
decimal, structures, and enumeration types.
Reference Types
Reference types do not contain data directly but only refer to data. Variables of reference
types store references to data, called objects. Two different variables can reference the
same object. Reference types are allocated on the managed heap and eventually get
destroyed through a process known as garbage collection.
Reference types include string, object, class types, array types, interfaces, and delegates.
Pointer Types
Pointer types are used only in unsafe code and will be discussed later in this chapter.
Value Types
In this section we survey all the value types, including the simple types, structures, and
enumerations.
Simple Types
The simple data types are general-purpose value data types, including numeric, character,
and Boolean.
• The sbyte data type is an 8-bit signed integer.

• The byte data type is an 8-bit unsigned integer.
• The short data type is a 16-bit signed integer.
• The ushort data type is a 16-bit unsigned integer.
• The int data type is a 32-bit signed integer.
• The uint data type is a 32-bit unsigned integer.
• The long data type is a 64-bit signed integer.
• The ulong data type is a 64-bit unsigned integer.
• The char data type is a Unicode character (16 bits).
• The float data type is a single-precision floating point.
• The double data type is a double-precision floating point.
• The bool data type is a Boolean (true or false).
• The decimal data type is a decimal type with 28 significant digits (typically used
for financial purposes).
Types in System Namespace
There is an exact correspondence between the simple C# types and types in the System
namespace. C# reserved words are simply aliases for the corresponding types in the
System namespace. Table 3-3 shows this correspondence.
Table 3-3. Types in C# and the System Namespace

C# Reserved Word Type in System Namespace
sbyte System.SByte
byte System.Byte
short System.Int16
ushort System.UInt16
int System.Int32
uint System.UInt32
long System.Int64
ulong System.UInt64
char System.Char
float System.Single
double System.Double
Table 3-3. Types in C# and the System Namespace
C# Reserved Word Type in System Namespace
bool System.Boolean
decimal System.Decimal
Structures
A struct is a value type which can group heterogeneous types together. It can also have
constructors and methods. In C++ the concept of class and struct is very close. In C++ a
class has default visibility of private and a struct has default visibility of public, and that
is the only difference. There is a more fundamental difference in C#.
In C# the key difference between a class and a struct is that a class is a reference type and
a struct a value type. A class must be instantiated explicitly using new. The new instance
is created on the heap, and memory is managed by the system through a garbage-
collection process. Since a default constructor will be created for a struct if none is
defined, a struct declared on the stack will be initialized. You may also use new. A new
instance of a struct is created on the stack, and the instance will be deallocated when it
goes out of scope.
There are different semantics for assignment, whether done explicitly or via call by value
mechanism in a method call. For a class, you will get a second object reference, and both
object references refer to the same data. For a struct, you will get a completely
independent copy of the data in the struct.
A struct is a convenient data structure to use for moving data across a process or machine
boundary, and we will use structs in our case study. For example, we will use a struct to
represent customer data.
public struct CustomerListItem

{
}
Enumeration Types
The final kind of value type is an enumeration type. An enumeration type is a distinct
type with named constants. Every enumeration type has an underlying type, which is one
of the following.
• byte
• short
• int
• long
An enumeration type is defined through an enum declaration.
public enum BookingStatus : byte

{
HotelNotFound, // 0 implicitly
RoomsNotAvailable, // 1 implicitly
Ok = 5 // explicit value
}
If the type is not specified, int is used. By default, the first enum member is assigned the
value 0, the second member 1, and so on. Constant values can be explicitly assigned.
You can make use of an enumeration type by declaring a variable of the type indicated in
the enum declaration (e.g., BookingStatus). You can refer to the enumerated values by
using the dot notation. Here is some illustrative code:
BookingStatus status;
status = hotel.ReserveRoom(name, date);
if (status == BookingStatus.HotelNotFound)
Console.WriteLine("Hotel not found");
...
Reference Types
A variable of a reference type does not directly contain its data but instead provides a
reference to the data stored in the heap. In C# there are the following kinds of reference
types:
• Class
• Array
• Interface
• Delegate
Reference types have a special value null, which indicates the absence of an instance.
We have already examined classes in some detail, and we will look at arrays later in this
chapter. Interfaces and delegates will be covered in Chapter 5.
Class Types
A class type defines a data structure that has fields, methods, constants, and other kinds of
members. Class types support inheritance. Through inheritance a derived class can extend
or specialize a base class. We will discuss inheritance in Chapter 4.
Two classes in the .NET Framework Class Library are so important that they have C#
reserved words as aliases for them: object and string.
Object
The object class type is the ultimate base type for all types in C#. Every C# type derives
directly or indirectly from object. The object keyword in C# is an alias for the predefined
System.Object class. System.Object has methods such as ToString, Equals, and Finalize,
which we will study later.
String
The string class encapsulates a Unicode character string. The string keyword is an alias
for the predefined System.String class. The string type is a sealed class. (A sealed class is
one that cannot be used as the base class for any other classes.)
The string class inherits directly from the root object class. String literals are defined
using double quotes. There are useful built-in methods for string. For now, note that the
Equals method can be used to test for equality of strings.
string a = "hello";
if (a.Equals("hello"))
Console.WriteLine("equal");
else
Console.WriteLine("not equal");
There are also overloaded operators:
if (a == "hello")
...
We will study string in detail later in this chapter.
Default Values
Several kinds of variables are automatically initialized to default values:
• Static variables
• Instance variables of class and struct instances
• Array elements
Local variables are not automatically initialized, and you will get a compiler error
message if you try to use a local variable that has not been initialized.
The default value of a variable of reference type is null.
The default value of a variable of value type is the value assigned in the default
constructor. For simple types this value corresponds to a bit pattern of all zeros:
• For integer types, the default value is 0

• For char, the default value is '\u0000'
• For float, the default value is 0.0f
• For double, the default value is 0.0d
• For decimal, the default value is 0.0m
• For bool, the default value is false
For an enum type, the default value is 0. For a struct type, the default value is obtained by
setting all value type fields to their default values, as described above, and all reference
type fields to null.
Boxing and Unboxing
One of the strong features of C# is that is has a unified type system. Every type, including
the simple built-in types such as int, derive from System.Object. In C# "everything is an
object."
A language such as Smalltalk also has such a feature but pays the price of inefficiency for
simple types. Languages such as C++ and Java treat simple built-in types differently from
objects, thus obtaining efficiency but at the cost of a unified type system.
C# enjoys the best of both worlds through a process known as boxing. Boxing converts a
value type such as int or a struct to an object reference and is done implicitly. Unboxing
converts a boxed value type (stored on the heap) back to an unboxed simple value (stored
on the stack). Unboxing is done through a type cast.
int x = 5;
object o = x; // boxing
x = (int) o; // unboxing
Strings
Characters and strings are very important data types in practical programming. C#
provides a string type, which is an alias for the String class in the System namespace. As
a class type, string is a reference type. Much string functionality, available in all .NET
languages, is provided by the String class. The C# compiler provides additional support
to make working with strings more concise and intuitive. In this section we will first look
at characters and then outline the main features of the String class. We will look at string
input, at the additional support provided by C#, and at the issues of string equality. The
section that follows surveys some of the useful methods of the String class. The section
after that discusses the StringBuilder class.
Characters
C# provides the primitive data type char to represent individual characters. A character
literal is represented by a character enclosed in single quotes.
char ch1 = 'a';
A C# char is represented internally as an unsigned two-byte integer. You can cast back
and forth between char and integer data types.
char ch1 = 'a';

int n = (int) ch1;
n++;
ch1 = (char) n; // ch1 is now 'b'
The relational operators ==, <, >, and so on apply to char.
char ch1 = 'a';

char ch2 = 'b'
if (ch1 < ch2) // expression is true
...
ASCII and Unicode
Traditionally, a one-byte character code called ASCII has been used to represent
characters. ASCII code is simple and compact. But ASCII cannot be employed to
represent many different alphabets used throughout the world.
Modern computer systems prefer to use a two-byte character code called Unicode. Most
modern (and many ancient) alphabets can be represented by Unicode characters. ASCII is
a subset of Unicode, corresponding to the first 255 Unicode character codes. For more
information on Unicode, you can visit the Web site www.unicode.org. C# uses Unicode
to represent characters.
Escape Sequences
You can represent any Unicode character in a C# program by using the special escape
sequence beginning with \u followed by hexadecimal digits.
char A = '\u0041'; // 41 (hex) is 65 (dec) or 'A'

Special escape sequences are provided for a number of standard non-printing characters
and for characters like quotation marks that would be difficult to represent otherwise.
Table 3-4 shows the standard escape sequences in C#.
Table 3-4. Escape Characters in C#

Escape Character Name Value
\' Single quote 0x0027
\" Double quote 0x0022
\\ Backslash 0x005C
\0 Null 0x0000
\a Alert 0x0007
\b Backspace 0x0008
\f Form feed 0x000C
\n New line 0x000A
\r Carriage return 0x000D
\t Horizontal tab 0x0009
\v Vertical tab 0x000B
String Class
The String class inherits directly from Object and is a sealed class, which means that you
cannot further inherit from String. We will discuss inheritance and sealed classes in
Chapter 4. When a class is sealed, the compiler can perform certain optimizations to
make methods in the class more efficient.
Instances of String are immutable, which means that once a string object is created, it
cannot be changed during its lifetime. Operations that appear to modify a string actually
return a new string object. If, for the sake of efficiency, you need to modify a stringlike
object directly, you can make use of the StringBuilder class, which we will discuss in a
later section.
A string has a zero-based index, which can be used to access individual characters in a
string. That means that the first character of the string str is str[0], the second character is
str[1], and so on.
By default, comparison operations on strings are case-sensitive, although there is an

overloaded version of the Compare method that permits case-insensitive comparisons.
The empty string should be distinguished from null. If a string has not been assigned, it
will be a null reference. Any string, including the empty string, compares greater than a
null reference. Two null references compare equal to each other.
Language Support
The C# language provides a number of features to make working with strings easier and
more intuitive.
String Literals and Initialization
You can define a string literal by enclosing a string of characters in double quotes.
Special characters can be represented using an escape sequence, as discussed earlier in
the chapter. You may also define a "verbatim" string literal using the @ symbol. In a
verbatim string, escape sequences are not converted but are used exactly as they appear.
If you want to represent a double quote inside a verbatim string, use two double quotes.
The proper way to initialize a string variable with a literal value is to supply the literal
after an equals sign. You do not need to use new as you do with other data types. Here are
some examples of string literals and initializing string variables.
string s1 = "bat";
string path1 = "c:\\OI\\NetCs\\Chap3\\Concat";
string path = @"c:\OI\NetCs\Chap3\Concat\";
string greeting = @"""Hello, world""";
Concatenation
The String class provides a method Concat for concatenating strings. In C# you can use
the operators + and += to perform concatenation. The following program illustrates string
literals and concatenation.
// Concat.cs
using System;
public class Concat

{
{
str[0],string s1 = "bat";
Console.WriteLine("s1 = {0}", s1);
string s2 = "man";
Console.WriteLine("s2 = {0}", s2);
s1 += s2;
Console.WriteLine(s1);
string path1 = "c:\\OI\\NetCs\\Chap3\\Concat";
Console.WriteLine("path1 = {0}", path1);
string path = @"c:\OI\NetCs\Chap3\Concat\";
string file = "Concat.cs";
path = path + file;
Console.WriteLine(path);
string greeting = @"""Hello, world""";
Console.WriteLine(greeting);
}
}
Here is the output:
s1 = bat
s2 = man
batman
path1 = c:\OI\NetCs\Chap3\Concat
c:\OI\NetCs\Chap3\Concat\Concat.cs
"Hello, world"
Index
You can extract an individual character from a string using a square bracket and a zero-
based index.
string s1 = "bat";
char ch = s1[0]; // contains 'b'
Relational Operators
In general, for reference types, the == and != operators check if the object references are
the same, not whether the contents of the memory locations referred to are the same.
However, the String class overloads these operators, so that the textual content of the
strings is compared. The program StringRelation illustrates using these relational
operators on strings. The inequality operators, such as <, are not available for strings; use
the Compare method.
String Equality
To fully understand issues of string equality, you should be aware of how the compiler
stores strings. When string literals are encountered, they are entered into an internal table
of string identities. If a second literal is encountered with the same string data, an object
reference will be returned to the existing string in the table; no second copy will be made.
As a result of this compiler optimization, the two object references will be the same, as
represented in Figure 3-3.
Figure 3-3. Object references to a string literal refer to the same storage.
You should not be misled by this fact to conclude that two object references to the same
string data will always be the same. If the contents of the string get determined at
runtime, for example, by the user inputting the data, the compiler has no way of knowing
that the second string should have an identical object reference. Hence you will have two
distinct object references, which happen to refer to the same data, as illustrated in Figure
3-4.
Figure 3-4. Two distinct object references, which happen to refer to the same
data.
As discussed, when strings are checked for equality, either through the relational operator
== or through the Equals method, a comparison is made of the contents of the strings, not
of the object references. So in both the previous cases the strings a and b will check out
as equal. You have to be more careful with other reference types, where reference
equality is not the same as content equality.
String Comparison
The fundamental way to compare strings for equality is to use the Equals method of the
String class. There are several overloaded versions of this function, including a static
version that takes two string parameters and a nonstatic version that takes one string
parameter that is compared with the current instance. These methods perform a case-
sensitive comparison of the contents of the strings. A bool value of true or false is
returned.
If you wish to perform a case-insensitive comparison, you may use the Compare method.
This method has several overloaded versions, all of them static. Two strings, s1 and s2,
are compared. An integer is returned expressing the lexical relationship between the two
strings, as shown in Table 3-5.
Table 3-5. Return Values of the Compare Method
Relationship Return Value
s1 less than s2 negative integer
s1 equal to s2 0
s1 greater than s2 positive integer
A third parameter allows you to control the case sensitivity of the comparison. If you use
only two parameters, a case-sensitive comparison is performed. The third parameter is a
bool. A value of false calls for a case-sensitive comparison, and a value of true calls for
ignoring case.
The program StringCompare illustrates a number of comparisons, using both the Equal
and Compare methods.
String Input
The Console class has methods for inputting characters and strings. The Read method
reads in a single character (as an int). The ReadLine method reads in a line of input,
terminated by a carriage return, line feed, or combination, and will return a string. In
general, the ReadLine method is the easier to use and synchronizes nicely with Write and
WriteLine. The program ReadStrings illustrates reading in a first name, a middle initial,
and a last name. All input is done via ReadLine. The middle initial as a character is
determined by extracting the character at position 0.
Our InputWrapper class has a method getString, which provides a prompt and reads in a
string.
String Methods and Properties
In this section we will survey a few useful methods and properties of the String class.
Many of the methods have various overloaded versions. We show a representative
version. Consult the online documentation for details on these and other methods. The
program StringMethods demonstrates all the examples that follow.
Length
public int Length {get;}
This property returns the length of a string. Notice the convenient shorthand notation that
is used for declaring a property.
string str = "hello";
int n = str.Length; // 5
ToUpper
public string ToUpper ();
This method returns a new string in which all characters of the original string have been
converted to uppercase.
str = "goodbye";
str = str.ToUpper(); // GOODBYE
ToLower
public string ToLower ();
This method returns a new string in which all characters of the original string have been
converted to lowercase.
str = str.ToLower(); // goodbye
Substring
public string Substring(int startIndex, int length);
This method returns a substring that starts from a specified index position in the value
and continues for a specified length. Remember that in C# the index of the first character
in a string is 0.
string sub = str.Substring(4,3); // bye
IndexOf
public int IndexOf(string value);
This method returns the index of the first occurrence of the specified string. If the string
is not found, -1 is returned.
str = "goodbye";
int n1 = str.IndexOf("bye"); // 4
int n2 = str.IndexOf("boo"); // -1
StringBuilder Class
As we have discussed, instances of the String class are immutable. As a result, when you
manipulate instances of String, you are frequently obtaining new String instances.
Depending on your applications, creating all these instances may be expensive. The .NET
library provides a special class StringBuilder (located in the System.Text namespace) in
which you may directly manipulate the underlying string without creating a new instance.
When you are done, you can create a String instance out of an instance of StringBuilder
by using the ToString method.
A StringBuilder instance has a capacity and a maximum capacity. These capacities can be
specified in a constructor when the instance is created. By default, an empty
StringBuilder instance starts out with a capacity of 16. As the stored string expands, the
capacity will be increased automatically. The program StringBuilderDemo provides a
simple demonstration of using the StringBuilder class. It shows the starting capacity and
the capacity after strings are appended. At the end, a String is returned.
// StringBuilderDemo.cs
using System;
using System.Text;
public class StringBuilderDemo

{
{
StringBuilder build = new StringBuilder();
Console.WriteLine("capacity = {0}", build.Capacity);
build.Append("This is the first sentence.\n");
build.Append("This is the second sentence.\n");
build.Append("This is the last sentence.\n");
string str = build.ToString();
Console.Write(str);
}
}
Here is the output:
capacity = 16
capacity = 34
capacity = 70
capacity = 142
This is the first sentence.
This is the second sentence.
This is the last sentence.
Arrays and Indexers
Arrays are another important data type in practical programming. In C# arrays are
objects. They are a reference data type. They are based on the class System.Array and so
inherit the properties and methods of this class. After examining one-dimensional arrays,
we examine two higher-dimensional varieties. A "jagged" array is an array of arrays, and
each row can have a different number of elements. In "rectangular" arrays, all rows have
the same number of elements. Arrays are a special kind of collection, which means that
the foreach loop can be used in C# for iterating through array elements.
We conclude the section with a discussion of indexers, which provides a way to access
encapsulated data in a class with an array notation.
Arrays
An array is a collection of elements with the following characteristics.
• All array elements must be of the same type. The element type of an array can be
any type, including an array type. An array of arrays is often referred to as a
jagged array.
• An array may have one or more dimensions. For example, a two-dimensional
array can be visualized as a table of values. The number of dimensions is known
as the array's rank.
• Array elements are accessed using one or more computed integer values, each
known as an index. A one-dimensional array has one index.
• In C# an array index starts at 0, as in other C family languages.
• The elements of an array are created when the array object is created. The
elements are automatically destroyed when there are no longer any references to
the array object.
One-Dimensional Arrays
An array is declared using square brackets [] after the type, not after the variable.
int [] a; // declares an array of int
Note that the size of the array is not part of its type. The variable declared is a reference
to the array.
You create the array elements and establish the size of the array using the new operator.
a = new int[10]; // creates 10 array elements

The new array elements start out with the appropriate default values for the type (0 for
int).
You may both declare and initialize array elements using curly brackets, as in C/C++.
int a[] = {2, 3, 5, 7, 11};
You can indicate you are done with the array elements by assigning the array reference to
null.
a = null;
The garbage collector is now free to deallocate the elements.
System.Array
Arrays are objects. System.Array is the abstract base class for all array types.
Accordingly, you can use the properties and methods of System.Array for any array. Here
are some examples:
• Length is a property that returns the number of elements currently in the array.
• Sort is a static method that will sort the elements of an array.
• BinarySearch is a static method that will search for an element in a sorted array,
using a binary search algorithm.
int [] array = {5, 2, 11, 7, 3};

Array.Sort(a); // sorts the array
for (int i = 0; i < a.Length; i++)
Console.Write("{0} ", a[i]);
Console.WriteLine();
int target = 5;
int index = Array.BinarySearch(a, target);
if (index < 0)
Console.WriteLine("{0} not found", target);
else
Console.WriteLine("{0} found at {1}", target, index);
A complete program containing the code shown above can be found in ArrayMethods.
Here is the output:
2 3 5 7 11
5 found at 2
Sample Program
The program ArrayDemo is an interactive test program for arrays. A small array is
created initially, and you can create new arrays. You can populate an array either with a
sequence of square numbers or with random numbers. You can sort the array, reverse the
array, and perform a binary search (which assumes that the array is sorted in ascending
order). You can destroy the array by assigning the array reference to null.
Interfaces for System.Array
If you look at the documentation for methods of System.Array, you will see many
references to various interfaces, such as IComparable. By using such interfaces you can
control the behavior of methods of System.Array. For example, if you want to sort an
array of objects of a class that you define, you must implement the interface IComparable
in your class so that the Sort method knows how to compare elements to carry out the
sort. The .NET Framework provides an implementation of IComparable for all the
primitive types. We will come back to this point after we discuss interfaces in Chapter 5.
Random-Number Generation
The ArrayDemo program contains the following code for populating an array with
random integers between 0 and 100.
Random rand = new Random();

for (int i = 0; i < size; i++)
{
array[i] = rand.Next(100);
}
The .NET Framework provides a useful class, Random, in the System namespace that can
be used for generating pseudorandom numbers for simulations.
Constructors
There are two constructors:
Random(); // uses default seed

Random(int seed); // seed is specified
The default seed is based on date and time, resulting in a different stream of random
numbers each time. By specifying a seed, you can produce a deterministic stream.
Next Methods
There are three overloaded Next methods that return a random int.
int Next();
int Next(int maxValue);
int Next(int minValue, int maxValue);
The first method returns an integer greater than or equal to zero and less than
Int32.MaxValue. The second method returns an integer greater than or equal to zero and
less than maxValue. The third method returns an integer greater than or equal to minValue
and less than or equal to maxValue.
NextDouble Method
The NextDouble method produces a random double between 0 and 1.
double NextDouble();
The return value r is in the range: 0 <= r < 1.
Jagged Arrays
You can declare an array of arrays, or a "jagged" array. Each row can have a different
number of elements.
int [][] binomial;
You then create the array of rows, specifying how many rows there are (each row is itself
an array).
binomial = new int [rows][];
Next you create the individual rows.
binomial[i] = new int [i+1];
Finally you can assign individual array elements.
binomial[0][0] = 1;
The example program Pascal creates and prints Pascal's triangle using a two-dimensional
jagged array. Higher-dimensional jagged arrays can be created following the same
principles.
Rectangular Arrays
C# also permits you to define rectangular arrays, where all rows have the same number of
elements. First you declare the array.
int [,] MultTable;
Then you create all the array elements, specifying the number of rows and columns.
MultTable = new int[rows, columns];
Finally you can assign individual array elements.
MultTable[i,j] = i * j;
The example program RectangularArray creates and prints out a multiplication table.
Higher-dimensional rectangular arrays can be created following the same principles.
Arrays as Collections
The class System.Array supports the IEnumerable interface. Hence arrays can be treated
as collections, a topic we will discuss in Chapter 5. This means that a foreach loop can be
used to iterate through the elements of an array.
The Pascal example code contains nested foreach loops to display the jagged array. The
outer loop iterates through all the rows, and the inner loop iterates through all the
elements within a row.
// Pascal.cs
...
Console.WriteLine(
"Pascal triangle via nested foreach loop");
foreach (int[] row in binomial)
{
foreach (int x in row)
{
Console.Write("{0} ", x);
}
}
Indexers
C# provides various ways to help the user of a class access encapsulated data. Earlier in
the chapter we saw how properties can provide access to a single piece of data associated
with a class, making it appear like a public field. In this section we will see how indexers
provide a similar capability for accessing a group of data items, using an array index
notation. Indexers can be provided when there is a private array or other collection.
The program TestHotel\Step3 provides an illustration. This version of the Hotel class
adds the capability to make hotel reservations, and the private array reservations stores a
list of reservations in the form of ReservationListItem structure instances. The Hotel class
provides the readonly property NumberReservations for the number of reservations in
this list, and it provides a read-write indexer for access to the elements in this list. Note
use of the keywords this and value in the indexer, which has a general syntax similar to
that of properties.
using System;
public struct ReservationListItem

{
public int ReservationId;
public string HotelName;
public string City;
public DateTime ArrivalDate;
public DateTime DepartureDate;
public int NumberDays;
}
...
public class Hotel

{
private int number;
private const int MAXDAY = 366;
private int[] numGuests;
private int nextReservation = 0;
private int nextReservationId = 1;
private const int MAXRESERVATIONS = 100;
private ReservationListItem[] reservations;
...
public int NumberReservations
{
get
{
return nextReservation;
}
}
public ReservationListItem this[int index]
{
get
{
return reservations[index];
}
set
{
reservations[index] = value;
}
}
The test program TestHotel.cs illustrates reading and writing individual array elements
using the index notation.
// Change the CustomerId of first reservation

ReservationListItem item = ritz[0];
item.CustomerId = 99;
ritz[0] = item;
ShowReservations(ritz);
More about Methods
In this section we look at several other topics pertaining to methods in C#:
• Parameter passing
• Variable-length parameter lists
• Method overloading
• Operator overloading
Parameter Passing
Programming languages have different mechanisms for passing parameters. In the C

family of languages the standard is "call by value." This means that the actual data values
themselves are passed to the method. Typically, these values are pushed onto the stack,
and the called function obtains its own independent copy of the values. Any changes
made to these values will not be propagated back to the calling program. C# provides this
mechanism of parameter passing as the default, but C# also supports "reference"
parameters and "output" parameters.
Some terminology will help us in the following discussion. Storage is allocated on the
stack for method parameters. This storage area is known as the activation record. It is
popped when the method is no longer active. The formal parameters of a method are the
parameters as seen within the method. They are provided storage in the activation record.
The actual parameters of a method are the expressions between commas in the parameter
list of the method call.
int sum = SimpleMath.Add(5, 7); // actual parameters

//are 5 and 7
...
public static int Add(int x, int y)
{ // formal parameters
//are x and y
...
}
Value Parameters
Parameter passing is the process of initializing the storage of the formal parameters by
the actual parameters. The default method of parameter passing in C# is call-by-value, in
which the values of the actual parameters are copied into the storage of the formal
parameters. Call-by-value is "safe," because the method never directly accesses the actual
parameters, only its own local copies. But there are drawbacks to call-by-value:
• There is no direct way to modify the value of an argument. You may use the
return type of the method, but that allows you to pass only one value back to the
calling program.
• There is overhead in copying a large object.
The overhead in copying a large object is borne when you pass a struct instance. If you
pass a class instance, or an instance of any other reference type, you are passing only a
reference and not the actual data itself. This may sound like "call-by-reference," but what
you are actually doing is passing a reference by value.
Reference Parameters
Consider a situation in which you want to pass more than one value back to the calling
program. C# provides a clean solution through reference parameters. You declare a
reference parameter with the ref keyword, which is placed before both the formal
parameter and the actual parameter. A reference parameter does not result in any copying
of a value. Instead, the formal parameter and the actual parameter refer to the same
storage location. Thus, changing the formal parameter will result in the actual parameter
changing, as both are referring to exactly the same storage location.
The program ReferenceMath illustrates using ref parameters. There is a single method,
Calculate, which passes back two values as reference parameters.
// ReferenceMath.cs
public class ReferenceMath

{
public static void Calculate(int x, int y,
ref int sum, ref int prod)
{
sum = x + y;
prod = x * y;
}
}
Notice the use of the ref keyword in front of the third and fourth parameters. Here is the
test program:
// TestReferenceMath.cs
using System;
public class TestReferenceMath

{
{
int sum = 0, product = 0;
MultipleMath.Calculate(5, 7, ref sum, ref product);
Console.WriteLine("sum = {0}", sum);
Console.WriteLine("product = {0}", product);
}
}
Again we need to have the ref keyword in front of the parameters. It is also necessary to
initialize the variables before using them as reference parameters.
Output Parameters
A reference parameter is really designed for two-way communication between the calling
program and the called program, both passing data in and getting data out. Thus there is a
requirement that reference parameters be initialized before their use. In the case we have
just looked at, where we are only obtaining output, initializing the variables only for them
to be assigned new values is rather pointless. C# provides for this case with output
parameters. Use the keyword out wherever you would use the keyword ref. Then you do
not have to initialize the variable before use. Within the method you must be sure to
assign the out parameter, and you could not use it before such an assignment. The
program OutputMath illustrates the use of output parameters.
Method Overloading
In a traditional programming language such as C, you need to create unique names for all
your methods. If methods do basically the same thing but apply only to different data
types, it becomes tedious to create unique names. For example, suppose you have a
FindMax method that can find the maximum of two int or two long or two string. If we
need to come up with a unique name for each method, we would have to create method
names such as FindMaxInt, FindMaxLong, and FindMaxString.
In C#, as in other object-oriented languages such as C++ and Java, you may overload
method names. That is, different methods can have different names, if they have different
signatures. Two methods have the same signature if they have the same number of
parameters, the parameters have the same data types, and the parameters have the same
modifiers (none, ref, or out). The return type does not contribute to defining the signature
of a method.
At runtime the compiler will resolve a given invocation of the method by trying to match
up the actual parameters with formal parameters. A match occurs if the parameters match
exactly or if they can match through an implicit conversion. For the exact matching rules,
consult the C# Language Specification.
The program OverloadDemo illustrates method overloading. The method FindMax is

overloaded to take either long or string parameters. The method is invoked three times,
for int, long, and string parameters. There is an exact match for the case of long and
string. The call with int actual parameters can resolve to the long version, because there is
an implicit conversion of int into long.
Modifiers as Part of the Signature
It is important to understand that if methods have identical types for their formal
parameters, but differ in a modifier (none, ref, or out), then the methods have different
signatures. The program OverloadHotel provides an illustration. We have two RaisePrice
methods. In the first, the hotel is passed as a value parameter. In the second, the hotel is
passed as a reference parameter. These methods have different signatures.
// HotelTest.cs
using System;
public class HotelDemo

{
{
Hotel ritz = new Hotel("Boston", "Ritz", 100,
200.00m);
Hotel flop = new Hotel("Podunk", "Flop", 50, 20.00m);
// The Ritz before and after
Console.WriteLine("Before price hike");
ritz.Show();
RaisePrice(ritz, 50.00m);
Console.WriteLine("After price hike");
ritz.Show();
// The Flop before and after -- use ref version
Console.WriteLine("Before price hike");
flop.Show();
RaisePrice(ref flop, 50.00m);
Console.WriteLine("After price hike");
flop.Show();
}
private static void RaisePrice(Hotel hotel,
decimal delta)
{
hotel.cost += delta;
Console.WriteLine("new cost = {0:C}", hotel.cost);
}
private static void RaisePrice(ref Hotel hotel,
decimal delta)
{
hotel.cost += delta;
Console.WriteLine("new cost = {0:C}", hotel.cost);
}
}
Variable-Length Parameter Lists
Our FindMax methods in the previous section were very specific with respect to the
number of parameters—there were always exactly two parameters. Sometimes you may
want to be able to work with a variable number of parameters—for example, to find the
maximum of two, three, four, or more numbers. C# provides a params keyword, which
you can use to indicate that an array of parameters is provided. Sometimes you may want
to provide both a general version of your method that takes a variable number of
parameters and also one or more special versions that take an exact number of
parameters. The special version will be called in preference, if there is an exact match.
The special versions are more efficient. The program VariableMax illustrates a general
FindMax method that takes a variable number of parameters. There is also a special
version that takes two parameters. Each method prints out a line identifying itself, so you
can see which method takes precedence. Here is the program:
// VariableMax.cs
using System;
public class VariableMax

{
{
Console.WriteLine("max of {0}, {1} = {2}",
5,7,FindMax(5,7));
Console.WriteLine("max of {0}, {1}, {2} = {3}",
500,5,7,FindMax(500,5,7));
Console.WriteLine("max of {0}, {1}, {2}, {3} = {4}",
500,5,7,80,FindMax(500,5,7,80));
}
static int FindMax(int a, int b)
{
Console.WriteLine("FindMax with Two Parameters");
return a < b ? b : a;
}
static int FindMax(params int[] args)
{
Console.WriteLine(
"FindMax with Variable Number of Parameters");
int imax = Int32.MinValue;
for (int i = 0; i < args.Length; i++)
{
if (args[i] > imax)
imax = args[i];
}
return imax;
}
}
Here is the output:
FindMax with Two Parameters

max of 5, 7 = 7
FindMax with Variable Number of Parameters
max of 500, 5, 7 = 500
FindMax with Variable Number of Parameters
max of 500, 5, 7, 80 = 500
Operator Overloading
C#, like C++ but unlike Java, supports operator overloading. The idea is that certain
method invocations can be implemented more concisely using operators rather than
method calls. Suppose we have a class Matrix that has static methods to add and multiply
matrices. Using methods, we could write a matrix expression like this:
Matrix a, b, c, d;
// code to initialize the object references
d = Matrix.Multiply(a, (Matrix.Add(b, c));
If we overload the operators + and *, we can write this code more succinctly:
d = a * (b + c);
You cannot create a brand new operator, but you can overload many of the existing C#
operators to be an alias for a static method. For example, given the static method Add in
the Matrix class:
class Matrix
{
...
public static Matrix Add(Matrix x, Matrix y)
{
you can write instead:
public static Matrix operator+(Matrix x, Matrix y)

All of the rest of the class implementation code stays the same, and you can then use
operator notation in client code. Operator declarations, such as operator+ shown above,
must obey the following rules:
• Operators must be public and static and may not have any other modifiers.
• Operators take only value parameters and not reference or output parameters.
• Operators must have a signature that differs from the signatures of all other
operators in the class.
The program OperatorOverloadDemo provides a simple example of operator

overloading. The + operator is overloaded in the HotelList class to add a hotel to an array
of hotels. In C# if you overload a binary operator such as +, the corresponding compound
assignment operator += will be overloaded for you automatically by the compiler. Thus,
in our test program, we add the hotel objects ritz and sheraton to the list of hotels using
the + and += operators.
// OperatorOverloadDemo.cs
using System;
public class OperatorOverloadDemo

{
{
HotelList list = new HotelList();
Hotel ritz = new Hotel("Atlanta", "Ritz");
Hotel sheraton = new Hotel("Boston", "Sheraton");
list = list + ritz;
list += sheraton;
list.ShowHotels();
}
}
Exceptions
An inevitable part of programming is dealing with error conditions of various sorts. This
section introduces the exception-handling mechanism of C#, beginning with a discussion
of the fundamentals of error processing and various alternatives that are available. We
then carefully go through the C# exception mechanism, which includes a try block, catch
handlers, and a finally block. You can raise exceptions by means of a throw statement.
The .NET class library provides an Exception class, which you can use to pass
information about an exception that occurred. To further specify your exception and to
pass additional information, you can derive your own class from Exception. When
handling an exception you may want to throw a new exception. In such a case you can
use the "inner exception" feature of the Exception class to pass the original exception on
with your new exception.
Exception Fundamentals
The traditional way to deal with runtime errors is to have the functions you call return a
status code. The status code may have a particular value for a good return and other
values to denote various error conditions. The calling function checks this status code,
and if an error was encountered, it performs appropriate error handling. This function in
return may pass an error code to its calling function, and so on up the call stack.
Although straightforward, this mechanism has a number of drawbacks. The basic

disadvantage is lack of robustness. The called function may have impeccable error-
checking code and return appropriate error information, but all this information is useless
if the calling function does not make use of it. The program may continue operation as if
nothing were amiss and some time later crash for some mysterious reason. Also, status
codes are non-standard. A 0 may indicate success in one case but failure in another. And
the caller and callee have to agree on return codes and their meaning.
Another disadvantage is that every function in the call stack must participate in the
process, or the chain of error information will be broken. Also, unusual flow control can
leave memory hanging without being deallocated.
Furthermore, in languages such as C# that have constructors and overloaded operators,

there isn't a return value for some operations.
.NET Exception Handling
C# provides an exception mechanism that can be used for reporting and handling errors.
An error is reported by "throwing" an exception. The error is handled by "catching" the
exception. This mechanism is similar in concept to that for exceptions in C++ and Java.
Exceptions are implemented in .NET by the Common Language Runtime, so exceptions

can be thrown in one .NET language and caught in another. The exception mechanism
involves the following elements:
• Code that might encounter an exception should be enclosed in a try block.

• Exceptions are caught in a catch block.
• An Exception object is passed as a parameter to catch. The data type of the
Exception object is System.Exception or a derived type.
• You may have multiple catch blocks. A match is made based on the data type of
the Exception object.
• An optional finally clause contains code that will be executed whether or not an
exception is encountered.
• In the called method, an exception is raised through a throw statement.
Exception Flow of Control
The general structure of code which might encounter an exception is shown below:
try
{
// code that might cause an exception to be thrown
}
catch (ExceptionClass1 e)
{
// code to handle this type of exception
}
catch (ExceptionClass2 e)
{
// code to handle this other type of exception
}
// possibly more catch handlers
// optional finally clause (discussed later)
// statements after try ... catch
finally
{
// cleanup code that is executed whether or not

// an exception is caught or if catch handler itself
// throws an exception
}
Each catch handler has a parameter specifying the data type of exception that it can
handle. The exception data type can be System.Exception or a class ultimately derived
from it. If an exception is thrown, the first catch handler that matches the exception data
type is executed, and then control passes to the statement just after the catch block(s). If
no handler is found, the exception is thrown to the next higher "context" (e.g., the
function that called the current one). If no exception is thrown inside the try block, all the
catch handlers are skipped.
Context and Stack Unwinding
As the flow of control of a program passes into nested blocks, local variables are pushed
onto the stack and a new "context" is entered. Likewise a new context is entered on a
method call, which also pushes a return address onto the stack.
If an exception is not handled in the current context, it is passed to successively higher

contexts until it is finally handled (or else is "uncaught" and is handled by a default
system handler).
When the higher context is entered, C# adjusts the stack properly, a process known as
stack unwinding. In C# exception handling, stack unwinding involves both setting the
program counter and cleaning up variables (popping stack variables and marking heap
variables as free, so that the garbage collector can deallocate them).
Example Program
Now let's look at some code that illustrates the principles we have discussed so far. We
will use a simplified version of our Hotel class. This hotel accepts reservations for only a
single date. There is a property Capacity and there are methods MakeReservation and
CancelReservation. A reservation has an id, a customer name, and the number of rooms
requested. (In this example we have added a feature. Previously, a customer could reserve
only a single room. We are now allowing multiple room requests. This is to simplify
exercising our program to bump against the exception condition of exceeding the
capacity of the hotel.) There is a property, NumberReservations, and an indexer to allow
the calling program to access the reservation list.
There are several possible exceptions:
• User does not request a positive number of rooms.

• Room request exceeds the capacity of the hotel.
• Index out of range when attempting to store reservation in array of reservations.
The first two exceptions are thrown explicitly by our Hotel class, and the index out-of-
range exception is thrown by the .NET library.
Our example program is in the directory HotelException\Step1.
// HotelException.cs - Step 1
using System;

{
public string CustomerName;
public int NumberRooms;
}
public class Hotel

{
private int capacity;
private int numGuests;
private int nextReservation = 0;
private int nextReservationId = 1;
private const int MAXRESERVATIONS = 3;
private ReservationListItem[] reservations;
public Hotel(int capacity)
{
this.capacity = capacity;
reservations =
new ReservationListItem[MAXRESERVATIONS];
}
public int MakeReservation(string cust, int rooms)
{
// Requested number of rooms should be positive
if (rooms <= 0)
throw new Exception(
"Please request a positive number of rooms");
// Check if rooms are available
if (numGuests + rooms > capacity)
throw new Exception("Rooms not available");
// Reserve the room for requested dates
numGuests += rooms;
// Fill in information for reservation
ReservationListItem item;
item.ReservationId = nextReservationId++;
item.CustomerName = cust;
item.NumberRooms = rooms;
// Add reservation to reservation list and return
// reservation id
reservations[nextReservation++] = item;
return item.ReservationId;
}
...
The next code fragment is the test program. Notice that we place the entire body of the
command-processing loop inside a try block. The catch handler prints an error message
that is passed within the exception object. Then, after either normal processing or
displaying an error message, a new command is read in. This simple scheme provides
reasonable error processing, as a bad command will not be acted upon, and the user will
have an opportunity to enter a new command.
// Test.cs
using System;

{
{
Hotel hotel = new Hotel(10);
ShowHotel(hotel);
string cmd;
Console.WriteLine("Enter command, quit to exit");
cmd = iw.getString("H> ");
while (! cmd.Equals("quit"))
{
try
{
if (cmd.Equals("new"))
{
int capacity = iw.getInt("capacity: ");
hotel = new Hotel(capacity);
ShowHotel(hotel);
}
else if (cmd.Equals("book"))
{
string customer =
iw.getString("customer name: ");
int rooms = iw.getInt("number of rooms: ");
int id =
hotel.MakeReservation(customer, rooms);
Console.WriteLine(
"Reservation has been booked");
Console.WriteLine(
"ReservationId = {0}", id);
}
else if (cmd.Equals("cancel"))
{
int id = iw.getInt("reservation id: ");
hotel.CancelReservation(id);
}
else if (cmd.Equals("show"))
ShowReservations(hotel);
else
hotelHelp();
}
catch (Exception e)
{
Console.WriteLine("Exception: {0}", e.Message);
}
}
}
Here is a transcript of a sample run. We try several kinds of errors.
The hotel has 10 rooms

Enter command, quit to exit
H> book
customer name: bob
number of rooms: xxx
Exception: Input string was not in a correct format.
H> book
customer name: bob
number of rooms: -5
Exception: Please request a positive number of rooms
H> book
customer name: bob
number of rooms: 5
Reservation has been booked
ReservationId = 1
H> book
customer name: mary
number of rooms: 6
Exception: Rooms not available
H> book
customer name: mary
number of rooms: 3
ReservationId = 2
H> book
customer name: david
number of rooms: 1
ReservationId = 3
H> show
1 bob 5
2 mary 3
3 david 1
H> book
customer name: ellen
number of rooms: 1
Exception: Exception of type
System.IndexOutOfRangeException was thrown.
H>
Notice that we threw two of the exceptions ourselves. A third (entering "xxx" for the
number of rooms) was caught by the .NET library inside our InputWrapper class. A
fourth (index out of range) was also caught by .NET, inside the Hotel class. Our catch
handler deals with all these different exceptions in a simple, uniform manner.
System.Exception
The System.Exception class provides a number of useful methods and properties for
obtaining information about an exception.
• Message returns a text string providing information about the exception. This
message is set when the exception object is constructed. If no message is
specified, a generic message will be provided indicating the type of the exception.
The Message property is read-only. (Hence, if you want to specify your own
message, you must construct a new exception object, as done in the example
above.)
• StackTrace returns a text string providing a stack trace at the place where the
exception arose.
• InnerException holds a reference to another exception. When you throw a new
exception, it is desirable not to lose the information about the original exception.
The original exception can be passed as a parameter when constructing the new
exception. The original exception object is then available through the
InnerException property of the new exception. (We will provide an example of
using inner exceptions later in this chapter.)
User-Defined Exception Classes
You can do basic exception handling using only the base Exception class, as previously
illustrated. In order to obtain finer-grained control over exceptions, it is frequently useful
to define your own exception class, derived from Exception. You can then have a more
specific catch handler that looks specifically for your exception type. You can also define
other members in your derived exception class, so that you can pass additional
information to the catch handler.
We will illustrate by enhancing the MakeReservation method of our Hotel class. We want
to distinguish between the two types of exceptions we throw. The one type is essentially
bad input data (a nonpositive value). We will continue to handle this exception in the
same manner as before (that is, bad input data gives rise to a format exception, thrown by
.NET library code). We will define a new exception class RoomException to cover the
case where the hotel does not have enough rooms to fulfill the request. (In this case we
want to allow the user an opportunity to submit another reservation request with fewer
rooms.) Our example program is HotelException\Step2. Here is the definition of our new
exception class. This class is defined using inheritance, which we will discuss in Chapter
4, where we will explain the "base(message)" syntax.
public class RoomException : Exception

{
private int available;
public RoomException(string message, int available)
: base(message)
{
this.available = available;
}
public int Available
{
get
{
return available;
}
}
}
Note that we define a property Available that can be used to retrieve the information
about how many rooms are available. The constructor of our exception class takes two
parameters. The first is an error message string, and the second is the number of rooms
available. We pass the message string to the constructor of the base class. We must also
modify the code of the Hotel class to throw our new type of exception when too many
rooms are requested.
// HotelException.cs - Step 2
...
public class Hotel
{
...
public int MakeReservation(string cust, int rooms)
{
// Requested number of rooms should be positive
if (rooms <= 0)
"Please request a positive number of rooms");
// Check if rooms are available
int available = capacity - numGuests;
if (rooms > available)
throw new RoomException(
"Rooms not available", available);
...
Finally we modify the code in our test program that processes the "book" command. We
place the call to MakeReservation inside another try block, and we provide a catch
handler for a RoomException. In this catch handler we allow the user an opportunity to
request fewer rooms. Here is the code:
...
else if (cmd.Equals("book"))
{
string customer = iw.getString("customer name: ");
int rooms = iw.getInt("number of rooms: ");
int id;
try
{
id = hotel.MakeReservation(customer, rooms);
}
catch (RoomException e)
{
Console.WriteLine(
"{0} rooms are available", e.Available);
// try again
rooms = iw.getInt("number of rooms: ");
id = hotel.MakeReservation(customer, rooms);
}
Console.WriteLine("Reservation has been booked");
Console.WriteLine("ReservationId = {0}", id);
...
Here is a transcript of a sample run of our program:

H> book
customer name: bob
number of rooms: 11
Exception: Rooms not available
10 rooms are available
number of rooms: 5
ReservationId = 1
Structured Exception Handling
One of the principles of structured programming is that a block of code should have a
single entry point and a single exit point. The single exit point is convenient, because you
can consolidate cleanup code in one place. The goto statement is usually bad, because it
facilitates breaking this principle. But there are other ways to violate the principle of a
single exit point, such as multiple return statements from a method.
Multiple return statements may not be too bad, because these may be encountered during
normal, anticipated flow of control. But exceptions can cause a particular difficulty, since
they interrupt the normal flow of control. In a common scenario you can have at least
three ways of exiting a method:
• No exception is encountered, and any catch handlers are skipped.

• An exception is caught, and control passes to a catch handler and then to the code
after the catch handlers.
• An exception is caught, and the catch handler itself throws another exception.
Then code after the catch handler will be bypassed.
The first two cases do not present a problem, as control passes to the code after the catch
handlers. But the third case is a source of difficulty.
Finally Block
The structured exception handling mechanism in C# resolves this problem with a finally
block. The finally block is optional, but if present must appear immediately after the
catch handlers. It is guaranteed, in all three cases described above, that the code in the
finally block will always execute before the method is exited.
We illustrate use of finally in the "cancel" command of our Hotel example. See the
directory HotelException\Step3. There are several ways to exit this block of code, and the
user might become confused about whether a cancellation was actually made or not. We
insert a finally block which will always display all the reservations. Here is the code:
else if (cmd.Equals("cancel"))
{
int id;
id = iw.getInt("reservation id: ");
try
{
}
catch (Exception e)
{
id = iw.getInt("reservation id: ");
}
finally
{
ShowReservations(hotel);
}
}
It is instructive to compare the "book" and "cancel" commands. In the "book" command
there is code after the catch handler. This code will be executed if the catch handler is
skipped (no exception). The code will also be executed if the catch handler exits normally
(user enters a small enough number of rooms). But if an exception is thrown inside the
catch handler, this code will be skipped. In the case of "cancel," there is a finally block.
The code inside the finally block will always be executed, even if the catch handler
throws an exception (user enters an invalid id a second time).
Inner Exceptions
In general it is wise to handle exceptions, at least at some level, near their source, because
you have the most information available about the context in which the exception
occurred. A common pattern is to create a new exception object that captures more
detailed information and throw this onto the calling program. So that information is not
lost about the original exception, you may pass the original exception as a parameter
when constructing the new exception. Then the calling program can gain access to both
exceptions through the InnerException property of the exception object.
The program HotelException\Step3 also illustrates using inner exceptions. In the

MakeReservation method we explicitly check for an IndexOutOfRangeException. We
throw a new exception, which we construct by passing the original exception as a
parameter.
// Add reservation to reservation list and return

// reservation id
try
{
reservations[nextReservation++] = item;
}
catch (IndexOutOfRangeException e)
{
"Reservation table size exceeded", e);
}
In the test program we make use of the InnerException property.
catch (Exception e)
{
if (e.InnerException != null)
{
Console.WriteLine(
"InnerException: {0}", e.InnerException.Message);
}
}
Multiple Catch Handlers
You may have several catch handlers for the same try block. Each catches a different type
of exception. The first catch handler that matches the exception object will be executed.
The program HotelException\Step3 also illustrates using multiple catch handlers. In the
test program we have handlers for both FormatException and Exception. Note that you
do not have to instantiate an exception object instance in the catch statement if you do not
use it. The catch statement can be used without any parameters if you want to catch any
exception and do not care about the exception object.
catch (FormatException)
{
Console.WriteLine(
"Please enter your data in correct format");
}
catch (Exception e)
{
if (e.InnerException != null)
{
Console.WriteLine(
"InnerException: {0}", e.InnerException.Message);
}
}
Here is a sample run of the program. When we use an incorrect format, the first catch
handler is invoked. When we use the correct format, but an illegal negative value for the
number of rooms, we don't get a match for the first catch handler, but we do get a match
for the second, since we are using the base Exception class.

H> book
customer name: bob
number of rooms: xxx
Please enter your data in correct format
H> book
customer name: bob
number of rooms: -1
Exception: Please request a positive number of rooms
H>
Unsafe Code
The mainstream use of C# is to write managed code, which runs on the Common
Language Runtime. As we shall see in Chapter 14, it is quite possible for a C# program to
call unmanaged code, such as a legacy COM component, which runs directly on the
operating system. This facility is important, because a tremendous amount of legacy code
exists, which is all unmanaged.
There is overhead in transitioning from a managed environment to an unmanaged one

and back again. C# provides another facility, called unsafe code, which allows you to
bypass the .NET memory management and get at memory directly, while still running on
the CLR. In particular, in unsafe code you can work with pointers, which we will discuss
later in this section.
Unsafe Blocks
The most circumspect use of unsafe code is within a block, which is specified using the
C# keyword unsafe. The program UnsafeBlock illustrates using the sizeof operator to
determine the size in bytes of various data types. You will get a compiler error if you try
to use the sizeof operator outside of unsafe code.
// UnsafeBlock.cs
using System;
struct Account
{
private int id;
private decimal balance;
}
public class UnsafeBlock

{
{
unsafe
{
Console.WriteLine("size of int = {0}",
sizeof(int));
Console.WriteLine("size of decimal = {0}",
sizeof(decimal));
Console.WriteLine("size of Account = {0}",
sizeof(Account));
}
}
}
To compile this program at the command line, open up a DOS window and navigate to
the directory c:\OI\NetCs\Chap3\UnsafeBlock. You can then enter the following
command to compile using the /unsafe compiler option.
csc /unsafe UnsafeBlock.cs
(You may ignore the warning messages, as our program does not attempt to use fields of
Account. It applies only the sizeof operator.) To run the program, type unsafeblock at the
command line, obtaining the output shown below:
C:\OI\NetCs\Chap3\UnsafeBlock>unsafeblock
size of int = 4
size of decimal = 16
size of Account = 20
To set the unsafe option in Visual Studio, perform the following steps:
1. Right-click over the project in the Solution Explorer and choose Properties.
2. In the Property Pages window that comes up, click on Configuration Properties and
then on Build.
3. In the dropdown for Allow unsafe code blocks choose True. See Figure 3-5.
Figure 3-5. Configuring a project for unsafe mode in Visual Studio.
4. Click OK. You can now compile your project in unsafe mode.
Pointers
Earlier in this chapter we saw that C# has three kinds of data types:
• Value types, which directly contain their data

• Reference types, which refer to data contained somewhere else
• Pointer types
Pointer types can be used only in unsafe code. A pointer is an address of an actual
memory location. A pointer variable is declared using an asterisk after the data type. To
refer to the data a pointer is pointing to, use the dereferencing operator, which is an
asterisk before the variable. To obtain a pointer from a memory location, apply the
address of operator, which is an ampersand in front of the variable. Here are some
examples.
int* p; // p is a pointer to an int

int a = 5; // a is an int, with 5 stored
p = &a; // p now points to a
*p = 12; // 12 is now stored in location pointed
// to by p. So a now has 12 stored
Pointers were widely used in the C programming language, because functions in C pass
data only by value. Thus, if you want a function to return data, you must pass a pointer
rather than the data itself. The program UnsafePointer illustrates a Swap method, which is
used to interchange two integer variables. Since the program is written in C#, we can pass
data by reference. We illustrate with two overloaded versions of the Swap method, one
using ref parameters and the other using pointers. Rather than using an unsafe block, this
program uses unsafe methods, which are defined by including unsafe among the
modifiers of the method. Both the Main method and the one Swap method are unsafe.
// UnsafePointer.cs
using System;
public class UnsafePointer

{
public static unsafe void Main()
{
int x = 55;
int y = 777;
Show("Before swap", x, y);
Swap(ref x, ref y);
Show("After swap", x, y);
Swap(&x, &y);
Show("After unsafe swap", x, y);
}
private static void Show(string s, int x, int y)
{
Console.WriteLine("{0}: x = {1}, y = {2}", s, x, y);
}
private static void Swap(ref int x, ref int y)
{
int temp = x;
x = y;
y = temp;
}
private static unsafe void Swap(int* px, int* py)
{
int temp = *px;
*px = *py;
*py = temp;
}
}
Again you should compile the program using the unsafe option, either at the command
line or in the Visual Studio project. Here is the output. The first swap interchanges the
values. The second swap brings the values back to their original state.
Before swap: x = 55, y = 777

After swap: x = 777, y = 55
After unsafe swap: x = 55, y = 777
Fixed Memory
When working with pointers there is a pitfall. Suppose you have obtained a pointer to a
region of memory that contains data you are working on. Since you have a pointer, you
are accessing memory directly. But suppose the garbage collector collects garbage and
moves data about in memory. Then your object may now reside at a different location,
and your pointer may no longer be valid.
To deal with such a situation, C# provides the keyword fixed, which declares that the
memory in question is "pinned" and cannot be moved by the garbage collector. Note that
you should use fixed only for temporary, local variables, and you should keep the scope
as circumscribed as possible. If too much memory is pinned, the CLR memory-
management system cannot manage memory efficiently.
The program UnsafeAccount illustrates working with fixed memory. This program
declares an array of five Account objects and then assigns them all the same value. The
attempt to determine the size of this array is commented out, because you cannot apply
the sizeof operator to a managed type such as Account[].
It also illustrates the arrow operator for dereferencing a field in a struct, when you have a
pointer to the struct. For example, if p is a pointer to an instance of the struct Account
shown below, the code that follows afterward will assign values to the account object
pointed to by p.
p->id = 101; // assign the id field

p->balance = 50.00m; // assign the balance field
Here is the code.
// UnsafeAccount.cs
using System;
struct Account
{
public int id;
public decimal balance;
public Account(int id, decimal balance)
{
this.id = id;
this.balance = balance;
}
}
public class UnsafeAccount

{
public static unsafe void Main()
{
int id = 101;
decimal balance = 50.55m;
Account acc = new Account(id, balance);
ShowAccount(&acc);
Account[] array = new Account[5];
//Console.WriteLine("size of Account[] = {0}",
// sizeof(Account[]));
ShowArray(array);
fixed (Account* pStart = array)
{
Account* pAcc = pStart;
for (int i = 0; i< array.Length; i++)
*pAcc++ = acc;
}
ShowArray(array);
}
private static unsafe void ShowAccount(Account* pAcc)
{
Console.WriteLine("id = {0}, balance = {1:C}",
pAcc->id, pAcc->balance);
}
private static void ShowAccount(Account acc)
{
Console.WriteLine("id = {0}, balance = {1:C}",
acc.id, acc.balance);
}
private static void ShowArray(Account[] array)
{
for (int i = 0; i < 5; i++)
{
ShowAccount(array[i]);
}
}
}
Summary
In this chapter we have covered the essentials of the C# language, which should equip
you to start writing nontrivial programs in C#. We surveyed variables, operators, control
structures, formatting, methods, and input/output. We examined classes in detail, and we
looked at some convenience features, such as properties. We covered the essentials of
data types in C#, which map to the Common Type System. We discussed the fundamental
distinction between value and reference types, and saw how to convert between them
using boxing and unboxing operations. We examined some standard types, such as string,
StringBuilder, and Array. We covered some additional topics concerning methods,
including parameter passing, variable length parameter lists, method overloading, and
operator overloading. We discussed exception handling in C# in some detail, including
the use of user defined exception classes and structured exception handling. We
concluded the chapter by looking at how you can have "unsafe" sections of C# code,
which can be used to work with pointers for efficiency or for interoperating with legacy
code.
A number of examples pertained to a hotel reservation system. In the next chapter we will
study object-oriented programming in C#, and we will extend our hotel reservation
example to a case study, which will be continued throughout the rest of the book.
Chapter 4. Object-Oriented Programming in C # Top
In this chapter we study in detail the object-oriented aspects of C#, with an emphasis on
inheritance. First we review the fundamentals of object-oriented programming. Next, the
Acme Travel Agency case study is introduced. This case study is developed throughout
the entire book, as we explain more about .NET. We consider some abstractions that will
enable us to implement a reservation system for a variety of resources, and we provide an
implementation of a hotel reservation system. The abstract base classes we define provide
reusable code that enables us to easily implement other kinds of reservation systems. The
key is finding the right abstractions.
We will see how C# language features facilitate object-oriented programming. Certain

details of C#, such as use of access control (public, private, and protected) and properties
can help express abstractions in a way that is safe and easy to use. We will then look at
other object-oriented features of C#, such as virtual methods, method hiding, method
overriding, and polymorphism. A problem in languages supporting inheritance is the
fragile base class problem, and we will see how C# helps in avoiding this pitfall.
This chapter is very much driven by our case study. We introduce object-oriented features
of C# as we elaborate the case study. At the end of the chapter we cover additional
concepts not illustrated by the case study.
Review of Object-Oriented Concepts
In this preliminary section we review the fundamentals of object-oriented programming.

If you are an experienced C++ or Java programmer, you may skim through this section as
a refresher and begin your careful reading with the next section, where we introduce the
case study.
Objects
Objects have both a real-world and a software meaning. The object model describes a
relationship between them.
Objects in the Real World
The term object has an intuitive real-world meaning. There are concrete, tangible objects,
such as a ball, an automobile, and an airplane. There are also more abstract objects that
have a definite intellectual meaning, such as a committee, a patent, or an insurance
contract.
Objects have both attributes (or characteristics) and operations that can be performed
upon them. A ball has a size, a weight, a color, and so on. Operations may be performed
on the ball, such as throw, catch, and drop.
There can be various types of relationships among classes of objects. One, for example, is
a specialization relationship, such as an automobile is a special kind of vehicle. Another
is a whole/part relationship, such as an automobile consists of an engine, a chassis,
wheels, and other parts.
Object Models
Objects can also be used in programs. Objects are useful in programming because you
can set up a software model of a real-world system. Software objects abstract the parts of
objects in the real world that are relevant to the problem being solved. The model can
then be implemented as software using a programming language. A software system
implemented in this way tends to be more faithful to the real system, and it can be
changed more readily when the real system is changed.
There are formal languages for describing object models. The most popular language is
UML (Unified Modeling Language), which is a synthesis of several earlier modeling
languages. Formal modeling languages are beyond the scope of this book, but we will
find that informal models are useful.
Reusable Software Components
Another advantage of objects in software is that they can facilitate reusable software
components. Hardware has long enjoyed significant benefits from reusable hardware
components. For example, computers can be created from power supplies, printed circuit
boards, and other components. Printed circuit boards in turn can be created from chips.
The same chip can be reused in many different computers, and new hardware designs do
not have to be done from scratch.
With appropriate software technology, similar reuse is feasible in software systems.

Objects provide the foundation for software reuse.
Objects in Software
An object is a software entity containing data (state) and related functions (behavior) as a
self-contained module. For example, a HotelBroker may contain a list of hotels (the state)
and provide operations to add a hotel and make a reservation (behavior).
Abstraction
An abstraction captures the essential features of a real-world object, suppressing

unnecessary details. All instances of an abstraction share these common features.
Abstraction helps us deal with complexity. For example, consider the problem of booking
a reservation. There are many different kinds of things you might want to reserve, such as
a hotel, an airplane flight, or a conference room. Such "reservables" have many
differences, but they have certain essentials in common., such as a capacity.
Encapsulation
The implementation of an abstraction should be hidden from the rest of the system, or
encapsulated. For example, the list of hotels may be contained in several different kinds
of data structures, such as an array, a collection, or a database. The rest of the system
should not need to know the details of the representation.
Classes
A class groups all objects with common behavior and common structure. A class allows
creation of new objects of the same type. An object is an instance of some class. We refer
to the process of creating an individual object as instantiation.
Classes can be related in various ways, such as by inheritance and by containment.
Inheritance
Inheritance is a key feature of the object-oriented programming paradigm. You abstract

out common features of your classes and put them in a highlevel base class. You can add
or change features in more specialized derived classes, which "inherit" the standard
behavior from the base class. Inheritance facilitates code reuse and extensibility.
Consider Reservable as a base class, with derived classes Hotel and Flight. All
reservables share some characteristics, such as a capacity. Different kinds of reservables
differ in other respects. For example, a hotel has a city and a name, while a flight has an
origin and a destination. Figure 4-1 illustrates the relationship among these different
kinds of reservables.
Figure 4-1. Inheritance relationship among different reservable classes.

Abstract Classes
Sometimes a class is not meant to be instantiated, but only to provide a template for
derived classes. The Reservable class is an example—it is too abstract to actually
instantiate. Only specific kinds of reservable classes, such as Hotel and Flight, may
actually be instantiated. We call a class such as Reservable that cannot be instantiated an
abstract class. A class that can be instantiated is called a concrete class.
Relationships Among Classes
Classes may be related to each other in various ways.

• The inheritance (IS-A) relationship specifies how one class is a special case of
another class. A Hotel (subclass or derived class) is a special kind of Reservable
(superclass or base class).
• The composition (HAS-A) relationship specifies how one class (the whole) is
made up of other classes (the parts). A HotelBroker (whole) has a list of Hotel
objects.
• A weaker kind of relationship (USES-A) can be identified when one class merely
makes use of some other class when carrying out its responsibilities.
Polymorphism
Consider the problem of generating a payroll for various categories of employees.

Different kinds of employees may have pay calculated in a different manner. A salaried
employee receives a fixed salary. A wage employee is paid according to the number of
hours worked. A sales employee is paid according to the commissions earned on sales
that were made.
A traditional approach is to maintain a type field in an employee structure and to perform

processing in a switch statement, with cases for each type. Such use of switch statements
is error prone and requires much maintenance when adding a new employee type.
An alternative is to localize the intelligence to calculate pay in each employee class,

which will support its own GetPay method. Generic payroll code can then be written that
will handle different types of employees and will not have to be modified to support an
additional employee type. Provide a GetPay method in the base class and an override of
this method in each derived class. Call GetPay through an object reference to a general
Employee object. Depending on the actual employee class referred to, the appropriate
GetPay method will be called.
The ability for the same method call to result in different behavior depending on the
object through which the method is invoked is referred to as polymorphism.
Polymorphism can greatly simplify complex systems and is an important part of the
object-oriented paradigm.
You should not try to coerce your design so that you can take advantage of
polymorphism. We will see in our Acme Travel Agency case study that we have three
different abstract base classes, but we do not need polymorphism to achieve quite general
behavior. On the other hand, the .NET Framework classes use polymorphism heavily, as
we shall see beginning in Chapter 5. Later in this chapter we will provide a small
example of polymorphism using an employee class hierarchy, as outlined above.
Acme Travel Agency Case Study: Design
The Acme Travel Agency provides various services, including the booking of hotel,
plane, and car rental reservations. We will use this simple theme of booking reservations
to illustrate various features of .NET throughout the book. In this chapter we design the
architecture of a general system for booking different kinds of reservations. We illustrate
the reservation system with an implementation of a hotel broker system that supports the
following basic features:
• Add a hotel to the list of hotels

• Show all the hotels
• Show all the hotels in a particular city
• Reserve a hotel room for a range of dates
• Show all the reservations
• Show all the reservations for a particular customer
The system also maintains a list of customers. Customers may register by giving their
name and email address, and they will be assigned a customer ID. The following features
are supported in the basic customer management subsystem:
• Register as a customer
• Change the email address of a customer
• Show a single customer or all the customers
In this chapter various lists, such as hotels, reservations, and customers, will be
maintained as arrays. In the next chapter we will use .NET collections in place of arrays,
and we will implement more features, such as the ability to delete a hotel, cancel a
reservation, and the like. In later chapters we will extend the case study in various ways,
such as providing a graphical user interface, storing all data in a database, deploying as a
Web application, and so on.
The code for our case study is in the CaseStudy folder for this chapter.
Designing the Abstractions
Bearing in mind that eventually we want to implement not only a hotel reservation
system, but also a system for other kinds of reservations, including plane and car rental, it
behooves us at the beginning to look for appropriate abstractions. The more functionality
we are able to put in base classes, the less work we will have to do in order to implement
a particular kind of reservation system. On the other hand, having more functionality in
the base classes can reduce the range of problems to which they are applicable. Good
design is a balancing act.
Another attribute of good abstractions is that they will survive major changes in
implementation. As we shall see later in this book, our C# abstractions of the hotel
reservation system remain intact as we implement the system on an SQL Server database.
These abstractions will be represented in C# by abstract classes, defined in the file

Broker.cs in the CaseStudy folder for this chapter.
Reservable
Our first abstraction is the thing we are looking to reserve. We will denote this abstraction
as simply Reservable. The basic issue in reservations is resource usage. There are a
limited number of reservable resources. Hence the key attribute of a Reservable is
capacity. For example, a hotel may have 100 rooms. A flight may have 250 seats. We will
also want a unique identifier for a Reservable, which we will denote by unitid. (The
shorter name unitid is used in preference to the longer, more awkward name reservableid.
Later we will see other uses of the terminology "unit." For example, the method to add a
reservable is called AddUnit.)
For our applications we are going to introduce an additional attribute, cost. There is a
room rate for a hotel, a ticket cost for a flight, and so on. Note that this attribute may not
be applicable to all things that are being reserved. For example, a conference room within
a company may not have a cost assigned to it. However, our applications are for
commercial customers, so we choose to include cost in our model.
Simplifications
Because our case study is designed to illustrate concepts in C# and .NET, we will
choose many simplifications in our design, so that we do not become bogged
down in too detailed coding. For example, in real life a hotel has several different
kinds of rooms, each having a different rate. Similarly, an airplane flight will
have different classes of seats. Here the situation in real life is even more
complicated, because the price of a seat may vary wildly depending on when the
reservation was made, travel restrictions, and so on. To make life simple for us,
we are assuming that each instance of a particular reservable will have the same
cost.
In C# we will represent a Reservable by an abstract class.
public abstract class Reservable

{
static private int nextid = 0;
protected int unitid;
internal protected int capacity;
internal protected decimal cost;
public Reservable(int capacity, decimal cost)
{
this.cost = cost;
unitid = nextid++;
}
}
A constructor allows us to specify the capacity and cost when the object is created. The
unitid is autogenerated by a static variable. This id starts out at 0, because it is also going
to be used in our implementation as an index in a two-dimensional array to track the
number of customers having a reservation at a given reservable on a given date.
We will discuss the role of the private, internal, and protected access control specifiers
later.
Reservation
When a customer books a reservation of a reservable, a record of the reservation will be

made. The Reservation class holds the information that will be stored.
public abstract class Reservation

{
public int UnitId;
public DateTime Date;
static private int nextReservationId = 1;
public Reservation()
{
ReservationId = nextReservationId++;
}
}
The ReservationId is autogenerated. The UnitId identifies the reservable that was booked.
Date is the starting date of the reservation, and NumberDays specifies the number of days
for which the reservation was made.
Broker
Our third abstraction, Broker, models a broker of any kind of reservable, and is also
represented by an abstract class. It maintains a list of reservables, represented by the array
units, and a list of reservations, represented by the array reservations. The two-
dimensional array numCust keeps track of the number of customers having a reservation
at a given reservable on a given day.
public abstract class Broker

{
private int MaxDay;
private const int MAXRESERVATION = 10;
private static int nextReservation = 0;
private static int nextUnit = 0;
private int[,] numCust;
protected Reservation[] reservations;
protected Reservable[] units;
public Broker(int MaxDay, int MaxUnit)
{
this.MaxDay = MaxDay;
numCust = new int[MaxDay, MaxUnit];
units = new Reservable[MaxUnit];
reservations = new Reservation[MAXRESERVATION];
}
...
ReservationResult
A simple structure is used for returning the result from making a reservation.
public struct ReservationResult

{
public decimal ReservationCost;
public decimal Rate;
public string Comment;
}
The Rate is the cost for one day, and ReservationCost is the total cost, which is equal to
the number of days multiplied by the cost for one day. The ReservationId is returned as -1
if there was a problem, and an explanation of the problem is provided in the Comment
field. This structure is created so that result information can be passed in distributed
scenarios, such as Web Services, where you cannot throw exceptions.
Base Class Logic
The base class Broker not only represents the abstraction of a broker of any kind of
reservable. It also contains general logic for booking reservations and maintaining a list
of reservations. Our ability to capture this logic abstractly gives the power to this base
class and will make implementing reservations in a derived class relatively simple.
Reserve
The core method of the Broker class is Reserve.
protected ReservationResult Reserve(Reservation res)

{
int unitid = res.UnitId;
DateTime dt = res.Date;
int numDays = res.NumberDays;
ReservationResult result = new ReservationResult();
// Check if dates are within supported range
int day = dt.DayOfYear - 1;
if (day + numDays > MaxDay)
{
result.ReservationId = -1;
result.Comment = "Dates out of range";
return result;
}
// Check if rooms are available for all dates
for (int i = day; i < day + numDays; i++)
{
if (numCust[i, unitid] >= units[unitid].capacity)
{
result.Comment = "Room not available";
return result;
}
}
// Reserve a room for requested dates
numCust[i, unitid] += 1;
// Add reservation to reservation list and
// return result
AddReservation(res);
result.ReservationId = res.ReservationId;
result.ReservationCost = units[unitid].cost * numDays;
result.Rate = units[unitid].cost;
result.Comment = "OK";
return result;
}
The Reserve method is designed to implement booking several different kinds of

reservations. Thus the Reservation object, which will be stored in the list of reservations,
is created in a more specialized class derived from Broker and is passed as a parameter to
Reserve. For example, a HotelBroker will book a HotelReservation, and so on. The
UnitId, Date, and NumberDays fields are extracted from the Reservation object, and a
ReservationResult object is created to be returned.
protected ReservationResult Reserve(Reservation res)

{
int unitid = res.UnitId;
DateTime dt = res.Date;
int numDays = res.NumberDays;
ReservationResult result = new ReservationResult();
...
Next we check that all the dates requested for the reservation are within the supported
range (which for simplicity we are taking as a single year). We make use of the DateTime
structure from the System namespace. We return an error if a date lies out of range.
// Check if dates are within supported range

int day = dt.DayOfYear - 1;
if (day + numDays > MaxDay)
{
result.Comment = "Dates out of range";
return result;
}
...
Now we check that space is available for each date, using the numCust array that tracks
how many customers currently have reservations for each day and comparing against the
capacity. The first dimension of this two-dimensional array indexes on days, and the
second dimension indexes on the unit id. (Note that for simplicity we have given our
fields and methods names suitable for our initial application, a HotelBroker.)

{
if (numCust[i, unitid] >= units[unitid].capacity)
{
return result;
}
}
...
Next we actually reserve the unit for the requested days, which is implemented by
incrementing the customer count in numCust for each day.

...
Finally, we add the reservation to the list of reservations and return the result.
// Add reservation to reservation list and

// return result
AddReservation(res);
result.ReservationId = res.ReservationId;
result.ReservationCost =
units[unitid].cost * numDays;
result.Rate = units[unitid].cost;
result.Comment = "OK";
return result;
}
Lists of Reservations and Reservables
The Broker class also maintains lists of reservations and reservables. For our simple array
implementation we only implement Add methods. In a later version we will provide logic
to remove elements from lists.
private void AddReservation(Reservation res)

{
reservations[nextReservation++] = res;
}
protected void AddUnit(Reservable unit)
{
units[nextUnit++] = unit;
}
Designing the Encapsulation
In our current implementation of Broker all lists are represented by arrays. Since this
implementation may not (and in fact will not) be preserved in later versions, we do not
want to expose the arrays themselves or the subscripts that are used for manipulating the
arrays. We provide public properties NumberUnits and NumberReservations to provide
read-only access to the private variables nextUnit and nextReservation.
public int NumberUnits

{
get
{
return nextUnit;
}
}
public int NumberReservations
{
get
{
return nextReservation;
}
}
In our Reservation class the simple fields ReservationId, UnitId, Date, and NumberDays
are not likely to undergo a change in representation, so we do not encapsulate them.
Later, if necessary, we could change some of these to properties, without breaking client
code. For now, and likely forever, we simply use public fields.
public abstract class Reservation

{
public int UnitId;
public DateTime Date;
...
Inheritance in C#
C# supports a single inheritance model. Thus a class may derive from a single base class,
and not from more than one. (In fact, as we saw in the previous chapter, every class in C#
ultimately derives from the root class System.Object. In C# we may use the alias object
for this root class.) This single inheritance model is simple and avoids the complexities
and ambiguities associated with multiple inheritance in C++. Although a C# class can
inherit only from a single base class, it may inherit from several interfaces, a topic we
will discuss in the next chapter.
In this section we discuss inheritance in connection with a further elaboration of our hotel
reservation case study. In the following section we will cover additional features of
inheritance in C#, illustrated by an employee class hierarchy.
Inheritance Fundamentals
With inheritance, you factor the abstractions in your object model, and put the more
reusable abstractions in a high-level base class. You can add or change features in more
specialized derived classes, which "inherit" the standard behavior from the base class.
Inheritance facilitates code reuse and extensibility. A derived class can also provide a
more appropriate interface to existing members of the base class.
Consider Reservable as a base class, with derived classes such as Hotel. All reservables
share some characteristics, such as an id, a capacity, and a cost. Different kinds of
reservables differ in other respects. For example, a hotel has a City and a HotelName.
C# Inheritance Syntax
You implement inheritance in C# by specifying the derived class in the class statement
with a colon followed by the base class. The file HotelBroker.cs in the CaseStudy folder
illustrates deriving a new class Hotel from the class Reservable.
// HotelBroker.cs
namespace OI.NetCs.Acme[1]
{
using System;
public class Hotel : Reservable

{
public string City;
public Hotel(string city, string name,

int number, decimal cost)
: base(number, cost)
{
City = city;
HotelName = name;
}
public int HotelId
{
get
{
return unitid;
}
}
public int NumberRooms
{
get
{
return capacity;
}
}
public decimal Rate
{
get
{
return cost;
}
}
}
[1]
We discuss creating a namespace with the namespace directive later in the chapter.
The class Hotel automatically has all the members of Reservable, and in addition has the
fields City and HotelName.
Changing the Interface to Existing Members
The base class Reservable has members unitid, capacity, and cost, which are designed for
internal use and are not intended to be exposed as such to the outside world. In the Hotel
class we provide public properties HotelId, NumberRooms, and Rate to give clients read-
only access to these fields. When we implement a property in this way, we can choose a
name that is meaningful, such as NumberRooms, in place of a more abstract name, such
as capacity, used in the base class.
Invoking Base Class Constructors
If your derived class has a constructor with parameters, you may wish to pass some of
these parameters along to a base class constructor. In C# you can conveniently invoke a
base class constructor by using a colon, followed by the base keyword and a parameter
list.
public Hotel(string city, string name,

: base(number, cost)
{
City = city;
HotelName = name;
}
Note that the syntax allows you to explicitly invoke a constructor only of an immediate
base class. There is no notation that allows you to directly invoke a constructor higher up
the inheritance hierarchy.
Access Control
C# has two means for controlling accessibility of class members. Access can be
controlled at both the class level and the member level.
Class Accessibility
An access modifier can be placed in front of the class keyword to control who can get at
the class at all. Access can be further restricted by member accessibility, discussed in the
next subsection.
Public
The most common access modifier of a class is public, which makes the class available to
everyone. Whenever we are implementing a class that anyone can use, we want to make
it public.
Internal
The internal modifier makes a class available within the current assembly, which can be
thought of as a logical EXE or DLL. (Assemblies were introduced in Chapter 2 and will
be discussed in more detail in Chapter 7.) All of our projects so far have built a single
assembly, with both the client test program and the class(es) in this assembly. That means
that if we had used internal for the class modifier, the programs would have still worked.
But later, if we put our classes into a DLL and tried to access them from a client program
in a separate EXE, any internal classes would not be accessible. So using public for class
accessibility is generally a good idea.
A common use of the internal modifier is for helper classes that are intended to be used
only within the current assembly, and not generally.
Note that if you omit the access modifier in front of a class, internal will be the default
used by the compiler.
Member Accessibility
Access to individual class members can be controlled by placing an access modifier such
as public or private in front of the member. Member access can only further restrict
access to a class, not widen it. Thus if you have a class with internal accessibility, making
a member public will not make it accessible from outside the assembly.
Public
A public member can be accessed from outside the class.
Private
A private member can be accessed only from within the class (but not from derived
classes).
Protected
Inheritance introduces a third kind of accessibility, protected. A protected member can be

accessed from within the class and from within any derived classes.
Internal
An internal member can be accessed from within classes in the same assembly but not
from classes outside the assembly.
Internal Protected
An internal protected member can be accessed from within the assembly and from
outside the assembly by a derived class.
Access Control in the Case Study
The Reservable class in the file broker.cs illustrates most of the member access-control
options that we have been discussing.
public abstract class Reservable

{
static private int nextid = 0;
protected int unitid;
internal protected int capacity;
internal protected decimal cost;
public Reservable(int capacity, decimal cost)
{
this.cost = cost;
unitid = nextid++;
}
}
The static member nextid is strictly private, because it is used for autogenerating an id
and has no use outside the class. The member unitid is protected because it is used in
derived classes, such as Hotel, but not elsewhere. The members capacity and cost are
used both in derived classes (such as Hotel) and in the class Broker, which is not a
derived class but is in the same assembly. The internal protected access-control
specification is ideal for this case. Note that if we had used just internal, the program
would have still compiled. But since later we may wish to implement derived classes in
other assemblies, internal protected is more appropriate. Finally, the constructor is public.
Acme Travel Agency Case Study: Implementation
With the abstractions Reservable, Reservation, and Broker already in place, it now
becomes very easy to implement a reservation system for a particular kind of reservable,
such as a Hotel. Figure 4-2 illustrates our inheritance hierarchy. Hotel derives from
Reservable, HotelReservation derives from Reservation, and HotelBroker derives from
Broker.
Figure 4-2. Class hierarchy for Acme hotel reservation system.
In this section we will examine key pieces of the implementation of the case study, which
is in the CaseStudy folder for this chapter.
Running the Case Study
Before proceeding with our code walkthrough, it would be a good idea to build and run
the case study. The program TestBroker.exe is a console application. By typing "help" at
the command prompt, you can obtain a list of commands:

H> help
The following commands are available:
hotels shows all hotels in a city
all shows all hotels
cities shows all cities
add adds a hotel
book book a reservation
bookings show all bookings
register register a customer
email change email address
show show customers
quit exit the program
H>
Experiment with this program until you have a clear understanding of its various features.
HotelReservation
HotelReservation is a simple class derived from Reservation. The code is in the file
hotelbroker.cs. It adds some additional public fields and provides the property
ArrivalDate as a more meaningful wrapper around the generic Date field of the base
class.
public class HotelReservation : Reservation

{
public string City;
public DateTime ArrivalDate
{
get
{
return Date;
}
set
{
Date = value;
}
}
}
HotelBroker
The heart of the implementation is the HotelBroker class, derived from Broker. The code
is also in the file hotelbroker.cs.
public class HotelBroker : Broker

{
private const int MAXUNIT = 10;
private const int MAXCITY = 5;
static private int nextCity = 0;
private string[] cities;
public HotelBroker() : base(MAXDAY, MAXUNIT)
{
cities = new String[MAXCITY];
AddHotel("Atlanta", "Dixie", 100, 115.00M);
AddHotel("Atlanta", "Marriott", 500,
70.00M);
AddHotel("Boston", "Sheraton", 250,
95.00M);
}
...
There are constants for various array definitions and a new array to hold the cities. The
constructor passes some array definitions to the base class, initializes the cities array, and
adds some starter hotels as test data.
The next part of the code defines a NumberCity property and provides a method to add a
hotel.
public int NumberCity

{
get
{
return nextCity;
}
}
public string AddHotel(string city, string name,
{
if (FindId(city, name) != -1)
return "Hotel is already on the list";
Hotel hotel = new Hotel(city, name, number, cost);
AddUnit(hotel);
AddCity(city);
return "OK";
}
...
Private helper functions are provided to find the id of a hotel and to add a city to the list
of cities. A city can be added only if it is not already on the list; duplicates are not
permitted.
private int FindId(string city, string name)

{
for (int i = 0; i < NumberUnits; i++)
{
Hotel hotel = (Hotel) units[i];
if ((hotel.City == city)
&& (hotel.HotelName == name))
return hotel.Id;
}
return -1;
}
private void AddCity(string city)
{
// check if city already on list, add if not
if (!Contains(city))
cities[nextCity++] = city;
}
private bool Contains(string city)
{
for (int i = 0; i < NumberCity; i++)
{
if (cities[i] == city)
return true;
}
return false;
}
Methods are provided to show all the hotels, all the hotels in a given city, and to show the
cities. You may wish to examine this code for a review of formatting in C#.
We finally come to the key method Reserve, which is used to book a hotel reservation.
public ReservationResult Reserve(int customerId,

string city, string name, DateTime dt, int numDays)
{
int id = FindId(city, name);
if (id == -1)
{
ReservationResult result =
new ReservationResult();
result.Comment = "Hotel not found";
return result;
}
HotelReservation res = new HotelReservation();
res.UnitId = id;
res.CustomerId = customerId;
res.HotelName = name;
res.City = city;
res.ArrivalDate = dt;
res.DepartureDate = dt + new TimeSpan(numDays, 0, 0, 0);
res.NumberDays = numDays;
return Reserve(res);
}
The code in this class is very simple, because it relies upon logic in the base class Broker.
An error is returned if the hotel cannot be found on the list of hotels. Then a
HotelReservation object is created, which is passed to the Reserve method of the base
class. We create the reservation object in the derived class, because we are interested in
all the fields of the derived HotelReservation class, not just the fields of the base
Reservation class. We have previously used the DateTime structure, and we now use the
TimeSpan structure in calculating the departure date by adding the number of days of the
stay to the arrival date. This calculation relies on the fact that the + operator is overloaded
in the DateTime structure.
Customers
No reservation system can exist without modeling the customers that use it. The
Customers class in the file customer.cs maintains a list of Customer objects. Again we use
an array as our representation. This code has very similar structure to code dealing with
hotels, and so we show it only in outline form, giving the data structures and the
declarations of the public methods and properties.
// Customer.cs
namespace OI.NetCs.Acme
{
using System;

{
static private int nextCustId = 1;
public Customer(string first, string last,
string email)
{
CustomerId = nextCustId++;
FirstName = first;
LastName = last;
}
}
public class Customers

{
private Customer[] customers;
static private int nextCust = 0;
public Customers(int MaxCust)
{
customers = new Customer[MaxCust];
RegisterCustomer("Rocket","Squirrel",
RegisterCustomer("Bullwinkle", "Moose",
}
public int NumberCustomers
...
public int RegisterCustomer(string firstName,
string lastName, string emailAddress)
...
public void ShowCustomers(int customerId)
...
public void ChangeEmailAddress(int id,
string emailAddress)
...
Namespace
All case study code is in the namespace OI.NetCs.Acme. All of the files defining classes
begin with a namespace directive. There is a corresponding using directive, which you
will see in the file TestHotel.cs.
// Customer.cs
namespace OI.NetCs.Acme
{
...
TestHotel
The TestHotel class in the file TestHotel.cs contains an interactive program to exercise
the hotel and customer classes, supporting the commands shown previously where we
suggested running the case study. There is a command loop to read in a command and
then exercise it. There is a big try block around all the commands with a catch handler
afterward. Note the using statement to gain access to the namespace.
// TestHotel.cs
using System;
using OI.NetCs.Acme;

{
{
const int MAXCUST = 10;
HotelBroker hotelBroker = new HotelBroker();
Customers customers = new Customers(MAXCUST);
string cmd;
Console.WriteLine("Enter command, quit to exit");
while (! cmd.Equals("quit"))
{
try
{
if (cmd.Equals("hotels"))
{
string city = iw.getString("city:");
hotelBroker.ShowHotels(city);
}
else if (cmd.Equals("all"))
hotelBroker.ShowHotels();
...
else
hotelhelp();
}
catch (Exception e)
{
Console.WriteLine(
"Exception: {0}", e.Message);
}
}
}
private static void hotelhelp()
{
Console.WriteLine(
"The following commands are available:");
...
}
}
More about Inheritance
Our case study has illustrated many important features of object-oriented programming,
but there is more to the story. Methods in a derived class may hide the corresponding
method in the base class, possibly making use of the base class method in their
implementation. Alternatively, the base class may have virtual methods, which are not
bound to an object at compile time but are bound dynamically at runtime. A derived class
may override a virtual method. This dynamic behavior enables polymorphic code, which
is general code that applies to classes in a hierarchy, and the specific class that determines
the behavior is determined at runtime.
C# provides keywords virtual and override that precisely specify in base and derived
classes, respectively, that the programmer is depending on dynamic binding. By
providing a mechanism to specify polymorphic behavior in the language, C# helps
programs deal with an issue known as the fragile base class problem, which can result in
unexpected behavior in a program when a base class in a library is modified but the
program itself is unchanged.
Employee Class Hierarchy
In this section we will use a much simpler class hierarchy to illustrate the important
concepts. The base class is Employee, which has a public field Name. There are two
derived classes. The SalaryEmployee class has a salary field. The WageEmployee class
has fields for an hourly rate of pay and for the number of hours worked. Figure 4-3
illustrates this simple class hierarchy.
Figure 4-3. Employee class hierarchy.

Method Hiding
A derived class inherits the methods of its base class, and these inherited methods are
automatically available "as is." Sometimes we may want the derived class to do
something a little different for some of the methods of the base class. In this case we will
put code for these changed methods in the derived class, and we say the derived class
"hides" the corresponding methods in the base class. Note that hiding a method requires
that the signatures match exactly. (As we discussed in Chapter 3, methods have the same
signature if they have the same number of parameters, and these parameters have the
same types and modifiers, such as ref or out. The return type does not contribute to
defining the signature of a method.)
In C#, if you declare a method in a derived class that has the same signature as a method
in the base class, you will get a compiler warning message. In such a circumstance, there
are two things you may wish to do. The first is to hide the base class method, which is
what we discuss in this section. The second is to override the base class method, which
we will discuss in the next section.
To hide a base class method, place the keyword new in front of the method in the derived
class. When you hide a method of the base class, you may want to call the base class
method within your implementation of the new method. You can do this by using the
keyword base, followed by a period, followed by the method name and actual parameters.
The example program HideEmployee illustrates method hiding. This program has the
Employee base class and the SalaryEmployee derived class. Each class has a Show
method. The derived class's Show method hides the Show method of the base class. But
the derived class can call the base class Show method through the base keyword. Here is
the code:
// Employee.cs
using System;
public class Employee

{
public string Name;
public Employee(string name)
{
Name = name;
}
public void Show()
{
Console.WriteLine("name = {0}", Name);
}
}
public class SalaryEmployee : Employee

{
private decimal salary;
public SalaryEmployee(string name,
decimal salary)
: base(name)
{
this.salary = salary;
}
new public void Show()
{
base.Show();
Console.WriteLine(
"salary = {0:C}", salary);
}
}
If you delete the new in the derived class Show method, you will get a compiler warning
message:
warning CS0108: The keyword new is required on
'SalaryEmployee.Show()' because it hides
inherited member 'Employee.Show()'
Static Binding
In C# the normal way methods are tied to classes is through static binding. That means
the object reference type is used at compile time to determine the class whose method is
called. The HideEmployee program we just looked at illustrates static binding, using a
simple Employee class and a derived SalaryEmployee class. Here is the test program:
// TestEmployee.cs
using System;
public class TestEmployee

{
{
Employee emp = new Employee("Ellen");
SalaryEmployee sal =
new SalaryEmployee("Sally", 100m);
emp.Show();
sal.Show();
//sal = emp;
emp = sal;
emp.Show();
}
}
In this program emp is an object reference of type Employee. Calling Show through this
object reference will always result in Employee.Show being called, no matter what kind
of object emp may actually be referring to. Here is the output. Notice that the second time
we call Show through emp we are still getting the Employee version of Show (only the
name is displayed).
name = Ellen
name = Sally
salary = $100.00
name = Sally
Press any key to continue
Type Conversions in Inheritance
This program also illustrates another feature of inheritance, type conversions. After the
objects emp and sal have been instantiated, the object references will be referring to
different objects, one of type Employee and the other of type SalaryEmployee. Note that
the SalaryEmployee object has an additional field, salary.
The test program tries two type conversions:
//sal = emp;
emp = sal;
The first assignment is illegal (as you can verify by uncommenting and trying to
compile). Suppose the assignment were allowed. Then you would have an object
reference of type SalaryEmployee referring to an Employee object. If the conversion
"down the hierarchy" (from a base class to a derived class) were allowed, the program
would be open to a bad failure at runtime. What would happen if the code tried to access
a nonexistent member, such as sal accessing the member salary?
The opposite assignment:
emp = sal;
is perfectly legal. We are converting "up the hierarchy." This is okay because of the IS-A
relationship of inheritance. A salary employee "is" an employee. It is a special kind of
employee. Everything that applies to an employee also applies to a salary employee.
There is no "extra field" in the Employee class that is not also present in the
SalaryEmployee class.
Virtual Methods
In C# you can specify that a method in C# will be bound dynamically. Only at runtime
will it be determined whether the base or derived class's method will be called. The
program VirtualEmployee illustrates this behavior. The file VirtualEmployee.cs contains
class definitions for a base class and a derived class, as before. But this time the Show
method is declared as virtual in the base class. In the derived class the Show method is
declared override (in place of new that we used before with method hiding). Now the
Show method in the derived class does not hide the base class method but overrides it.
// VirtualEmployee.cs
using System;

{
public string Name;
{
Name = name;
}
virtual public void Show()
{
Console.WriteLine("name = {0}", Name);
}
}
{
public SalaryEmployee(string name,
decimal salary)
: base(name)
{
}
override public void Show()
{
base.Show();
Console.WriteLine(
"salary = {0:C}", salary);
}
}
We use the same test program. Here is the output. Now, the second time we call Show
through sal, we will be getting the SalaryEmployee.Show method, showing the salary as
well as the name.
name = Ellen
name = Sally
salary = $100.00
name = Sally
salary = $100.00
Virtual Methods and Efficiency
Virtual method invocation (dynamic binding) is slightly less efficient than calling an
ordinary nonvirtual method (static binding). With a virtual method call, there is some
overhead at runtime associated with determining which class's method will be invoked.
C# allows you to specify in a base class whether you want the flexibility of a virtual
method or the slightly greater efficiency of a nonvirtual method. You simply decide
whether or not to use the keyword virtual. (In some languages all methods are virtual, and
you don't have this choice.)
Method Overriding
The override keyword in C# is very useful for making programs clearer. In some
languages, such as C++, there is no special notation for overriding a method in a derived
class. You simply declare a method with the same signature as a method in the base class.
If the base class method is virtual, the behavior is to override. If the base class method is
not virtual, the behavior is to hide. In C# this behavior is made explicit.
The Fragile Base Class Problem
One subtle pitfall in object-oriented programming is the fragile base class problem.
Suppose the override keyword syntax did not exist. Suppose further that you derive a
class from a third-party class library, and you have a method in the derived class that does
not hide or override any method in the base class.
Now a new version of the class library comes out, and the base class has a new virtual
method whose signature happens to match one of the methods in your class. Now you can
be in trouble! Classes that derive from your class may now behave in unexpected ways.
Code that was "expected" to call the new method in the class library—or in code in a
derived class that deliberately overrides this method—may now call your method that has
nothing whatever to do with the method in the class library.
This situation is rare, but if it occurs it can be extremely vicious. Fortunately, C# helps
you avoid such situations by requiring you to use the override keyword if you are indeed
going to perform an override. If you do not specify either override or new and a method
in your derived class has the same signature as a method in a base class, you will get a
compiler error or warning. Thus, if you build against a new version of the class library
that introduces an accidental signature match with one of your methods, you will get
warned by the compiler.
COM and the Fragile Base Class Problem

There is no inheritance in Microsoft's Component Object Model (COM).
Microsoft used the fragile base class problem as a rationale for not providing
inheritance. The issue is much more important for binary components, such as
COM objects, than for traditional class libraries distributed in source code,
because if the problem arises and you have no source for the library, your options
are limited. The real killer is for the problem not to reveal itself in the
development lab, but to crop up only in the field after the application has been
deployed.
Microsoft .NET has similar aims to COM in providing binary components in

multiple languages. The C# override concept uses a corresponding feature of
.NET, so .NET is able to effectively utilize inheritance with less vulnerability
than COM would have had.
Polymorphism
Virtual functions make it easy to write polymorphic code in C#. Our employee example
illustrates the concept of polymorphic code. Imagine a large system with a great many
different kinds of employees. How will you write and maintain code that deals with all
these different employee types?
A traditional approach is to have a "type field" in an employee structure. Then code that
manipulates an employee can key off this type field to determine the correct processing to
perform, perhaps using a switch statement. Although straightforward, this approach can
be quite tedious and error-prone. Introducing a new kind of employee can require
substantial maintenance.
Polymorphism can offer a cleaner solution. You organize the different kinds of employees
in a class hierarchy, and you structure your program so that you write general-purpose
methods that act upon an object reference whose type is that of the base class. Your code
calls virtual methods of the base class. The call will be automatically dispatched to the
appropriate class, depending on what kind of employee is actually being referenced.
You trade off some slight degradation in runtime performance for more reliable code
development.
The program PolyEmployee\Step1 provides an illustration. The GetPay method is virtual,

and methods in the derived class will override it. Here is the code for the base class:
// Employee.cs

{
public string Name;
{
Name = name;
}
virtual public decimal GetPay()
{
return 1.0m;
}
}
Methods in the derived classes override the virtual method in the base class. Here is the
code for SalaryEmployee:
// SalaryEmployee.cs

{
public SalaryEmployee(string name, decimal salary)
: base(name)
{
}
override public decimal GetPay()
{
return salary;
}
}
The WageEmployee class provides its own override of GetPay, where pay is calculated
differently.
// WageEmployee.cs
using System;
public class WageEmployee : Employee

{
private double hours;
public WageEmployee(string name, decimal rate,
double hours)
: base(name)
{
this.rate = rate;
this.hours = hours;
}
override public decimal GetPay()
{
return rate * Convert.ToDecimal(hours);
}
}
The payoff comes in the client program, which can now call GetPay polymorphically.
Here is the code for the test program:
// TestPoly.cs
using System;
public class TestPoly

{
private static Employee[] employees;
private const int MAXEMPLOYEE = 10;
private static int nextEmp = 0;
{
employees = new Employee[MAXEMPLOYEE];
AddSalaryEmployee("Amy", 500.00m);
AddWageEmployee("Bob", 15.00m, 40);
AddSalaryEmployee("Charlie", 900.00m);
PayReport();
}
private static void AddSalaryEmployee(
string name, decimal salary)
{
employees[nextEmp++] =
new SalaryEmployee(name, salary);
}
private static void AddWageEmployee(
string name, decimal rate, double hours)
{
employees[nextEmp++] =
new WageEmployee(name, rate, hours);
}
private static void PayReport()
{
for (int i = 0; i < nextEmp; i++)
{
Employee emp = employees[i];
string name = emp.Name.PadRight(10);
string pay = string.Format("{0:C}",
emp.GetPay());
string str = name + pay;
Console.WriteLine(str);
}
}
}
Here is the output:
Amy $500.00
Bob $600.00
Charlie $900.00
Abstract Classes
Sometimes it does not make sense to instantiate a base class. Instead, the base class is
used to define a standard template to be followed by the various derived classes. Such a
base class is said to be abstract, and it cannot be instantiated. In C# you can designate a
base class as abstract by using the keyword abstract. The compiler will then flag an error
if you try to instantiate the class.
An abstract class may have abstract methods, which are not implemented in the class but
only in derived classes. The purpose of an abstract method is to provide a template for
polymorphism. The method is called through an object reference to the abstract class, but
at runtime the object reference will actually be referring to one of the concrete derived
classes. The keyword abstract is also used to declare abstract methods. In place of curly
brackets and implementation code, you simply provide a semicolon after the declaration
of the abstract method.
An abstract class can be used to provide a cleaner solution of our polymorhphic payroll
example. In the Step 1 solution we discussed previously, there was a virtual function
GetPay in the base class which returned an arbitrary amount of $1.00. We know that this
method is going to be overridden, and in fact the Employee class will itself never be
instantiated. Hence we make Employee an abstract class and GetPay an abstract method.
This solution is illustrated in PolyEmployee\Step2.
// Employee.cs
using System;
abstract public class Employee

{
public string Name;
{
Name = name;
}
abstract public decimal GetPay();
}
Sealed Classes
At the opposite end of the spectrum from abstract classes are sealed classes. While you
must derive from an abstract class, you cannot derive from a sealed class. A sealed class
provides functionality that you can use as is, but you cannot derive from the class and
hide or override some of the methods. An example in the .NET Framework class library
of a sealed class is System.String.
Marking a class as sealed protects against unwarranted class derivations. It can also make
the code a little more efficient, because any virtual functions inherited by the sealed class
are automatically treated by the compiler as nonvirtual.
In C# you use the sealed keyword to mark a class as sealed.
Heterogeneous Collections
A class hierarchy can be used to implement heterogeneous collections that can be treated
polymorphically. For example, you can create an array whose type is that of a base class.
Then you can store within this array object references whose type is the base class, but
which actually may refer to instances of various derived classes in the hierarchy. You may
then iterate through the array and call a virtual method. The appropriate method will be
called for each object in the array.
The program PolyEmployee example illustrates a heterogeneous array of three

employees, which are a mixture of salary and wage employees.
Summary
In this chapter we studied, in detail, the object-oriented aspects of C#, with an emphasis
on inheritance. After a review of the fundamentals of object-oriented programming, we
introduced the Acme Travel Agency case study, which runs as a strand throughout the
entire book. We examined the suitable abstractions that enable us to implement a
reservation system for a variety of resources that must be reserved, and we provided an
implementation of a hotel reservation system. The abstract base classes we defined
provide reusable code that can enable us to easily implement other kinds of reservation
systems. The key is finding the right abstractions.
We saw how C# language features facilitate object-oriented programming. Certain details

of C#, such as use of access control (public, private and protected) and properties can
help express abstractions in a way that is safe and easy to use.
We concluded the chapter by looking at other object-oriented features of C#, such as

virtual methods, method hiding, method overriding, and polymorphism. A pitfall in
languages supporting inheritance is the fragile base class problem, and we have seen how
C# helps in avoiding this pitfall.
Chapter 5. C# in the .NET Framework Top
C# as a language is elegant and powerful. To fully use its capabilities you need to
understand how it works within the .NET Framework. We begin with the root class
object. Collections are examined next, including the methods of the object class that
should be overridden to tap into the functionality provided by the .NET Framework. We
then introduce interfaces, which allow you to rigorously define a contract for a class or
struct to implement. In C# a class can implement multiple interfaces, even though it can
inherit from only one class. Interfaces allow for dynamic programming; you can query a
class at runtime to see whether it supports a particular interface.
The interfaces supporting collections are examined in detail. We investigate issues

involved in copying objects, such as shallow copy and deep copy. Instead of using copy
constructors as in C++, in C# you implement the ICloneable interface. We explore
generic interfaces in the .NET Framework programming model and compare the .NET
and COM component models. A further illustration of generic interfaces is provided by
sorting in different orders with the IComparable interface. The examples offer insight into
the workings of frameworks, which are more than class libraries. In a framework, you
call the framework, and the framework calls you. Your code can be viewed as the middle
layer of a sandwich. This key insight can help you grasp what makes .NET programming
"tick." While callback functions have been used for years in programming, C# uses this
concept in delegates and events. Two simple and intuitive examples are presented: a stock
market simulation and an online chat room. The chapter concludes with a discussion of
attributes, which are pervasive in the .NET Framework.
System.Object
As we have already seen, every type in C#, whether it is a value type or a reference type,
ultimately inherits from the root class System.Object. C# provides object as a keyword
alias for this root class. The class ValueType inherits directly from object. ValueType is
the root for all value types, such as structures and simple types like int and decimal.
Public Instance Methods of object
There are four public instance methods of object, three of which are virtual and
frequently overridden by classes.
Equals
public virtual bool Equals(object obj);
This method compares an object with the object passed as a parameter and returns true if
they are equal. object implements this method to test for reference equality. ValueType
overrides the method to test for content equality. Many classes override the method to
make equality behave appropriately for the particular class.
ToString
public virtual string ToString();
This method returns a human-readable string representation of the object. The default
implementation returns the type name. Derived classes frequently override this method to
return a meaningful string representation of the particular object.
GetHashCode
public virtual int GetHashCode();
This method returns a hash value for an object, suitable for use in hashing algorithms and
hash tables. You should normally override this method if you override ToString. (The C#
compiler will give you a warning message if you override one and not the other.)
GetType
public Type GetType();
This method returns type information for the object. This type information can be used to
get the associated metadata through reflection, a topic we discuss in Chapter 8.
Protected Instance Methods
There are two protected instance methods, which can be used only within derived classes.
MemberWiseClone
protected object MemberwiseClone();
This method creates a shallow copy of the object. To perform a deep copy, you should
implement the ICloneable interface We will discuss shallow and deep copy later in this
chapter.
Finalize
~Object();
This method allows an object to free resources and perform other cleanup operations
before it is reclaimed by garbage collection. In C# the Finalize method is represented by
"destructor" notation like that used in C++. But note that the semantics are totally
different. In C++, destructors are invoked in a deterministic manner, which the
programmer can depend upon. In C#, finalization is nondeterministic, dependent upon the
garbage collector. We discuss finalization in Chapter 8.
Generic Interfaces and Standard Behavior
If you are used to a language like Smalltalk, the set of behaviors specified in object may
seem quite limited. Smalltalk, which introduced the concept of a class hierarchy rooted in
a common base class, has a very rich set of methods defined in its Object class. I counted
38 methods! [1] These additional methods support features such as comparing objects and
copying objects. The .NET Framework class library has similar methods, and many more.
But rather than putting them all in a common root class, .NET defines a number of
standard interfaces, which classes can optionally support. This kind of organization,
which is also present in Microsoft's Component Object Model (COM) and in Java, is very
flexible. We will study interfaces later in this chapter, and we will discuss some of the
generic interfaces of the .NET Framework.
[1]
The methods of Smalltalk's Object class are described in Chapters 6 and 14 of
Smalltalk-80: The Language and its Implementation, by Adele Goldberg and David
Robson.
Using object Methods in the Customer Class
As a simple illustration of object methods, let's look at our Customer class before and
after overriding the Equals, ToString, and GetHashCode methods.
Default Methods of Object
If our class does not provide any overrides of the virtual instance methods of object, our
class will inherit the standard behavior. This behavior is demonstrated in
CustomerObject\Step1.
// Customer.cs

{
public Customer(int id, string first, string last,
string email)
{
CustomerId = id;
FirstName = first;
LastName = last;
}
}
// TestCustomer.cs
using System;

{
{
cust1 = new Customer(99, "John", "Doe",
cust2 = new Customer(99, "John", "Doe",
ShowCustomerObject("cust1", cust1);
ShowCustomerObject("cust2", cust2);
CompareCustomerObjects(cust1, cust2);
}
private static void ShowCustomerObject(string label,
Customer cust)
{
Console.WriteLine("---- {0} ----", label);
Console.WriteLine("ToString() = {0}",
cust.ToString());
Console.WriteLine("GetHashCode() = {0}",
cust.GetHashCode());
Console.WriteLine("GetType() = {0}",
cust.GetType());
}
private static void CompareCustomerObjects(
Customer cust1, Customer cust2)
{
Console.WriteLine("Equals() = {0}",
cust1.Equals(cust2));
}
}
Run the test program and you will see this output:
---- cust1 ----

ToString() = Customer
GetHashCode() = 4
GetType() = Customer
---- cust2 ----
ToString() = Customer
GetHashCode() = 6
Equals() = False
The default implementation is not at all what we want for our Customer object. ToString
returns the name of the class, not information about a particular customer. Equals checks
for reference equality. In our example, we have two different references to Customer
objects with the same content, and Equals return false.
Overriding Methods of Object
The version of the project in CustomerObject\Step2 demonstrates overriding these virtual

methods. Our override of Equals tests for content equality.
// Customer.cs

{
public Customer(int id, string first, string last,
string email)
{
CustomerId = id;
FirstName = first;
LastName = last;
}
public override bool Equals(object obj)
{
Customer cust = (Customer) obj;
return (cust.CustomerId == CustomerId);
}
public override int GetHashCode()
{
return CustomerId;
}
public override string ToString()
{
return FirstName + " " + LastName ;
}
}
The test program is identical. Here is the new output:
---- cust1 ----

ToString() = John Doe
GetHashCode() = 99
---- cust2 ----
ToString() = John Doe
GetHashCode() = 99
Equals() = True
Collections
The .NET Framework class library provides an extensive set of classes for working with
collections of objects. These classes are all in the System.Collections namespace and
implement a number of different kinds of collections, including lists, queues, stacks,
arrays, and hashtables. The collections contain object instances. Since all types derive
ultimately from object, any built-in or user-defined type may be stored in a collection.
In this section we will look at a representative class in this namespace, ArrayList, and see
how to use array lists in our programs.
ArrayList Example
To get our bearings, let's begin with a simple example of using the ArrayList class. An
array list, as the name suggests, is a list of items stored like an array. An array list can be
dynamically sized and will grow as necessary to accommodate new elements being
added.
Collection classes are made up of instances of type object. We will create and manipulate
a collection of Customer objects. We could just as easily create a collection of any other
built-in or user-defined type. If our type were a value type, such as int, the instance would
be boxed before being stored in the collection. When the object is extracted from the
collection, it will be unboxed back to int.
Our example program is CustomerCollection. It initializes a list of customers and then

lets the user show the customers, register a new customer, unregister a customer, and
change an email address. A simple "help" method displays the commands that are
available:

H> help
The following commands are available:
register register a customer
unregister unregister a customer
email change email address
show show customers
Before examining the code it would be a good idea to run the program to register a new
customer, show the customers, change an email address, unregister a customer, and show
the customers again. Here is a sample run of the program:
H> show
id (-1 for all): -1
1 Rocket Squirrel [email protected]
2 Bullwinkle Moose [email protected]
H> register
first name: Bob
last name: Oberg
email address: [email protected]
id = 3
H> email
customer id: 1
email address: [email protected]
H> unregister
id: 2
H> show
id (-1 for all): -1
1 Rocket Squirrel
[email protected]
3 Bob Oberg
[email protected]
Customers Class
All the code for this project is in the folder CustomerCollection. The file customer.cs has
code for the Customer and Customers classes. The code for Customer is almost identical
to what we looked at previously. The only addition is a special constructor that
instantiates a Customer object with a specified id. We use this constructor in the
Customers class when we remove an element and when we check if an element is present
in the collection.

{
...
public Customer(int id)
{
CustomerId = id;
FirstName = "";
LastName = "";
EmailAddress = "";
}
...
}
The Customers class contains a list of customers, represented by an ArrayList.
public class Customers

{
private ArrayList customers;
public Customers()
{
customers = new ArrayList();
RegisterCustomer("Rocket", "Squirrel",
RegisterCustomer("Bullwinkle", "Moose",
}
{
Customer cust = new Customer(firstName, lastName,
emailAddress);
customers.Add(cust);
return cust.CustomerId;
}
public void UnregisterCustomer(int id)
{
Customer cust = new Customer(id);
customers.Remove(cust);
}
{
foreach (Customer cust in customers)
{
if (cust.CustomerId == id)
{
cust.EmailAddress = emailAddress;
return;
}
}
throw new Exception("id " + id + " not found");
}
public void ShowCustomers(int id)
{
if (!CheckId(id) && id != -1)
return;
foreach (Customer cust in customers)
{
if (id == -1 || id == cust.CustomerId)
{
string sid =
cust.CustomerId.ToString().PadLeft(4);
string first = cust.FirstName.PadRight(12);
string last = cust.LastName.PadRight(12);
string email = cust.EmailAddress.PadRight(20);
string str = sid + " " + first + " " +
last + " " + email;
}
}
}
private bool CheckId(int id)
{
Customer cust = new Customer(id);
return customers.Contains(cust);
}
}
The lines in the listing in bold show the places where we are using collection class
features. In Chapter 3 we have already used foreach with arrays. The reason foreach can
be used with arrays is that the Array class, like ArrayList, implements the IEnumerable
interface that supports foreach syntax. We will discuss IEnumerable and the other
collection interfaces later in this chapter.
The Add and Remove methods, as their names suggest, are used for adding and removing
elements from a collection. The Remove method searches for an object in the collection
that Equals the object passed as a parameter. Our special constructor creates an object
having the id of the element we want to remove. Since we provided an override of the
Equals method that bases equality on CustomerId, the proper element will be removed.
Similarly, the Contains method used in our CheckId helper method also relies on the
override of the Equals method.
Compare the code in this program with the use of arrays in the code in the previous
chapter's case study. The collection code is much simpler. Using collections makes it easy
to remove elements as well as add them. Using arrays, you would have to write special
code to move array elements to fill in the space where an element was deleted. Also,
collections are not declared to have a specific size, but can grow as required.
Interfaces
Interface is a very fundamental concept in computer programming. A large system is

inevitably decomposed into parts, and it is critical to precisely specify the interfaces
between these parts. Interfaces should be quite stable, as changing an interface affects
multiple parts of the system. In C# interface is a keyword and has a very precise meaning.
An interface is a reference type, similar to an abstract class, that specifies behavior as a
set of methods, properties, indexers, and events. [2] An interface is a contract. When a
class or struct implements an interface, it must adhere to the contract.
[2]
We discuss events later in this chapter.
Interfaces are a useful way to partition functionality. You should first specify interfaces
and then design appropriate classes to implement the interfaces. While a class in C# can
inherit from only one other class, it can implement multiple interfaces.
Interfaces facilitate dynamic programs—you can query a class at runtime to see whether
it supports a particular interface, and take action accordingly. Interfaces in C# and .NET
are conceptually very similar to interfaces in Microsoft's Component Object Model, but
as we will see, they are much easier to work with.
In this section we will study the fundamentals of interfaces and provide illustrations using
some small sample programs. Then we will restructure our Acme case study to take
advantage of interfaces and explore their use in detail. After that we will examine several
important generic interfaces in the .NET library, which will help us gain an understanding
of how C# and the .NET library support each other to help us develop powerful and
useful programs.
Interface Fundamentals
Object-oriented programming is a useful paradigm for helping to design and implement

large systems. Using classes helps us to achieve abstraction and encapsulation. Classes
are a natural decomposition of a large system into manageable parts. Inheritance adds
another tool for structuring our system, enabling us to factor out common parts into base
classes, helping us to accomplish greater code reuse.
The main purpose of an interface is to specify a contract independently of

implementation. It is important to understand that conceptually the interfaces come first.
Interfaces in C#
In C# interface is a keyword, and you define an interface in a manner similar to defining

a class. Like classes, interfaces are reference types. The big difference is that there is no
implementation code in an interface; it is pure specification. Also note that an interface
can have properties as well as methods (it could also have other members, such as
indexers). As a naming convention, interface names usually begin with a capital I.
The IAccount interface specifies operations to be performed on a bank account.
interface IAccount
{
void Deposit(decimal amount);
void Withdraw(decimal amount);
decimal Balance {get;}
void Show();
}
This interface illustrates the syntax for declaring the read-only Balance property—you
specify the data type, the property name, and in curly brackets which of set and get apply
(only get in this case, because the property is read-only).
Implementing an Interface
In C# you specify that a class or struct implements an interface by using the colon
notation that is employed for class inheritance. A class can also inherit both from a class
and from an interface. In this case the base class should appear first in the derivation list
after the colon.
public class AccountC : Account, IAccount

{
public void Show()[3]
{
Console.WriteLine("balance = {0}", Balance);
}
}
[3]
Note that we do not need the override keyword when our class implements the Show
method of the IAccount interface. Unlike overriding a virtual method in a class, we are
implementing a method which was only specified but not implemented in the interface
definition.
In our example the class AccountC inherits from the class Account, and it implements the
interface IAccount. The methods of the interface must all be implemented by Account,
either directly or in one of the base classes in its inheritance hierarchy.
We will examine a full-blown example of interfaces with the reservation-broker

inheritance hierarchy later in the chapter, when we implement Step 2 of the case study.
As a small example, consider the program InterfaceDemo. The interface IAccount is

defined, and two different classes, AccountC and AccountW, implement the interface.
These implementations differ only in the Show method. The AccountC implementation
performs console output to display the account balance, and AccountW uses a Windows
message box. [4] The Deposit and Withdraw methods and the Balance property are all
implemented in the Account base class.
[4]
We will discuss Windows programming in Chapter 6. The example program has all
needed references to libraries, and all you need to do to display a message box is to call
the Show method of the MessageBox class.
// Account.cs
using System;
using System.Windows.Forms;
interface IAccount
{
void Show();
}
public class Account
{
public Account()
{
balance = 100;
}
public void Deposit(decimal amount)
{
balance += amount;
}
public void Withdraw(decimal amount)
{
balance -= amount;
}
public decimal Balance
{
get
{
return balance;
}
}
}
public class AccountC : Account, IAccount

{
public void Show()
{
}
}
public class AccountW : Account, IAccount

{
public void Show()
{
MessageBox.Show("balance = " + Balance);
}
}
Using an Interface
You may call methods of an interface through an object reference to the class, or you may
obtain an interface reference and call the methods through this interface reference. [5] The
test program in the file InterfaceDemo.cs demonstrates both. We obtain the interface
reference iacc by an implicit cast when we do the assignment to the object reference acc
or accw. Note the polymorphic behavior of the call to Show, using console or Windows
output depending on which object is being used.
[5]
As we will see later in the chapter when we discuss "explicit interface
implementation," you can force a client program to use an interface reference and not a
class reference.
// InterfaceDemo.cs
using System;
class InterfaceDemo
{
{
// Use an object reference
AccountC acc = new AccountC();
acc.Deposit(25);
acc.Show();
// Use an interface reference
IAccount iacc = acc;
iacc.Withdraw(50);
iacc.Show();
// Use interface reference for another class
// that implements IAccount
AccountW accw = new AccountW();
iacc = accw;
iacc.Show();
}
}
Multiple Interfaces
Our first example illustrated two classes providing different implementations of the same
interface. Another common scenario is for a class to implement multiple interfaces, and
in C# it is easy to test at runtime which interfaces are implemented by a class.
Our example program is MultipleInterfaces, which also illustrates interface inheritance.

The interfaces IBasicAccount, IDisplay, and IAccount are defined in the file
AccountDefs.cs.
// AccountDefs.cs
interface IBasicAccount
{
}
interface IDisplay
{
void Show();
}
interface IAccount : IBasicAccount, IDisplay

{
}
Interface Inheritance
Interfaces can inherit from other interfaces. Unlike classes in C#, for which there is only
single inheritance, there can be multiple inheritance of interfaces. In our example, the
interface IAccount is declared by inheriting from the two smaller interfaces,
IBasicAccount and IDisplay. The advantage of factoring the original interface into two
smaller interfaces is an increase in flexibility. For example, a class implementing
IBasicAccount may run on a server, where it would not be appropriate to implement
IDisplay.
When declaring a new interface using interface inheritance, you can also introduce
additional methods, as illustrated for IAccount2.
interface IAccount2 : IBasicAccount, IDisplay

{
void NewMethod();
}
Implementing Multiple Interfaces
A class implements multiple interfaces by mentioning each interface in its inheritance list
and by providing code for the methods of each interface. A method may be implemented
through inheritance from a base class. The file Account.cs in the MultipleInterfaces
project illustrates two classes. BasicAccount implements only the interface
IBasicAccount, and Account implements the two interfaces, IBasicAccount and IDisplay.
// Account.cs
using System;
public class BasicAccount : IBasicAccount

{
public BasicAccount()
{
balance = 100;
}
{
balance += amount;
}

{
balance -= amount;
}
{
get
{
return balance;
}
}
}
public class Account : BasicAccount, IBasicAccount,

IDisplay
{
public void Show()
{
}
}
Using Multiple Interfaces
The test program MultipleInterfaces.cs illustrates using (or trying to use) the two
interfaces with an Account object and a BasicAccount object. Both interfaces can be used
with Account, but we cannot use the IDisplay interface with BasicAccount. If we
attempted to do an implicit cast from BasicAccount to IDisplay, the compiler would flag
an error message. In our code we perform an explicit cast within a try block. The code
compiles, but we get a runtime InvalidCast exception, which we catch. The program also
illustrates that we can sometimes take a reasonable, alternative course of action if the
desired interface is not available. In our case, we are able to perform the output ourselves,
making use of the Balance property of the IBasicAccount interface.
// MultipleInterfaces.cs
using System;
class MultipleInterfaces
{
{
IBasicAccount iacc;
IDisplay idisp;
// Use an Account object, which has full functionality
Account acc = new Account();
iacc = acc;
idisp = acc;
iacc.Deposit(25);
idisp.Show();
// Use BasicAccount object, with reduced functionality
BasicAccount bacc = new BasicAccount();
iacc = bacc;
iacc.Withdraw(50);
try
{
idisp = (IDisplay) bacc;
idisp.Show();
}
catch (InvalidCastException e)
{
Console.WriteLine("IDisplay is not supported");
Console.WriteLine(e.Message);
// Display the balance another way
Console.WriteLine("balance = {0}", iacc.Balance);
}
}
}
Here is the output from running the program:
balance = 125
IDisplay is not supported
Exception of type System.InvalidCastException was thrown.
balance = 50
Dynamic Use of Interfaces
A powerful feature of interfaces is their use in dynamic scenarios, allowing us to write

general code that can test whether an interface is supported by a class. If the interface is
supported, our code can take advantage of it; otherwise our program can ignore the
interface. We could in fact implement such dynamic behavior through exception
handling, as illustrated previously. Although entirely feasible, this approach is very
cumbersome and would lead to programs that are hard to read. C# provides two
operators, as and is, that facilitate working with interfaces at runtime.
As an example, consider the program DynamicInterfaces, which uses the interface

definitions and class implementations from our previous example. The test program
illustrates using each of the C# as and is operators to check whether the IDisplay
interface is supported.
// DynamicInterfaces.cs
using System;
class DynamicInterfaces
{
{
IBasicAccount iacc;
IDisplay idisp;
BasicAccount bacc = new BasicAccount();
iacc = bacc;
iacc.Withdraw(50);
// Check IDisplay via C# "as" operator
idisp = bacc as IDisplay;
if (idisp != null)
idisp.Show();
else
{
}
// Check IDisplay via C# "is" operator
if (bacc is IDisplay)
{
idisp.Show();
}
else
{
}
}
}
Here is the output from running the test program:

balance = 50
balance = 50
As Operator [6]
[6]
The C# as operator is similar to dynamic_cast in C++.
The as operator is used to convert one reference type to another reference type. A
common application is to convert an object reference or an interface reference to another
interface reference. Unlike performing the conversion by a cast operation, the as operator
never throws an exception. If the conversion fails, the result value is null.
idisp = bacc as IDisplay;

if (idisp != null)
// idisp is a valid interface reference
The as operator can also be used to explicitly convert a value type to a reference type by a
boxing operation. Again, null is returned if the conversion fails.
Is Operator [7]
[7]
The C# is operator is similar to type_id in C++.
The is operator dynamically checks if the runtime type of an object is compatible with a
given type. The result is a boolean value. The is operator can be used to check if an object
refers to a class supporting a given interface, as illustrated in our DynamicInterfaces
program.
if (bacc is IDisplay)
{
idisp.Show();
}
The is operator is not the most efficient solution, as a check of the type is made twice.
The first time is when the is operator is invoked. But the check is made all over again
when the cast operation is performed, because the runtime will throw an exception if the
interface is not supported. For this situation, as is more efficient, since you obtain the
interface reference directly.
The is operator is useful if you want to check whether an interface is supported but you
don't need to directly call a method of the interface. Later in the chapter we will see an
example of this situation, when we discuss the IComparable interface. If the elements of a
collection support IComparable, you will be able to call a Sort method on the collection.
The Sort method calls the CompareTo method of IComparable, although your own code
does not.
Interfaces in C# and COM

There are many similarities between .NET and COM. In both, the concept of
interface plays a fundamental role. Interfaces are useful for specifying contracts.
Interfaces support a very dynamic style of programming.
In COM you must yourself provide a very elaborate infrastructure in order to

implement a COM component. You must implement a class factory for the
creation of COM objects. You must implement the QueryInterface method of
IUnknown for the dynamic checking of interfaces. You must implement AddRef
and Release for proper memory management.
With C# (and other .NET languages) the Common Language Runtime does all
this for you automatically. You create an object via new. You check for an
interface via is or as and obtain the interface by a cast. The garbage collector
takes care of memory management for you.
Explicit Interface Implementation
When working with interfaces, an ambiguity can arise if a class implements two
interfaces and each has a method with the same name and signature. As an example,
consider the following versions of the interfaces IAccount and IStatement. Each interface
contains the method Show.
interface IAccount
{
void Show();
}
interface IStatement
{
int Transactions {get;}
void Show();
}
How can the class specify implementations of these methods? The answer is to use the
interface name to qualify the method, as illustrated in the program Ambiguous. The
IAccount version IAccount.Show will display only the balance, and IStatement.Show
will display both the number of transactions and the balance.
// Account.cs (project "Ambiguous")
...
public class Account : IAccount, IStatement
{
int numXact = 0;
public Account(decimal balance)
{
this.balance = balance;
}
{
balance += amount;
++numXact;
}
{
balance -= amount;
++numXact;
}
{
get
{
return balance;
}
}
void IAccount.Show()
{
Console.WriteLine("balance = {0}", balance);
}
public int Transactions
{
get
{
return numXact;
}
}
void IStatement.Show()
{
Console.WriteLine("{0} transactions, balance = {1}",
numXact, balance);
}
}
You will notice that in the definition of the class Account, the qualified methods
IAccount.Show and IStatement.Show do not have an access modifier such as public.
Such qualified methods cannot be accessed through a reference to a class instance. They
can only be accessed through an interface reference of the type explicitly shown in the
method definition. The test program shows that we cannot call the IAccount.Show
method through an Account object reference but only through an IAccount interface
reference.
By obtaining an IStatement interface reference, we can call IStatement.Show.
// Ambiguous.cs
using System;
public class Ambiguous

{
{
Account acc = new Account(100);
// acc.Show(); // illegal - MUST go through an
// interface
IAccount iacc = (IAccount) acc;
IStatement istat = (IStatement) acc;
iacc.Show();
istat.Show();
iacc.Deposit(25);
iacc.Withdraw(10);
iacc.Show();
istat.Show();
}
}
Even when there is no ambiguity, you may wish to use explicit interface implementation,
in order to force client programs to use interfaces to call the methods specified in the
interfaces. This approach makes it very clear that the client code is programming against
specific interfaces and not against a large amorphous collection of methods of a class.
The code will be easily adaptable to using different classes that implement the same
interfaces.
Acme Travel Agency Case Study: Step 2
We will now apply our knowledge of interfaces to a little restructuring of the Acme case
study. A major benefit of using interfaces is that they raise the level of abstraction
somewhat, helping you to understand the system by way of the interface contacts,
without worrying about how the system is implemented.
As usual, our case study code is in the CaseStudy directory for this chapter.
The Contracts
There are two main sets of contracts in the Acme Travel Agency Case Study. The first
specifies operations on customers, and the second, operations involving hotels.
Customer Contract
The ICustomer interface shown below specifies the methods to be used by clients in the
Acme Travel Agency system.
public interface ICustomer

{
int RegisterCustomer(string firstName, string lastName,
string emailAddress);
void UnregisterCustomer(int id);
ArrayList GetCustomer(int id);
void ChangeEmailAddress(int id, string emailAddress);
}
The RegisterCustomer, UnregisterCustomer, and ChangeEmailAddress method

definitions are exactly the same as the methods we implemented in the Customers class.
The GetCustomer method is new. Previously, we had a ShowCustomers method, which
displayed a list of customers to the console. This method was strictly temporary. For
general use we want to return data and let the client decide what to do with it. The
GetCustomer method returns information about one or all customers in an array list. If -1
is passed for the id, the list will contain all the registered customers. Otherwise, the list
will contain the customer information for the customer with the given id. If no customer
has that id, the list will be empty.
Hotel Contracts
We next look at the functionality of the class HotelBroker. The methods divide fairly
naturally into three groups.
• Hotel information, such as the cities where hotels are available and the hotels
within a city
• Hotel administration, such as adding or deleting a hotel, or changing the number
of rooms and rate of a hotel
• Hotel reservations, such as booking or canceling a reservation or obtaining a list
of reservations
Accordingly we create three interfaces for the HotelBroker. These interfaces are defined
in AcmeDefinitions.cs.
public interface IHotelInfo

{
ArrayList GetCities();
ArrayList GetHotels();
ArrayList GetHotels(string city);
}
public interface IHotelAdmin

{
string AddHotel(string city, string name,
int numberRooms, decimal rate);
string DeleteHotel(string city, string name);
string ChangeRooms(string city, string name,
int numberRooms, decimal rate);
}
public interface IHotelReservation

{
ReservationResult MakeReservation(int customerId,
string city, string hotel, DateTime checkinDate,
int numberDays);
void CancelReservation(int id);
ArrayList FindReservationsForCustomer(int customerId);
}
The Implementation
We examined the Step 1 implementation of the hotel brokerage system in detail in

Chapter 4. The Step 2 implementation uses collections in place of arrays, and it passes
information to the client rather than displays information directly.
Structures
One detail of our implementation concerns the data structures used to pass lists to the
client. We use the ArrayList class. But what do we store in each array list? We could use
Customer objects and Hotel objects. The problem here is that these classes have
implementation-specific data in them, as well as the information fields that the client
program cares about. To obtain implementation neutral representations, we introduce
several structures.
In Customers.cs we define the CustomerListItem structure for passing customer
information.

{
}
In AcmeDefinitions.cs we define structures for hotels, reservations, and reservation

results.
public struct HotelListItem

{
public string City;
public int NumberRooms;
}

{
public string City;
public DateTime ArrivalDate;
}
public struct ReservationResult

{
public decimal ReservationCost;
public string Comment;
}
The ReservationResult returns a ReservationId of -1 if there is a problem, giving an

explanation of the problem in the Comment field. Otherwise "OK" is returned in the
Comment field.
We invite you to examine the code in the CaseStudy folder and to build and run the
program.
Generic Interfaces in .NET
The .NET Framework exposes much standard functionality through generic interfaces,
which are implemented in various combinations by classes in the Framework itself, and
which can also be implemented by your own classes in order to tap into standard
functionality defined by the Framework. In this section we will look at several categories
of operations that are supported by these standard, generic interfaces,
• Collections
• Copying objects
• Comparing objects
Our survey of generic interfaces is by no means exhaustive, but our sampling should give
you a good understanding of how generic interfaces work in the .NET Framework.
Collection Interfaces
Now that we understand the concept of interfaces, we are equipped to take a closer look
at collections, and in particular the ArrayList class that we have used so heavily in the
case study. If we look at the definition of ArrayList, we see that it implements four
standard interfaces.
public class ArrayList : IEnumerable, ICollection,

IList, ICloneable
The first three interfaces form a simple interface hierarchy, as shown in Figure 5-1. As
you go down the hierarchy, additional methods are added, until IList specifies a fully
featured list.
Figure 5-1. Interface hierarchy for lists.

The fourth interface, ICloneable, is independent and is used to support deep copying. As
a simple illustration of the collection interfaces we provide the program StringList. Here
is the Main method. We'll look at the individual helper methods as we examine the
various collection interfaces.
// StringList.cs
using System;
using System.Collections;
public class StringList

{
private static ArrayList list;
{
// Initialize strings and show starting state
list = new ArrayList(4);
ShowCount();
AddString("Amy");
AddString("Bob");
AddString("Charlie");
ShowEnum(list); // enumerator
ShowCount();
// Add two more strings and show state again
AddString("David");
AddString("Ellen");
ShowList(list); // foreach
ShowCount();
// Remove two strings from list and show state
RemoveString("David");
RemoveAt(0);
ShowArray(list); // index notation
ShowCount();
// Try to remove two strings not in list
RemoveString("Amy");
RemoveAt(3);
}
...
Here is the output:
list.Count = 0
list.Capacity = 4
Amy
Bob
Charlie
list.Count = 3
list.Capacity = 4
array[0] = Amy
array[1] = Bob
array[2] = Charlie
array[3] = David
array[4] = Ellen
list.Count = 5
list.Capacity = 8
Bob
Charlie
Ellen
list.Count = 3
list.Capacity = 8
List does not contain Amy
No element at index 3
Interface Documentation
Interfaces are documented in the online .NET Framework SDK Documentation. Figure 5-
2 illustrates the documentation of the IEnumerable interface. The right-hand pane has a
language filter button which we have used to show only C# versions. If you are using
the interface in one of the .NET Framework classes that implement the interface, you do
not need to implement any of the methods yourself. If you are creating your own class
that supports an interface, you must provide implementations of all the methods of the
interface. In either case, the documentation describes the methods for you.
Figure 5-2. NET Framework SDK documentation for IEnumerable interface.

IEnumerable And IEnumerator
The basic interface that must be supported by collection classes is IEnumerable, which
has a single method, GetEnumerator.
interface IEnumerable
{
IEnumerator GetEnumerator();
}
GetEnumerator returns an interface reference to IEnumerator, which is the interface used

for iterating through a collection. This interface has the property Current and the methods
MoveNext and Reset.
interface IEnumerator
{
object Current {get;}
bool MoveNext();
void Reset();
}
The enumerator is initially positioned before the first element in the collection, and it
must be advanced before it is used. The ShowEnum method (in the StringList example)
illustrates using an enumerator to iterate through a list.
private static void ShowEnum(ArrayList array)
{
IEnumerator iter = array.GetEnumerator();
bool more = iter.MoveNext();
while (more)
{
string str = (string) iter.Current;
more = iter.MoveNext();
}
}
This pattern of using an enumerator to iterate through a list is so common that C#

provides a special kind of loop, foreach, that can be used for iterating through the
elements of any collection. Here is the comparable code using foreach.
private static void ShowList(ArrayList array)

{
foreach (string str in array)
{
}
}
ICollection
The ICollection interface is derived from IEnumerable and adds a Count property and a
CopyTo method.
interface ICollection : IEnumerable

{
int Count {get;}
bool IsSynchronized {get;}
object SyncRoot {get;}
void CopyTo(Array array, int index);
}
There are also synchronization properties that can help you deal with thread safety issues.
"Is it thread safe?" is a question frequently asked about library code. The short answer to
this question for the .NET Framework class library is "No." This does not mean that the
designers of the Framework did not think about thread safety issues. On the contrary,
there are many mechanisms to help you write thread-safe code when you need to. The
reason that collections are not automatically thread safe is that your code should not have
to pay the performance penalty to enforce synchronization when it is not running in a
multithreading scenario. If you do need thread safety, you may use the thread-safety
properties to easily implement. We discuss the .NET mechanisms for thread
synchronization in Chapter 8.
Our StringList program illustrates use of the Count property of ICollection.
private static void ShowCount()

{
Console.WriteLine("list.Count = {0}", list.Count);
Console.WriteLine("list.Capacity = {0}",
list.Capacity);
}
IList
The IList interface is derived from ICollection and provides methods for adding an item
to a list, removing an item, and so on. An indexer is provided that enables array notation
to be used. (We discussed indexers in Chapter 3.)
interface IList : ICollection

{
object this[int index] {get; set;}
int Add(object value);
void Clear();
bool Contains(object value);
int IndexOf(object value);
void Insert(int index, object value);
void Remove(object value);
void RemoveAt(int index);
}
Our StringList sample code illustrates using the indexer and the Add, Contains, Remove,
and RemoveAt methods.
private static void ShowArray(ArrayList array)

{
for (int i = 0; i < array.Count; i++)
{
Console.WriteLine("array[{0}] = {1}", i, array[i]);
}
}
private static void AddString(string str)
{
if (list.Contains(str))
throw new Exception("list contains " + str);
list.Add(str);
}
private static void RemoveString(string str)
{
if (list.Contains(str))
list.Remove(str);
else
Console.WriteLine("List does not contain {0}", str);
}
private static void RemoveAt(int index)
{
try
{
list.RemoveAt(index);
}
catch (ArgumentOutOfRangeException)
{
Console.WriteLine("No element at index {0}", index);
}
}
Copy Semantics and ICloneable
Sometimes you have to make a copy of an object. When you copy objects that contain
objects and object references, you have to be aware of. the copy semantics of C#. We will
compare reference copy, shallow memberwise copy, and deep copy. We will see that by
implementing the ICloneable interface in your class, you can make a deep copy.
Recall that C# has value types and reference types. A value type contains all its own data,
while a reference type refers to data stored somewhere else. If a reference variable gets
copied to another reference variable, both will refer to the same object. If the object
referenced by the second variable is changed, the first variable will also reflect the new
value. Sometimes you want this behavior, but sometimes you do not.
Shallow Copy and Deep Copy
A struct in C# automatically implements a "memberwise" copy, sometimes known as a

"shallow copy." The object root class has a protected method, MemberwiseClone, which
will perform a memberwise copy of members of a class.
If one or more members of a class are of a reference type, this memberwise copy may not
be good enough. The result will be two references to the same data, not two independent
copies of the data. To actually copy the data itself and not merely the references, you will
need to perform a "deep copy." Deep copy can be provided at either the language level or
the library level. In C++ deep copy is provided at the language level through a copy
constructor. In C# deep copy is provided by the .NET Framework through a special
interface, ICloneable, which you can implement in your classes in order to enable them to
perform deep copy.
Example Program
We will illustrate all these ideas in the program CopyDemo. This program makes a copy
of a Course instance. The Course class consists of a title and a collection of students.
// Course.cs
using System;
public class Course : ICloneable
{
public string Title;
public ArrayList Roster;
public Course(string title)
{
Title = title;
Roster = new ArrayList();
}
public void AddStudent(string name)
{
Roster.Add(name);
}
public void Show(string caption)
{
Console.WriteLine("-----{0}-----", caption);
Console.WriteLine("Course : {0} with {1} students",
Title, Roster.Count);
foreach (string name in Roster)
{
Console.WriteLine(name);
}
}
public Course ShallowCopy()
{
return (Course) this.MemberwiseClone();
}
public object Clone()
{
Course course = new Course(Title);
course.Roster = (ArrayList) Roster.Clone();
return course;
}
}
The test program constructs a Course instance c1 and then makes a copy c2 by various
methods.
Reference Copy by Assignment
The first way the copy is performed is by the straight assignment c2 = c1. Now we get
two references to the same object, and if we make any change through the first reference,
we will see the same change through the second reference. The first part of the test
program illustrates such an assignment.
// CopyDemo.cs
using System;
public class CopyDemo

{
private static Course c1, c2;
{
Console.WriteLine("Copy is done via c2 = c1");
InitializeCourse();
c1.Show("original");
c2 = c1;
c2.Title = ".NET Programming";
c2.AddStudent("Charlie");
c2.Show("copy with changed title and new student");
...
}
private static void InitializeCourse()
{
c1 = new Course("Intro to C#");
c1.AddStudent("John");
c1.AddStudent("Mary");
}
}
We initialize with the title "Intro to C#" and two students. We make the assignment c2 =
c1 and then change the title and add another student for c2. We then show both c1 and c2,
and we see that both reflect both of these changes. Here is the output from this first part
of the program:
Copy is done via c2 = c1

-----original-----
Course : Intro to C# with 2 students
John
Mary
-----copy-----
John
Mary
-----copy with changed title and new student-----
Course : .NET Programming with 3 students
John
Mary
Charlie
-----original-----
John
Mary
Charlie
Memberwise Clone
Next we will illustrate doing a memberwise copy, which can be accomplished using the
MemberwiseClone method of object. Since this method is protected, we cannot call it
directly from outside our Course class. Instead, in Course we define a method,
ShallowCopy, which is implemented using MemberwiseClone.
// Course.cs
using System;

{
...
public Course ShallowCopy()
{
return (Course) this.MemberwiseClone();
}
...
}
Here is the second part of the test program, which calls the ShallowCopy method. Again
we change the title and a student in the second copy.
// CopyDemo.cs
using System;

{
...
Console.WriteLine(
"\nCopy is done via c2 = c1.ShallowCopy()");
InitializeCourse();
c2 = c1.ShallowCopy();
...
Here is the output of this second part of the program. Now the Title field has its own
independent copy, but the Roster collection is just copied by reference, so each copy
refers to the same collection of students.
Copy is done via c2 = c1.ShallowCopy()

John
Mary
Charlie
-----original-----
John
Mary
Charlie
Using ICloneable
The final version of copy relies on the fact that our Course class supports the ICloneable
interface and implements the Clone method. To clone the Roster collection we use the
fact that ArrayList also implements the ICloneable interface, as discussed earlier in the
chapter. Note that the Clone method returns an object, so we must cast to ArrayList
before assigning to the Roster field.
// Course.cs
using System;

{
...
public object Clone()
{
Course course = new Course(Title);
course.Roster = (ArrayList) Roster.Clone();
return course;
}
}
Here is the third part of the test program, which calls the Clone method. Again we change
the title and a student in the second copy.
// CopyDemo.cs
using System;

{
...
Console.WriteLine(
"\nCopy is done via c2 = c1.Clone()");
InitializeCourse();
c2 = (Course) c1.Clone();
...
Here is the output from the third part of the program. Now we have completely
independent instances of Course. Each has its own title and set of students.
Copy is done via c2 = c1.Clone()

John
Mary
Charlie
-----original-----
John
Mary
Comparing Objects
We have quite exhaustively studied copying objects. We now examine comparing objects.
To compare objects, the .NET Framework uses the interface IComparable. In this section
we use the interface IComparable to sort an array.
Sorting an Array
The System.Array class provides a static method, Sort, that can be used for sorting an
array. The program ArrayName illustrates applying this Sort method to an array of Name
objects, where the Name class simply encapsulates a string through a read-only property
Text. Here is the main program.
// ArrayName.cs
...
public class ArrayName

{
{
Name[] array = new Name[10];
array[0] = new Name("Michael");
array[1] = new Name("Charlie");
array[2] = new Name("Peter");
array[3] = new Name("Dana");
array[4] = new Name("Bob");
if (array[0] is IComparable)
Array.Sort(array);
else
Console.WriteLine(
"Name does not implement IComparable");
foreach (Name name in array)
{
if (name != null)
}
}
}
Implementing IComparable
In order for the Sort method to function, there must be a way of comparing the objects
that are being sorted. This comparison is achieved through the CompareTo method of the
interface IComparable. Thus to sort an array of a type you define, you must implement
IComparable for your type.
public interface IComparable

{
int CompareTo(object object);
}
Here is the implementation of the Name class, with its implementation of IComparable.
public class Name : IComparable

{
private string text;
public Name(string text)
{
this.text = text;
}
public string Text
{
get
{
return text;
}
}
public int CompareTo(object obj)
{
string s1 = this.Text;
string s2 = ((Name) obj).Text;
return String.Compare(s1, s2);
}
}
Understanding Frameworks
Our example offers some insight into the workings of frameworks. A framework is more
than a library. In a typical library, you are concerned with your code calling library
functions. In a framework, you call into the framework and the framework might call
you. Your program can be viewed as the middle layer of a sandwich.
• Your code calls the bottom layer.

• The top layer calls your code.
The .NET Framework is an excellent example of such an architecture. There is rich
functionality that you can call directly. There are many interfaces, which you can
optionally implement to make your program behave appropriately when called by the
framework, often on behalf of other objects.
Delegates
Interfaces facilitate writing code so that your program can be called into by some other
code. This style of programming has been available for a long time, under the guise of
"callback" functions. In this section we examine del-egates in C#, which can be thought
of as type-safe and object-oriented callback functions. Delegates are the foundation for a
design pattern, known as events, which we'll look at in the next section.
A callback function is one which your program specifies and "registers" in some way, and
which then gets called by another program. In C and C++ callback functions are
implemented by function pointers.
In C# you can encapsulate a reference to a method inside a delegate object. A delegate

can refer to either a static method or an instance method. When a delegate refers to an
instance method, it stores both an object instance and an entry point to the instance
method. The instance method can then be called through this object instance. When a
delegate object refers to a static method, it stores just the entry point of this static method.
You can pass this delegate object to other code, which can then call your method. The
code that calls your delegate method does not have to know at compile time which
method is being called.
In C# a delegate is considered a reference type that is similar to a class type. A new

delegate instance is created just like any other class instance, using the new operator. In
fact, C# delegates are implemented by the .NET Framework class library as a class,
derived ultimately from System.Delegate.
Delegates are object oriented and type safe, and they enjoy the safety of the managed
code execution environment.
Declaring a Delegate
You declare a delegate in C# using a special notation with the keyword delegate and the
signature of the encapsulated method. A naming convention suggests that your name
should end with "Callback."
We illustrate delegates in the sample program DelegateAccount. Here is an example of a

delegate declaration from the file DelegateAccount.cs.
public delegate void NotifyCallback(decimal balance);

Defining a Method
When you instantiate a delegate, you will need to specify a method, which must match
the signature in the delegate declaration. The method may be either a static method or an
instance method. Here are some examples of methods that can be hooked to the
NotifyCallback delegate:
private static void NotifyCustomer(decimal balance)

{
Console.WriteLine("Dear customer,");
Console.WriteLine(
" Account overdrawn, balance = {0}", balance);
}
private static void NotifyBank(decimal balance)
{
Console.WriteLine("Dear bank,");
Console.WriteLine(
}
private void NotifyInstance(decimal balance)
{
Console.WriteLine("Dear instance,");
Console.WriteLine(
}
Creating a Delegate Object
You instantiate a delegate object with the new operator, just as you would with any other
class. The following code illustrates creating two delegate objects. The first one is hooked
to a static method, and the second to an instance method. The second delegate object
internally will store both a method entry point and an object instance that is used for
invoking the method.
NotifyCallback custDlg =
new NotifyCallback(NotifyCustomer);
...
DelegateAccount da = new DelegateAccount();
NotifyCallback instDlg =
new NotifyCallback(da.NotifyInstance);
Calling a Delegate
You "call" a delegate just as you would a method. The delegate object is not a method,
but it has an encapsulated method. The delegate object "delegates" the call to this
encapsulated method, hence the name "delegate." In the following code the delegate
object notifyDlg is called whenever a negative balance occurs on a withdrawal. In this
example the notifyDlg delegate object is initialized in the method SetDelegate.
private NotifyCallback notifyDlg;
...
public void SetDelegate(NotifyCallback dlg)
{
notifyDlg = dlg;
}
...
{
balance -= amount;
if (balance < 0)
notifyDlg(balance);
}
Combining Delegate Objects
A powerful feature of delegates is that you can combine them. Delegates are "multicast,"
in which they have an invocation list of methods. When such a delegate is called, all the
methods on the invocation list will be called in the order they appear in the invocation
list. The + operator can be used to combine the invocation methods of two delegate
objects. The - operator can be used to remove methods.
NotifyCallback bankDlg = new NotifyCallback(NotifyBank);
NotifyCallback currDlg = custDlg + bankDlg;
In this example we construct two delegate objects, each with an associated method. We
then create a new delegate object whose invocation list will consist of both the methods
NotifyCustomer and NotifyBank. When currDlg is called, these two methods will be
invoked. Later on in the code we may remove a method.
currDlg -= bankDlg;
Now NotifyBank has been removed from the delegate, and the next time currDlg is
called, only NotifyCustomer will be invoked.
Complete Example
The program DelegateAccount illustrates using delegates in our bank account scenario.
The file DelegateAccount.cs declares the delegate NotifyCallback. The class
DelegateAccount contains methods matching the signature of the delegate. The Main
method instantiates delegate objects and combines them in various ways. The delegate
objects are passed to the Account class, which uses its encapsulated delegate object to
invoke suitable notifications when the account is overdrawn.
Observe how this structure is dynamic and loosely coupled. The Account class does not
know or care which notification methods will be invoked in the case of an overdraft. It
simply calls the delegate, which in turn calls all the methods on its invocation list. These
methods can be adjusted at runtime.
Here is the code for the Account class:
// Account.cs

{
private NotifyCallback notifyDlg;
public Account(decimal bal, NotifyCallback dlg)
{
balance = bal;
notifyDlg = dlg;
}
public void SetDelegate(NotifyCallback dlg)
{
notifyDlg = dlg;
}
{
balance += amount;
}
{
balance -= amount;
if (balance < 0)
notifyDlg(balance);
}
{
get
{
return balance;
}
}
}
Here is the code declaring and testing the delegate:
// DelegateAccount.cs
using System;
public delegate void NotifyCallback(decimal balance);
public class DelegateAccount

{
{
NotifyCallback bankDlg =
new NotifyCallback(NotifyBank);
NotifyCallback currDlg = custDlg + bankDlg;
Account acc = new Account(100, currDlg);
Console.WriteLine("balance = {0}", acc.Balance);
acc.Withdraw(125);
acc.Deposit(200);
acc.Withdraw(125);
currDlg -= bankDlg;
acc.SetDelegate(currDlg);
acc.Withdraw(125);
DelegateAccount da = new DelegateAccount();
NotifyCallback instDlg =
new NotifyCallback(da.NotifyInstance);
currDlg += instDlg;
acc.SetDelegate(currDlg);
acc.Withdraw(125);
}
private static void NotifyCustomer(decimal balance)
{
Console.WriteLine("Dear customer,");
Console.WriteLine(
}
private static void NotifyBank(decimal balance)
{
Console.WriteLine("Dear bank,");
Console.WriteLine(
}
private void NotifyInstance(decimal balance)
{
Console.WriteLine("Dear instance,");
Console.WriteLine(
}
}
Here is the output from running the program. Notice which notification methods get
invoked, depending upon the operations that have been performed on the current delegate
object.
balance = 100
Dear customer,
Account overdrawn, balance = -25
Dear bank,
balance = -25
balance = 50
Dear customer,
Dear customer,
Dear instance,
Stock Market Simulation
As a further illustration of the use of delegates, consider the simple stock-market

simulation, implemented in the directory StockMarket. The simulation consists of two
modules:
• The Admin module provides a user interface for configuring and running the
simulation. It also implements operations called by the simulation engine.
• The Engine module is the simulation engine. It maintains an internal clock and
invokes randomly generated operations, based on the configuration parameters
passed to it.
Figure 5-3 shows the high-level architecture of the simulation. The following
operations are available:
Figure 5-3. Architecture of stock-market simulation.
• PrintTick: shows each clock tick.

• PrintTrade: shows each trade.
The following configuration parameters can be specified:
• Ticks on/off
• Trades on/off
• Count of how many ticks to run the simulation
Running the Simulation
Build and run the example program in StockMarket. Start with the default configuration:
Ticks are OFF, Trades are ON, Run count is 100. (Note that the results are random and
will be different each time you run the program.)
Ticks are OFF
Trades are ON
Run count = 100
: run
2 ACME 23 600
27 MSFT 63 400
27 IBM 114 600
38 MSFT 69 400
53 MSFT 75 900
62 INTC 27 800
64 MSFT 82 200
68 MSFT 90 300
81 MSFT 81 600
83 INTC 30 800
91 MSFT 73 700
99 IBM 119 400
:
The available commands are listed when you type "help" at the colon prompt. The
commands are:
count set run count

ticks toggle ticks
trades toggle trades
config show configuration
run run the simulation
The output shows clock tick, stock, price, volume.
Delegate Code
Two delegates are declared in the Admin.cs file.
public delegate void TickCallback(int ticks);

public delegate void TradeCallback(int ticks, string stock,
int price, int volume);
As we saw in the previous section, a delegate is similar to a class, and a delegate object is
instantiated by new.
TickCallback tickDlg = new TickCallback(PrintTick);

TradeCallback tradeDlg = new TradeCallback(PrintTrade);
A method is passed as the parameter to the delegate constructor. The method signature
must match that of the delegate.
public static void PrintTick(int ticks)
{
Console.Write("{0} ", ticks);
if (++printcount == LINECOUNT)
{
printcount = 0;
}
}
Passing the Delegates to the Engine
The Admin class passes the delegates to the Engine class in the constructor of the Engine
class.
Engine engine = new Engine(tickDlg, tradeDlg);
Random-Number Generation
The heart of the simulation is the Run method of the Engine class. At the core of the Run
method is assigning simulated data based on random numbers. We use the
System.Random class, which we discussed in Chapter 3.
double r = rangen.NextDouble();
if (r < tradeProb[i])
{
int delta = (int) (price[i] * volatility[i]);
if (rangen.NextDouble() < .5)
{
delta = -delta;
}
price[i] += delta;
int volume = rangen.Next(minVolume, maxVolume) * 100;
tradeOp(tick, stocks[i], price[i], volume);
}
Using the Delegates
In the Engine class, delegate references are declared:
TickCallback tickOp;
TradeCallback tradeOp;
The delegate references are initialized in the Engine constructor:
public Engine(TickCallback tickOp, TradeCallback tradeOp)

{
this.tickOp = tickOp;
this.tradeOp = tradeOp;
}
The method that is wrapped by the delegate object can then be called through the
delegate reference:
if (showTicks)
tickOp(tick);
Events
Delegates are the foundation for a design pattern known as events. Conceptually, servers
implement incoming interfaces, which are called by clients. In a diagram, such an
interface may be shown with a small bubble (a notation used in COM). Sometimes a
client may wish to receive notifications from a server when certain "events" occur. In
such a case the server will specify an outgoing interface. The server defines the interface
and the client implements it. In a diagram, such an interface may be shown with an arrow
(again, a notation used in COM). Figure 5-4 illustrates a server with one incoming and
one outgoing interface. In the case of the outgoing interface, the client will implement an
incoming interface, which the server will call.
Figure 5-4. A server with an incoming interface and an outgoing interface.
A good example of a programming situation with events is a graphical user interface. An

event is some external action, typically triggered by the user, to which the program must
respond. Events include user actions such as clicking a mouse button or pressing a key on
the keyboard. A GUI program must contain event handlers to respond to or "handle"
these events. We will see many examples of GUI event handling in Chapter 6, where we
discuss Windows Forms.
Events in C# and .NET
The .NET Framework provides an easy-to-use implementation of the event paradigm

built on delegates. C# simplifies working with .NET events by providing the keyword
event and operators to hook up event handlers to events and to remove them. The
Framework also defines a base class EventArgs to be used for passing arguments to event
handlers. There are a number of derived classes defined by the Framework for specific
types of events, such as MouseEventArgs, ListChangedEventArgs, and so forth. These
derived classes define data members to hold appropriate argument information.
An event handler is a delegate with a specific signature,
public delegate void EventHandler(

object sender,
EventArgs e);
The first argument represents the source of the event, and the second argument contains
data associated with the event.
We will examine this event architecture through salient code from the example program
EventDemo, which illustrates a chat room.
Server-Side Event Code
We begin with server-side code, in ChatServer.cs. The .NET event architecture uses
delegates of a specific signature:
public delegate void JoinHandler(object sender,

ChatEventArg e);
The first parameter specifies the object that sent the event notification. The second
parameter is used to pass data along with the notification. Typically, you will derive a
class from EventArg to hold your specific data.
public class ChatEventArg : EventArgs

{
public string Name;
public ChatEventArg(string name)
{
Name = name;
}
}
A delegate object reference is declared using the keyword event.

public event JoinHandler Join;
A helper method is typically provided to facilitate calling the delegate object(s) that have
been hooked up to the event.
protected void OnJoin(ChatEventArg e)

{
if (Join != null)
{
Join(this, e);
}
}
A test for null is made in case no delegate objects have been hooked up to the event.
Typically, access is specified as protected, so that a derived class has access to this helper
method. You can then "fire" the event by calling the helper method.
public void JoinChat(string name)

{
members.Add(name);
OnJoin(new ChatEventArg(name));
}
Client-Side Event Code
The client provides event handler functions.
public static void OnJoinChat(object sender,

ChatEventArg e)
{
Console.WriteLine(
"sender = {0}, {1} has joined the chat",
sender, e.Name);
}
The client hooks the handler to the event, using the += operator.
ChatServer chat = new ChatServer("OI Chat Room");

// Register to receive event notifications from the server
chat.Join += new JoinHandler(OnJoinChat);
The event starts out as null, and event handlers get added through +=. All of the
registered handlers will get invoked when the event delegate is called. You may
unregister a handler through -=.
Chat Room Example
The chat room example in EventDemo illustrates the complete architecture on both the
server and client sides. The server provides the following methods:
• JoinChat
• QuitChat
• ShowMembers
Whenever a new member joins or quits, the server sends a notification to the client. The
event handlers print out an appropriate message. Here is the output from running the
program:
sender = OI Chat Room, Michael has joined the chat

sender = OI Chat Room, Bob has joined the chat
sender = OI Chat Room, Sam has joined the chat
--- After 3 have joined---
Michael
Bob
Sam
sender = OI Chat Room, Bob has quit the chat
--- After 1 has quit---
Michael
Sam
Client Code
The client program provides event handlers. It instantiates a server object and then hooks
up its event handlers to the events. The client then calls methods on the server. These
calls will trigger the server, firing events back to the client, which get handled by the
event handlers.
// ChatClient.cs
using System;
class ChatClient
{
public static void OnJoinChat(object sender,
ChatEventArg e)
{
Console.WriteLine(
"sender = {0}, {1} has joined the chat",
sender, e.Name);
}
public static void OnQuitChat(object sender,
ChatEventArg e)
{
Console.WriteLine(
"sender = {0}, {1} has quit the chat",
sender, e.Name);
}
{
ChatServer chat = new ChatServer("OI Chat Room");
// Register to receive event notifications from the
// server
chat.Join += new JoinHandler(OnJoinChat);
chat.Quit += new QuitHandler(OnQuitChat);
// Call methods on the server
chat.JoinChat("Michael");
chat.JoinChat("Bob");
chat.JoinChat("Sam");
chat.ShowMembers("After 3 have joined");
chat.QuitChat("Bob");
chat.ShowMembers("After 1 has quit");
}
}
Server Code
The server provides code to store in a collection the names of people who have joined the
chat. When a person quits the chat, the name is removed from the collection. Joining and
quitting the chat triggers firing an event back to the client. The server also contains the
"plumbing" code for setting up the events, including declaration of the delegates, the
events, and the event arguments. There are also helper methods for firing the events.
// ChatServer.cs
using System;
public class ChatEventArg : EventArgs

{
public string Name;
public ChatEventArg(string name)
{
Name = name;
}
}
public delegate void JoinHandler(object sender,

ChatEventArg e);
public delegate void QuitHandler(object sender,
ChatEventArg e);
public class ChatServer

{
private ArrayList members = new ArrayList();
private string chatName;
public event JoinHandler Join;
public event QuitHandler Quit;
public ChatServer(string chatName)
{
this.chatName = chatName;
}
override public string ToString()
{
return chatName;
}
protected void OnJoin(ChatEventArg e)
{
if (Join != null)
{
Join(this, e);
}
}
protected void OnQuit(ChatEventArg e)
{
if (Quit != null)
{
Quit(this, e);
}
}
public void JoinChat(string name)
{
members.Add(name);
OnJoin(new ChatEventArg(name));
}
public void QuitChat(string name)
{
members.Remove(name);
OnQuit(new ChatEventArg(name));
}
public void ShowMembers(string msg)
{
Console.WriteLine("--- " + msg + "---");
foreach (string member in members)
{
Console.WriteLine(member);
}
}
}
It may appear that there is a fair amount of such "plumbing" code, but it is much simpler
than the previous connection-point mechanism used by COM for events. Also, in certain
areas various wizards and other tools (such as the Forms designers) will generate the
infrastructure for you automatically. We will see how easy it is to work with events in
Windows programming in Chapter 6.
Attributes
A modern approach to implementing complex code is to let the system do it for you.
There must be a way for the programmer to inform the system of what is desired. In the
.NET Framework such cues can be given to the system by means of attributes.
Microsoft introduced attribute-based programming in Microsoft Transaction Server. The

concept was that MTS, not the programmer, would implement complex tasks such as
distributed transactions. The programmer would "declare" the transaction requirements
for a COM class, and MTS would implement it. This use of attributes was greatly
extended in the next generation of MTS, known as COM+. In MTS and COM+ attributes
are stored in a separate repository, distinct from the program itself.
Attributes are also used in Interface Definition Language (IDL), which gives a precise
specification of COM interfaces, including the methods and signatures. Part of the
function of IDL is to make it possible for a tool to generate proxies and stubs for
remoting a method call across a process boundary or even across a network. When
parameters are passed remotely, it is necessary to give more information than when they
are passed within the same process. For example, within a process, you can simply pass a
reference to an array. But in passing an array across a process boundary, you must inform
the tool of the size of the array. This information is communicated in IDL by means of
attributes, which are specified using a square-bracket notation. Here is an example of IDL
that shows the use of attributes.
[
object,
uuid(AAA19CDE-C091-47BF-8C96-C80A00989796),
dual,
pointer_default(unique)
]
interface IAccount : IDispatch
{
[id(1)] HRESULT Deposit([in] long id, [in] long
amount);
[id(2)] HRESULT Withdraw([in] long id, [in] long
amount);
[id(3)] HRESULT GetBalance([in] long id,
[out] long *pBal);
[id(4)] HRESULT GetAllBalances([in, out] long* pCount,
[out, size_is(*pCount)] long balances[]);
};
If you are experienced with COM, such IDL will be familiar to you. If not, just notice the
general structure of how attributes are used. Attributes such as object and uuid are applied
to the interface, the id attribute is applied to methods, and the attributes in, out, and
size_is are applied to parameters.
A problem with attributes in both MTS/COM+ and IDL is that they are separate from the
program source code. When the source code is modified, the attribute information may
get out of sync with the code.
Attributes in .NET
In .NET, attributes are declared with square brackets, as in IDL. But unlike IDL, the
attributes are part of the program source code. When compiled into intermediate
language, the attributes become part of the metadata. There are some predefined
attributes in C#, there are many attributes associated with various .NET classes, and there
is a mechanism to create custom attributes for your own classes. In this section we look at
the general characteristics of how attributes are used, beginning with a simple example of
using one of the predefined attributes in C#. In later chapters attributes associated with
specific .NET classes will be used extensively, and in Chapter 8, after we've discussed
Reflection, we will see how to create and use custom attributes.
The AttributeDemo program provides a simple example of using the predefined C#

attribute Conditional, which is used to mark a method to be executed only if a
preprocessor symbol is defined.
// AttributeDemo.cs
#define LINUX
using System;
using System.Diagnostics;
public class AttributeDemo

{
{
Notice();
MultiNotice();
Console.WriteLine("Goodbye");
}
[ConditionalAttribute("UNIX")]
private static void Notice()
{
Console.WriteLine("Notice: Unix version");
}
[Conditional("UNIX")] [Conditional("LINUX")]
private static void MultiNotice()
{
Console.WriteLine("Notice: Some version of Unix");
}
}
Conditional is one of three predefined attributes in C#. [8] Its full name is
ConditionalAttribute, but C# has the convenience feature that when an attribute's name
ends with the Attribute suffix, you may drop the suffix. Conditional is used to mark a
method with a symbol. If that symbol is defined by the preprocessor, calls to the method
will be included, otherwise calls will be omitted. The Conditional attribute is multiuse,
which means that it may be used several times in front of a method. For example, in the
code above the MultiNotice method is conditioned on either "UNIX" or "LINUX," and
calls to this method will be included if either symbol is defined. The preprocessor #define
directive [9] defines cthe symbol "LINUX." The "UNIX" symbol is not defined (unless
done via a compiler option, which we'll look at shortly). The Conditional attribute
requires the namespace System.Diagnostics. (We will discuss .NET diagnostic support in
detail in Chapter 13.)
[8]
The other two predefined attributes in C# are Obsolete and AttributeUsage. Obsolete is
used to mark a program entity that should not be used, causing the compiler to issue a
warning or error message if it is used. We will discuss AttributeUsage in Chapter 8 in
connection with custom attributes.
[9]
C#, unlike C and C++, does not allow use of preprocessor directives to define macros.
Running the program produces the following output:
Notice: Some version of Unix

Goodbye
The call to Notice is omitted, but the call to MultiNotice is included. You may experiment
with this program by defining no symbols, defining "UNIX," etc.
Preprocessor Symbols Using Compiler Option
Besides using a #define preprocessor directive in your source code, you can also define
preprocessor symbols using the /define command-line option of the C# compiler. For
example, you can define the symbol "UNIX" using the following command:
csc /define:UNIX AttributeDemo.cs
You can also specify preprocessor directives in Visual Studio. In Solution Explorer right-
click on the solution. From the context menu choose Properties. Select Build from
Configuration Properties, and enter your desired string in the Conditional Compilation
Constant section, as illustrated in Figure 5-5.
Figure 5-5. Specifying a preprocessor symbol in Visual Studio.

Using Attributes
The example program demonstrated an attribute with a single string parameter. Attributes
can take multiple parameters, and there can also be named parameters. Named parameters
are useful when there are many different parameters, and in a particular case you may use
only some of them. Named parameters can appear in any order.
As an example, the DllImport attribute takes a single positional parameter (the name of
the DLL) and several positional parameters. Here is an example of using the DllImport
attribute, with named parameters CharSet and CallingConvention:
[DllImport("KERNEL32.DLL", CharSet=CharSet.Unicode,
CallingConvention=CallingConvention.StdCall)]
We will see examples of the use of DllImport in Chapter 14, when we discuss the
Platform Invocation Service (or PInvoke), which enables you to call unmanaged code
through functions implemented in a DLL.
Attribute Targets
An attribute may be applicable to different kinds of entities. In the COM IDL example we
saw examples of attributes for interfaces, methods, and parameters. In .NET attributes
may be applied to many different kinds of entities, including
• assembly
• module
• class
• struct
• interface
• method
• parameter
and many more. The specification of legal entities to which an attribute may be applied is
part of the definition of an attribute, and you will get a compiler error message if you
attempt to use an attribute on the wrong kind of entity. When we discuss custom
attributes in Chapter 8, we will see how to specify the legal attribute targets for our own
attributes.
Summary
This chapter explored several important interactions between C# and the .NET
Framework, beginning with the root class object. We examined collections, including the
methods of the object class that should be overridden to tap into the functionality
provided by the .NET Framework. We introduced interfaces, which allow you to
rigorously define a contract for a class to implement. While a class in C# can inherit from
only one other class, it can implement multiple interfaces. Another benefit of interfaces is
that they facilitate very dynamic programs. C# provides convenient facilities to query a
class at runtime to see whether it supports a particular interface.
The interfaces supporting collections were examined in detail, and copy semantics were
explored. While C++ relies on a language feature of a copy constructor, in C# you
provide the capability by implementing a special interface, ICloneable. This led to an
exploration of the role of generic interfaces in the .NET Framework programming model
and to a comparison of the .NET and COM component models. A further illustration of
programming with generic interfaces was provided by sorting in different orders with the
IComparable interface. The examples offered insight into the workings of frameworks,
which are more than class libraries. In a framework, you call the framework, and the
framework calls you. Your code can be viewed as the middle layer of a sandwich. This
key insight can help you grasp what makes .NET programming "tick."
This behavior of being called into has been around for a long time in the form of callback
functions. The chapter included a careful examination of delegates and events. Two
simple and intuitive examples were presented: a stock market simulation and an online
chat room.
Finally, we covered attributes, which can be used to modify the behavior of entities of our
program according to our specifications.
This chapter concludes our exploration of the C# programming language. In the next
chapter we begin our detailed examination of the .NET Framework with a study of user
interface programming using Windows Forms.
Chapter 6. User Interface Programming Top
A fundamental feature of modern user interaction with a computer is a graphical user

interface or GUI. In this chapter we learn how to implement a GUI using the Windows
Forms classes of the .NET Framework. Practical Windows programming involves
extensive use of tools and wizards that greatly streamline the process. But all this
automation can obscure the fundamentals of what is going on. Hence we begin with the
basics, employing the .NET Framework SDK to create simple Windows applications
from scratch, without use of any special tools. We describe the fundamentals of drawing
in Windows Forms, using a font and a brush. We explain the principles of event handling
in Windows Forms and implement handlers for mouse events. We implement menus in
Windows Forms and corresponding event handlers. Controls are introduced.
At this point we switch over to using Visual Studio.NET, which makes it easy to create a
starter project, draw controls using a Forms Designer, create menus, add event handlers,
and perform other useful tasks. Dialog boxes are covered, and the listbox control is
introduced. We illustrate GUI programming by constructing a GUI for our Acme Travel
Agency case study.
Windows Forms Hierarchy
Windows Forms is that part of the .NET Framework that supports building traditional
GUI applications on the Windows platform. Windows Forms provides a large set of
classes that make it easy to create sophisticated user interfaces. These classes are
available to all .NET languages.
Your application will typically have a main window implemented by deriving from the
Form class. Figure 6-1 illustrates how your class derives from the Windows Forms
hierarchy.
Figure 6-1. Simplified Windows Forms class hierarchy.

Simple Forms Using .NET SDK
To gain insight into the fundamentals of Windows Forms it will be helpful to build a
simple application using only the .NET Framework SDK. See the program SimpleForm
with several progressive steps. None of these steps has a Visual Studio project. There is a
simple batch file build.bat that you should run at the command prompt.
Step 0:. A Simple Form
Here is a bare-bones Windows application. It is Step 0 of the example SimpleForm.
using System;
public class Form1: Form

{
public Form1()
{
Size = new System.Drawing.Size(300,200);
Text = "Simple Form - Step 0";
}
{
Application.Run(new Form1());
}
}
Our Form1 class inherits from System.Windows.Forms. The class System.Application

has static methods, such as Run and Exit, to control an application. The Main method
instantiates a new form and runs it as the main window.
The constructor of the form does initializations: The Size field sets the size of the new
form in pixels. The Text field specifies the caption to be shown in the title bar of the new
form.
The key to Windows Forms programming is the Form base class. This class contains a
great deal of functionality, which is inherited by form classes that we design.
You can build the application at the command line using the batch file build.bat. To run
the batch file, open up a DOS window and navigate to the SimpleForm\Step0 directory
and type build. Remember that you must have the environment variables set up properly,
which you can ensure by running the Visual Studio.NET Command Prompt.
csc /target:winexe /r:System.dll /r:System.Drawing.dll

/r:System.Windows.Forms.dll SimpleForm.cs
The target is a Windows executable, and there are references to the required .NET
libraries, System.dll, System.Drawing.dll, and System.Windows.Forms.dll.
After you have built the application using the batch file, you can run it by typing
SimpleForm at the command line. You can also double-click on the file SimpleForm.exe
in Windows Explorer. Figure 6-2 shows this simple application. Although trivial, it
already has a great deal of functionality, which is inherited from the Form base class. You
can drag the window around, resize it, minimize it, maximize it, open the system menu
(click in top left of the window), and so forth.
Figure 6-2. A bare-bones Windows Forms application (Step 0).

Windows Messages
Visual Studio.NET supplies a tool called Spy++, which can be used to "spy" on windows,
gaining some inkling of things taking place under the hood. Spy++ can be started from
the Visual Studio Tools menu. With the Step 0 version of SimpleForm.exe running, start
Spy++. Bring up the Find Window dialog from the menu Spy | Find Window. Click on
the Messages radio button. See Figure 6-3.
Figure 6-3. The Finder Tool lets you select a window to spy upon.
Using the left mouse button, drag the Finder Tool over the window of the SimpleForm
application and release the button. Now. as you interact with the SimpleForm window,
you will see windows messages displayed in a window of Spy++, as illustrated in Figure
6-4.
Figure 6-4. The Finder Tool lets you select a window to spy upon.
Windows applications are structured to handle events. The Windows operating system
sends messages to applications in response to user actions such as clicking a mouse
button, selecting a menu, typing at the keyboard, and so on. A Windows application must
be structured so that it can respond to such messages.
The nice thing about Windows programming using the .NET Framework classes is that
you program at a much higher level of abstraction. We have already seen how simple the
Step 0 application is. In the next several sections we will progressively implement some
basic features, illustrating the fundamentals of GUI programming using the Windows
Forms classes.
Step 1:. Drawing Text on a Form
Step 1 illustrates drawing text on a form. Figure 6-5 shows a run of the application.
Figure 6-5. Drawing text on a simple form (Step 1).

Performing output in Windows programs is very different from outputting in console
applications, where we simply use methods such as Console.WriteLine. Drawing output
in a window is referred to as "painting." Painting is done in response to a special kind of
message, a "paint" message WM_PAINT. This "on-demand" style of painting ensures that
the output of a window will be shown correctly even if the window is covered up and
uncovered again.
Another difference in output in Windows programs is that you have to specify details,
such as the coordinates at which it is drawn, a "brush" to draw with, a font for text, and so
forth. Here is the code for Step 1.
// SimpleForm.cs - Step 1
using System;
using System.Drawing;

{
private float x, y;
private Brush stdBrush;
public Form1()
{
x = y = 10;
stdBrush = new SolidBrush(Color.Black);
}
protected override void OnPaint(PaintEventArgs e)
{
e.Graphics.DrawString("Hello, Window Forms",
Font, stdBrush, x, y);
}
}
...
To draw in Windows Forms, you must override the virtual method OnPaint. The class
PaintEventArgs has a Graphics object as a read-only property. The Graphics class, part of
the System.Drawing namespace, has methods for drawing.
The DrawString method has parameters for:
• The string to be drawn

• The font (Font is a property of Form that gives the default font for the form)
• The brush to be used
• The pixel coordinates (as float numbers)
A black SolidBrush is constructed as our standard brush.
Windows Forms Event Handling
GUI applications are event-driven: The application executes code in response to user
events, such as clicking the mouse, choosing a menu item, and so on. Each form or
control has a predefined set of events. For example, every form has a MouseDown event.
Windows Forms employs the .NET event model, [1] which uses delegates to bind events to
the methods that handle them. The Windows Forms classes use multicast delegates. A
multicast delegate maintains a list of the methods it is bound to. When an event occurs in
an application, the control raises the event by calling the delegate for that event. The
delegate then calls all the methods it is bound to.
[1]
You may wish to review the discussion of delegates and events in Chapter 5.
C# provides the overloaded += operator for adding a delegate to an event. The following
code adds the Form1_MouseDown method to the MouseDown event.
MouseDown += new MouseEventHandler (Form1_MouseDown);
We will see this code in context shortly.
Events Documentation
You can find all the events associated with a class in the .NET Framework Reference.
The screen shot in Figure 6-6 shows the predefined events associated with the Form
class.
Figure 6-6. Documentation of events in the Form class.

MouseDown Event
One of the predefined events in the Control class, from which the Form class derives, is
MouseDown.
public event MouseEventHandler MouseDown;
Here is the declaration of MouseEventHandler:
public delegate void MouseEventHandler(

object sender,
MouseEventArgs e
);
The event handler receives a MouseEventArgs (derived from EventArgs), which has
read-only properties to provide information specific to this event:
• Button specifies which button (left, right, or the like) was pressed.
• Clicks indicates how many times the button was pressed and released.
• Delta provides a count of rotations of a mouse wheel.
• X and Y provide the coordinates where the mouse button was pressed.
Step 2:. Handling a Mouse Event
In Step 2 a mouse click (any button) will reposition the location of the greeting string.
Figure 6-7 shows the string relocated after we have clicked the mouse.
Figure 6-7. Clicking the mouse repositions the text (Step 2).
using System;
using System.Drawing;

{
private void InitializeComponent()
{
}
private float x, y;
private Brush stdBrush;
public Form1()
{
InitializeComponent();
x = y = 10;
}
protected void Form1_MouseDown (object sender,
MouseEventArgs e)
{
x = e.X;
y = e.Y;
Invalidate();
}
...
As part of its initialization, our program registers the Form1_MouseDown method with
the MouseDown event. This method sets the x and y coordinates of our text to the
location where the mouse was clicked. To understand the role of Invalidate, comment out
the code and build again. Click the mouse to relocate the greeting string. What happens?
The string is not relocated. Now cover the SimpleForm window with some other window
and then uncover it. Now you should see the string relocated.
The Invalidate method is defined in the Control base class. There are several overloaded
versions of this method. Each invalidates some region of the control and causes a paint
message to be sent to the control. The method with no parameters causes the entire
control to be invalidated. To minimize the amount of redrawing done, a more
sophisticated application might invalidate just a rectangle.
Step 2M:. Multiple Event Handlers
Step 2M illustrates tying two different event handlers to the MouseDown event. The
second handler merely displays a message box. [2]
[2]
A message box is a special kind of dialog box and will be discussed later in this
chapter.
// SimpleForm.cs - Step 2M
...

{
{
MouseDown += new MouseEventHandler (ShowClick);
}
...
MouseEventArgs e)
{
x = e.X;
y = e.Y;
Invalidate();
}
protected void ShowClick(object sender,
MouseEventArgs e)
{
MessageBox.Show("Mouse clicked!!!");
}
...
}
Step 3:. MouseDown and KeyPress Events
Step 3 of our demonstration illustrates handling an additional event, KeyPress, and also
distinguishing between left and right buttons in MouseDown.
HAndling Left and Right Buttons
We can distinguish between left and right buttons by using the Button property of the
MouseEventArgs parameter. Right button down is used for clearing the message string,
which is now stored in a StringBuilder data member str.

MouseEventArgs e)
{
if (e.Button == MouseButtons.Left)
{
x = e.X;
y = e.Y;
}
else if (e.Button == MouseButtons.Right)
{
str = new StringBuilder();
}
Invalidate();
}
KeyPress Event
Step 3 also illustrates handling a KeyPress event. Every time the user presses a key, the
corresponding character is appended to the greeting string. Note use of the StringBuilder
class, which is more efficient in this context than string. String is immutable, and hence
string objects would be continually created and destroyed while we appended characters.
private StringBuilder str;

...
protected void Form1_KeyPress (object sender,
KeyPressEventArgs e)
{
str.Append(e.KeyChar);
Invalidate();
}
As with Step 2 we call Invalidate to force a repaint after we have made a change in the
data to be displayed. Figure 6-8 illustrates our SimpleForm window after the starting text
has been cleared and some new text typed in.
Figure 6-8. Exercising mouse and key press events (Step 3).
Menus
As a user of Windows applications you should be acquainted with menus, which provide
a simple mechanism for choosing commands. In .NET menus are implemented in code.
There is no separate resource file.
Step 4:. A Menu to Exit the Program
Step 4 of our SimpleForm program illustrates adding a simple menu. File | Exit is used to
exit the program. See Figure 6-9.
Figure 6-9. A File | Exit menu is added to our form (Step 4).
Menu Code
...
private MenuItem menuExit;
private MenuItem menuFile;
private MainMenu mainMenu1;
public Form1()
{
x = y = 10;
str = new StringBuilder("Hello, Windows Forms");
}
{
mainMenu1 = new MainMenu ();
menuFile = new MenuItem ();
menuExit = new MenuItem ();
// mainMenu1
mainMenu1.MenuItems.Add(menuFile);
// menuFile
menuFile.Index = 0;
menuFile.MenuItems.Add(menuExit);
menuFile.Text = "File";
// menuExit
menuExit.Index = 0;
menuExit.Text = "Exit";
menuExit.Click += new EventHandler(menuExit_Click);
Menu = mainMenu1;
...
The code in InitializeComponent builds up the hierarchical menu structure, represented

by an instance of the MainMenu class. A menu is composed of MenuItem objects that
represent the individual menu commands in the menu structure. Each MenuItem can be a
command for your application or a parent menu for other submenu items. You bind the
MainMenu to the Form that will display it by assigning the MainMenu to the Menu
property of the Form.
When we discuss the Forms Designer later in the chapter, we will see that it is easy to
create a menu by dragging a MainMenu control from the toolbox to the form. The Forms
Designer will take care of generating appropriate boilerplate code.
Menu Event Code
A delegate is hooked to the event, as with other Windows Forms events. Clicking on a
menu item causes the corresponding command to be executed.

{
...
menuExit.Click +=
new EventHandler(this.menuExit_Click);
...
}
private void menuExit_Click(object sender,

EventArgs e)
{
Application.Exit();
}
...
Controls
In the program we have just discussed, mainMenu1 is an example of a control. It is an

instance of the MainMenu class. A control is an object that is contained within a form and
is used to add functionality to the form. A control can perform many tasks automatically
on behalf of its parent form. It simplifies programming, as you do not have to be
concerned with painting, invalidating, working with graphics elements, and so forth. The
simple menu that we just illustrated would have required a substantial amount of code if
we had to implement it from scratch. Controls provide rich reusable code—a big benefit
from programming with objects.
Step 5:. Using a TextBox Control
Step 5 of our SimpleForm application illustrates using a TextBox control to display our
greeting text. As with earlier versions of the application, you can reposition the greeting
by clicking the left mouse button, and you can clear the greeting by clicking the right
mouse button. You can also type in your own greeting text. Now you have full editing
capability. You can insert characters wherever you wish in the control, cut and paste
(Ctrl+X and Ctrl+V), and so forth. All of this editing capability is provided by the
TextBox control. Figure 6-10 illustrates the application after the greeting has been
repositioned and we have typed in some text of our own.
Figure 6-10. The greeting text is now displayed using a control (Step 5).
Here is the new version of our program. Note that it has both greater simplicity and more
functionality. We no longer need member variables for the coordinates or text of the
greeting string (this information is now stored in the TextBox control txtGreeting). We do
not need OnPaint any longer, either, because the text box knows how to paint itself. We
can then also get rid of the brush. We don't need to handle KeyPress events, because this
functionality is handled (in a much more full-blown way) by the TextBox control.
...

{
private TextBox txtGreeting;
private MenuItem menuExit;
private MenuItem menuFile;
private MainMenu mainMenu1;
public Form1()
{
}
{
mainMenu1 = new MainMenu ();
menuFile = new MenuItem ();
menuExit = new MenuItem ();
// mainMenu1
mainMenu1.MenuItems.Add(menuFile);
// menuFile
menuFile.Index = 0;
menuFile.MenuItems.Add(menuExit);
menuFile.Text = "File";
// menuExit
menuExit.Index = 0;
menuExit.Text = "Exit";
menuExit.Click += new EventHandler(menuExit_Click);
Menu = mainMenu1;
// txtGreeting
txtGreeting = new TextBox();
txtGreeting.Location = new Point(10, 10);
txtGreeting.Size = new Size(150, 20);
txtGreeting.Text = "Hello, Windows Forms";
Controls.Add(txtGreeting);
this.MouseDown +=
new MouseEventHandler (Form1_MouseDown);
}
MouseEventArgs e)
{
if (e.Button == MouseButtons.Left)
{
txtGreeting.Location = new Point(e.X, e.Y);
}
else if (e.Button == MouseButtons.Right)
{
txtGreeting.Text = "";
}
}
private void menuExit_Click(object sender, EventArgs e)
{
Application.Exit();
}
{
Application.Run(new Form1());
}
}
Using the TextBox control is very easy. As part of the initialization we instantiate it and
assign the Location, Size, and Text properties. We add our new control to the Controls
collection of our form. In the mouse event handler we reposition the control by assigning
the Location property. We clear the text by assigning the Text property.
Visual Studio.NET and Forms
Although it is perfectly feasible to create Windows Forms applications using only the
command-line tools of the .NET Framework SDK, in practice it much easier to use
Visual Studio.NET. You can get started by creating a Windows Application project, which
provides starter code and sets up references to the required .NET libraries. You can then
use the Forms Designer to drag and drop controls from a toolbox onto your forms. The
Forms Designer inserts all the needed boilerplate code to make your controls work within
your forms. There is a Properties window which makes it easy to set properties of your
controls at design time. You can, of course, also set properties at runtime, which is what
we did with our txtGreeting text box in the code shown previously.
The same Forms Designer can be used in all .NET languages. A similar Designer is
available for visually drawing Web Forms, which we will discuss in Chapter 10 on
ASP.NET.
Windows Forms Demonstration
The best way to become acquainted with using Visual Studio.NET to create Windows
applications is to build a small application from scratch yourself. Our demonstration
creates a Windows application to make deposits and withdrawals from a bank account.
Do all your work in the Demos directory for this chapter.
1. Create a new C# project BankGui of type Windows Application in the Demos folder.
See Figure 6-11.
Figure 6-11. Creating a new Windows Application project.

2. Open up the Toolbox by dragging the mouse over the vertical Toolbox tab on the left
side of the main Visual Studio window. If the Toolbox tab does not show, you can
open it from the menu View | Toolbox. You can make the Toolbox stay open by
clicking on the "push-pin" next to the X on the title bar of the Toolbox. (The little
yellow box will say "Auto Hide" when you pause the mouse over the push-pin.)
3. From the Toolbox, drag two labels, two textboxes, and two buttons to the form. See
Figure 6-12.
Figure 6-12. Dragging controls from the Toolbox onto a form.

4. Click on label1 in the Forms Designer. This will select that control in the Properties
window, just beneath the Solution Explorer. You can use the Properties window to
make changes to properties of controls. Change the Text property to Amount. After
you type the desired value, hit the carriage return. You will then see the new text
shown on the form. Figure 6-13 shows the Properties window after you have changed
the Text property of the first label.
Figure 6-13. Changing property values in the Properties window.

5. Similarly, change the text of label2 to Balance.
6. Enter property values for the textboxes and buttons, as shown in Table 6-1.
Table 6-1. Property Values for Textboxes and Buttons

Name Text
txtAmount (blank)
txtBalance (blank)
cmdDeposit Deposit
cmdWithdraw Withdraw
1. Resize the form by dragging the sizing handles on the middle of each side.
Reposition the controls as desired by dragging with the mouse, and resize the
controls with the mouse, if you wish. When you are satisfied with the appearance
of your form, save the project. Your form should now look similar to Figure 6-14.
Figure 6-14. Form for BankGui application.

2. Add event handlers for the buttons by double-clicking on each button.
3. Add the following code:
4. public class Form1 : System.Windows.Forms.Form
5. {
6. ...
7. public Form1()
8. {
9. //
10. // Required for Windows Form Designer support
11. //
12. InitializeComponent();
13.
14. //
15. // TODO: Add any constructor code after
16. // InitializeComponent call
17. //
18. txtAmount.Text = "25";
19. txtBalance.Text = "100";
20. }
21. ...
22. /// <summary>
23. /// The main entry point for the application.
24. /// </summary>
25. [STAThread]
26. static void Main()
27. {
28. Application.Run(new Form1());
29. }
30.
31. private void cmdDeposit_Click(object sender,
32. System.EventArgs e)
33. {
34. int amount = Convert.ToInt32(txtAmount.Text);
35. int balance = Convert.ToInt32(txtBalance.Text);
36. balance += amount;
37. txtBalance.Text = Convert.ToString(balance);
38. }
39.
40. private void cmdWithdraw_Click(object sender,
42. {
43. int amount = Convert.ToInt32(txtAmount.Text);
44. int balance = Convert.ToInt32(txtBalance.Text);
45. balance -= amount;
46. txtBalance.Text = Convert.ToString(balance);
}
47. Build and run the application. It should behave like a standard Windows
application. You should be able to make deposits and withdrawals. Figure 6-15
illustrates the running application.
Figure 6-15. The BankGui Windows application.
Design Window and Code Window
The most important thing to understand about navigating Windows Forms projects in
Visual Studio is switching between the Design window, where you work with controls on
a form, and the Code window, where you work with source code. We can illustrate these
two windows from the Demos\VsForm project, where we have provided starter code
corresponding to VsForm\Step1 in the main directory for this chapter. The starter project
simply displays a fixed greeting string. The state of the project at various points in the
demonstration is captured in other numbered steps.
If you double-click on VsForm.sln (Demos directory) in the Solution Explorer, you will
bring up the Design window, as shown in Figure 6-16.
Figure 6-16. The Design window in a Windows Forms project.

To bring up the Code window, click on the "View Code" toolbar button in the
Solution Explorer. This will open up the source code, and you will see horizontal tabs at
the top of the principal window area, allowing you to select among the open windows.
Now the Design window and the Code window for this one form are open. You may also
go back to the Design window by clicking on the "View Designer" toolbar button.
Figure 6-17 shows the open Code window.
Figure 6-17. The Code window in a Windows Forms project.

Adding an Event
1. Build and run the starter program. This is a completely static application—it merely
displays a greeting at a fixed location.
2. Open up the Design window of the form and click on the Events button of the
Properties window.
3. Find the MouseDown event. See Figure 6-18.
Figure 6-18. Adding an event by using the Events button.

4. Double-click. This will automatically generate code to register a delegate for the
event and provide a skeleton for a method tied to the delegate. [3]
[3]
If you cannot see this Windows Form Designer generated code, click on the little
"+" on the extreme left of the editor window to open up the hidden "region."
{
...
this.MouseDown +=
new System.WinForms.MouseEventHandler
(this.Form1_MouseDown);
}

System.WinForms.MouseEventArgs e)
{
}
...
Code for Event Handler

5. Add the highlighted code to the mouse down event handler to set the coordinates of
the greeting message. Don't forget to call Invalidate!
System.WinForms.MouseEventArgs e)
{
x = e.X;
y = e.Y;
Invalidate();
}
6. Build and run. You should now be able to relocate the greeting by clicking the mouse
button (either button will work). The project now corresponds to VsForm\Step2.
Using the Menu Control

1. Open up the Toolbox if not already open (click on the Toolbox vertical tab) and drag
the MainMenu control onto the form.
2. Type "File" and "Exit," creating a popup menu File with a menu item Exit. See Figure
6-19.
Figure 6-19. Use the Menu Control to add a menu to a form.
3. In the Properties window change the names of your two menu items to "menuFile"
and "menuExit."
4. Double-click on "Exit" to add code for a File | Exit event handler.
5. Add code to the handler to exit the application.

protected void menuExit_Click (object sender,
System.EventArgs e)
{
Application.Exit();
}
6. Build and run. Your menu should be operational. The project now corresponds to
VsForm\Step3.
Closing a Form
As an interesting modification to our program, let us arrange it so that whenever the user
attempts to close the application, the user will be queried on whether to really close.
There are several ways a window can be closed:
• From the "X" at top right of the window

• From the system menu at the top left of the window
• By the keyboard Alt + F4
• In our application, by File | Exit
When a form is about to close, the Closing event is raised. You may stop the closing by
setting the Cancel property in the handler for this event. (First add a handler for the event
Closing in the usual way.) Just type in the MessageBox code as shown.
protected void Form1_Closing (object sender,

System.ComponentModel.CancelEventArgs e)
{
DialogResult status = MessageBox.Show(
"Do you want to close",
"Simple Form (VS)", MessageBoxButtons.YesNo);
if (status == DialogResult.No)
{
e.Cancel = true;
}
}
To tap into this behavior, in your handler for File | Exit you should not exit the
application but instead close the main window by calling the Close method:
protected void menuExit_Click (object sender,

System.EventArgs e)
{
//Application.Exit();
Close();
}
The project now corresponds to VSForm\Step4. Run your program and try closing in
various ways. You should always see the dialog box shown in Figure 6-20.
Figure 6-20. Dialog box that queries the user whether or not to close.
Dialog Boxes
Dialog boxes provide a more elaborate way for a user to interact with a Windows
application. A dialog box can provide a number of controls to facilitate data input. The
code in the previous section illustrated use of a simple message box dialog that allowed
the user to answer a yes or no question. This kind of dialog can be created using the
MessageBox class. You can implement more general dialog boxes by creating forms for
them.
We will illustrate a number of dialogs through a graphical user interface to our Acme
Travel Agency case study. As usual, the case study code is in the CaseStudy folder for
this chapter. Let's begin by examining a simple dialog that is used for adding a new hotel
to our list of hotels. Build and run the case study. In the main form click the "Add..."
button. [4] The "New Hotel" dialog is brought up, as illustrated in Figure 6-21.
[4]
The three dots are a Windows UI style that indicates the program will not carry out the
action immediately but will prompt the user for additional input, typically through a
dialog box.
Figure 6-21. Dialog box for adding a new hotel.
The user can now enter data. Clicking the OK button will cause the information to be
accepted. Clicking the Cancel button will cause the new data to be ignored. This dialog
box (like the message box in the previous section) is a modal dialog, which means that
the user cannot work elsewhere in the application until the dialog is closed. If you try do
something else on the main form while the "New Hotel" dialog is open—for example,
click another button—you will hear a beep. The other kind of dialog is modeless, which
will allow the user to work elsewhere in the application while the dialog is open.
Dialog boxes normally have special characteristics as forms. For example, they typically
do not have a system menu, they have no minimize or maximize buttons, and they have a
border that does not permit them to be resized. You can examine these features with the
"New Hotel" dialog.
Continuing the demonstration, enter some data for a new hotel and click OK. You will
now be brought back to the main form, and your new hotel will be shown in the list of
hotels, as illustrated in Figure 6-22. The main form also illustrates some additional GUI
features, such as a list box for displaying a list of hotels and a multiline text box that can
display text that is too long to fit on one line.
Figure 6-22. Main form for hotel administration.
.NET Dialog Documentation
Dialogs are explained clearly in the Documentation in the .NET Framework. Look in
"Dialog Boxes in Windows Forms" under "Introduction to Windows Forms." It is
noteworthy that the principles of dialog boxes are the same in all .NET languages. This is
in sharp contrast to the days before .NET, where, for example, dialogs in Visual Basic and
in Microsoft Foundation Classes were totally different. Figure 6-23 shows the entry point
to this documentation.
Figure 6-23. Documentation on dialog boxes using the .NET Framework.
Dialog Box Demonstration
We will demonstrate the implementation details of a dialog box by creating a dialog to

change hotel information in a simplified version of our case study. Do your work in the
folder Demo\HotelAdmin. The starter code is backed up in the folder HotelAdmin\Step1
in the main folder for this chapter. The completed program is in HotelAdmin\Step3. You
may run either the case study or the Step 3 solution to see what the completed dialog
should look like. In the main form select a hotel by clicking in the list box of hotels. Then
click on the "Change..." button. This brings up the "Change Hotel Information" dialog, as
illustrated in Figure 6-24. Notice that the City and Hotel Name are grayed out. These
items are read-only and cannot be changed. The user can enter new information for the
Rooms and Rate.
Figure 6-24. Dialog for changing hotel information.

Creating a Modal Dialog
The first part of our demonstration illustrates how to create a modal dialog box. We show
how to set properties appropriately for the dialog and how to return a dialog result
through use of OK and Cancel buttons.
1. Build and run the starter application. The "Add..." and "Delete" buttons work, but
there is only a stub for "Change...", which brings up an empty form. This form is
ordinary, with system menu, minimize and maximize buttons, resizability, and so on.
2. Open up ChangeHotelDialog.cs in Design mode. In the Properties window, change

the FormBorderStyle property to FixedDialog.
3. Set the ControlBox, MinimizeBox, and MaximizeBox properties to False. If you like,
you may build and run the application at this point. The dialog now is not resizable,
and there is no system menu and no "X" in top right to close the window. [5]
[5]
You may use Alt+F4 to close the window.
4. The next job is to enter labels and text boxes for the hotel information, plus OK and
Cancel buttons. You may practice using the Toolbox to add these controls.
Alternatively, you may copy and paste from NewHotelDialog.cs (open both files in
Design mode).
5. If you used copy and paste, the controls will have proper Name and Text properties.
Otherwise, assign values as shown in Table 6-2.
Table 6-2. Property Values for Textboxes and Buttons for
ChangeHotelDialog.cs
Name Text
txtCity (blank)
txtHotelName (blank)
txtNumberRooms (blank)
txtrate (blank)
cmdOK OK
cmdCancel Cancel
6. Change the ReadOnly property of txtCity and txtHotelName to true.
7. Resize the form to better fit the controls we have added.
8. Set the DialogResult property of the OK button to OK. Similarly set the property of
the Cancel button to Cancel. Save ChangeHotelDialog.cs.
9. In MainAdminForm.cs, add temporary code to the cmdChange_Click handler to

display "OK" or "Cancel" in the Messages text box, depending on whether the dialog
was closed by clicking OK or Cancel. Notice that a dialog is brought up by the
method ShowDialog in place of Show, which is used for ordinary forms. ShowDialog
returns a result as an enum of type DialogResult.
private void cmdChange_Click(object sender,
System.EventArgs e)
{
ChangeHotelDialog dlg = new ChangeHotelDialog();
DialogResult status = dlg.ShowDialog();
if (status == DialogResult.OK)
{
txtMessages.Text = "OK";
}
else
{
txtMessages.Text = "Cancel";
}
}
10. Build and test. You should now be able to bring up the dialog from the menu, and
either the OK or Cancel button will close the dialog, and a corresponding message
will be displayed. You can verify that the dialog is modal by trying to click elsewhere
in the application. The program is now at Step 2.
Passing Information Between Parent Form and a Dialog
The second part of our demonstration shows how to pass information to a dialog and how
to retrieve information from a dialog. The .NET Framework classes do not provide a
built-in mechanism for this purpose, but there is a design pattern you can follow. You
create a property in the dialog class for each piece of information you wish to pass
between the parent form and the dialog.
In our example we implement write-only [6] properties for City and HotelName and read-
write properties for Rate and NumberRooms.
[6]
The properties are write-only from the perspective of the dialog class, because we pass
information a dialog instance. The corresponding controls are read-only, because the user
is not allowed to enter new information.
1. Add code to ChangeHotelDialog.cs to implement these properties.

2. public string City
3. {
4. set
5. {
6. txtCity.Text = value;
7. }
8. }
9. public string HotelName
10.{
11. set
12. {
13. txtHotelName.Text = value;
14. }
15.}
16.public int NumberRooms
17.{
18. get
19. {
20. return Convert.ToInt32(txtNumberRooms.Text);
21. }
22. set
23. {
24. }
25.}
26.public decimal Rate
27.{
28. get
29. {
30. return Convert.ToDecimal(txtRate.Text);
31. }
32. set
33. {
34. txtRate.Text = value.ToString();
35. }
}
2.Now add code to the main form MainAdminForm.cs to set these properties prior to
bringing up the dialog and to use the properties if the dialog box closes via an OK.
Comment out or delete your previous test code that displays "OK" or "Cancel" in the
Messages box.
36.private void cmdChange_Click(object sender,

38.{
39. ChangeHotelDialog dlg = new ChangeHotelDialog();
40. if (currHotel.HotelName != "")
41. {
42. dlg.City = currHotel.City;
43. dlg.HotelName = currHotel.HotelName;
44. dlg.NumberRooms = currHotel.NumberRooms;
45. dlg.Rate = currHotel.Rate;
46. }
47. else
48. {
49. MessageBox.Show("Please select a hotel",
50. "Hotel Broker Administration",
51. MessageBoxButtons.OK,
52. MessageBoxIcon.Exclamation
53. );
54. return;
55. }
56. DialogResult status = dlg.ShowDialog();
57. if (status == DialogResult.OK)
58. {
59. string comment = hotelBroker.ChangeRooms(
60. currHotel.City,
61. currHotel.HotelName,
62. dlg.NumberRooms,
63. dlg.Rate);
64. if (comment == "OK")
65. {
66. ShowHotelList(hotelBroker.GetHotels());
67. txtMessages.Text = "Hotel " + currHotel.HotelName
68. + " has been changed";
69. }
70. else
71. txtMessages.Text = comment;
72. }
}
The structure currHotel holds the fields of the hotel that is currently selected in
the list box. In the next section we will see how to extract information from a list
box and how to populate a list box.
3.Build and test. Your dialog should now be fully operational. Your project should
now correspond to HotelAdmin\Step3.
ListBox Control
The .NET Framework provides a number of controls that you can use to display lists of
items to the user. These controls also allow the user to select an item from the list,
typically by clicking on the item to be selected. In this section we examine the ListBox
control.
Our example program is HotelAdmin\Step3. The main form in MainAdminForm.cs

contains the listbox listHotels, which maintains a list of hotels. Each hotel is represented
by a string with values separated by commas.
Populating a ListBox
When the HotelAdmin program starts up, it populates the listbox listHotels with a list of
hotels as part of the initialization in the MainAdminForm constructor.
public MainAdminForm()
{
//
// Required for Windows Form Designer support
//
//
// TODO: Add any constructor code after
// InitializeComponent call
//
hotelBroker = new HotelBroker();
ShowHotelList(hotelBroker.GetHotels());
}
The ShowHotelList method displays an array list of hotels in a listbox. This array list is
obtained by calling HotelBroker.GetHotels. Here is the code for ShowHotelList.
private void ShowHotelList(ArrayList array)

{
listHotels.Items.Clear();
if (array == null)
{
return;
}
for each(HotelListItem hotel in array)
{
string city = hotel.City.Trim();
string name = hotel.HotelName.Trim();
string rooms = hotel.NumberRooms.ToString();
string rate = hotel.Rate.ToString();
string str = city + "," + name + ","
+ rooms + "," + rate;
listHotels.Items.Add(str);
}
}
A ListBox has a property Items which maintains a collection of object references. We

first call Items.Clear to clear out the listbox of items currently being displayed. We then
loop through the hotels in the array list and build up a string consisting of the fields of the
hotel structure, separated by commas. This string is added to the listbox by calling
Items.Add.
Selecting an Item From a ListBox
An item in a listbox is selected by clicking on the item, generating a

SelectedIndexChanged event. You can access the selected item through the SelectedIndex
and SelectedItem properties. If no item is selected, SelectedIndex is -1. Here is the code
for the event handler for SelectedIndexChanged.
private void listHotels_SelectedIndexChanged(object sender,

System.EventArgs e)
{
if (listHotels.SelectedIndex != -1)
{
string selected = (string) listHotels.SelectedItem;
char[] sep = new char[] {','};
string[] fields;
fields = selected.Split(sep);
currHotel = new HotelListItem();
currHotel.City = fields[0];
currHotel.HotelName = fields[1];
currHotel.NumberRooms = Convert.ToInt32(fields[2]);
currHotel.Rate = Convert.ToDecimal(fields[3]);
}
else
{
currHotel.HotelName = "";
}
}
Since the items in a listbox are stored as object references, we cast the selected item to a
string. We use String.Split to extract the fields that are separated by commas and store
them in the fields string array. The values are then moved from the array and stored in
currHotel. In the previous section we saw currHotel used to initialize the "New Hotel"
and "Change Hotel Information" dialog boxes.
Acme Travel Agency Case Study—Step 3
The Acme Travel Agency case study was introduced in Chapter 4, where we used arrays
as our data structures for storing lists of hotels, customers, and reservations. In Chapter 5
we changed the implementation to use collections in place of arrays. We also specified a
number of interfaces, and we passed lists as ArrayList object references. We provided a
command-line user interface. In the CaseStudy folder of the present chapter we provide a
graphical user interface, implemented by using Windows Forms.
We have already looked at the main window (see Figure 6-22), which is the same as in
the simplified HotelAdmin [7] program we used to illustrate dialog boxes. The "Add..."
button lets us add a new hotel (Figure 6-21), and the "Change..." button (Figure 6-24) lets
us change the number of rooms and the rate of a hotel. The "Delete" button will delete the
currently selected hotel.
[7]
The HotelAdmin program provides only empty forms as stubs for the "Customers…"
and "Reservations…" buttons.
The "Customers..." button brings up a "Customer Management" form, which shows a list
of currently registered customers. You may select a customer by clicking in the listbox.
Figure 6-25 shows this form after selecting a customer.
Figure 6-25. Customer Management form.
The Id of the selected customer is shown in a textbox. You may unregister this customer
by clicking "Unregister." You may change the email address of this customer by clicking
"Change Email," which will bring up a dialog box. You may display the information for
just this one customer by clicking "One Customer." The "All Customers" button will
again show all the customers in the listbox. The "Register" button lets you add a new
customer.
The third major form of our user interface is "Hotel Reservations," which is brought up
from the main administration form by clicking "Reservations...." To make a reservation,
enter the Customer Id, Checkin Date, and Number of Days. You may specify the City and
Hotel Name by selecting a hotel from the listbox. To make the reservation, you then
simply click the "Make Reservation" button. To show all the reservations for a customer
with a particular Customer Id, [8] click "Show Reservations." Figure 6-26 shows this form
after the customer whose Id is 1 has made a reservation and we have shown the
reservations for this customer.
[8]
A Customer Id of -1 will show the reservations for all customers.
Figure 6-26. Hotel Reservations form.
You may clear the reservations listbox by clicking the "Clear Reservations" button. The
"Cancel Reservation" will cancel the reservation with a particular Reservation Id, which
may either be typed in or selected by clicking in the Reservations listbox.
The Acme Travel Agency case study is used extensively in the following chapters, so you
may wish to experiment with it at this point. The graphical user interface makes
exercising the case study much easier than our previous command-line interface. On the
other hand, the command-line interface and a simple global try block around the whole
command loop made it easy to check for all exceptions. Such an approach is not feasible
for a GUI program. In an industrial-strength application you should check for exceptions
wherever they may occur. Our case study is simplified for instructional purposes, and we
have not attempted to be thorough in catching exceptions. Another simplification we
made is not checking that a Customer Id used in making a reservation corresponds to a
real, registered customer. The database implementation in Chapter 9 does provide such a
check.
Summary
In this chapter we learned how to implement a GUI using the Windows Forms classes of
the .NET Framework. We began with first principles, using the .NET Framework SDK to
create simple Windows applications from scratch, without use of any special tools.
Drawing is done in an override of OnPaint using a font and a brush. The .NET event
mechanism is used to handle user interaction such as mouse events and pressing keys.
Controls simplify Windows programming. A menu control makes it easy to add menus to
a Windows program. Visual Studio.NET greatly simplifies Windows programming. The
Forms Designer lets you drag controls from the Toolbox onto your forms, and you can set
properties of the controls at design time. You can also easily add event handlers. Dialog
boxes are a special kind of form, and you can pass information between a parent form
and a dialog through use of properties in the dialog. The listbox control makes it easy to
display lists of information.
We concluded the chapter by presenting a graphical user interface for our Acme Travel
Agency Case Study.
Chapter 7. Assemblies and Deployment Top
Deployment makes the programmer's hard work available to the customer. .NET
assemblies make deployment much simpler and much more reliable than current
Windows deployment. Private assembly deployment is as simple as copying the
component assembly into the same directory as the client program. Alternatively, shared
assembly deployment places the component with a unique name (known as a strong
name) in the global assembly cache, which makes it available for general use.
This chapter begins with a look at assemblies, which are the fundamental unit of
deployment in .NET. Private assembly deployment and shared assembly deployment are
described next. Versioning and digital signing of assemblies are discussed in the context
of shared deployment. Finally, the Visual Studio.NET deployment and setup wizards are
introduced. Throughout our discussion we illustrate a number of useful tools that are part
of the .NET Framework SDK.
Assemblies
In .NET, Assemblies are components. Assemblies, which may be composed of one or

more DLL or EXE files, are the unit of deployment. You do not deploy individual DLLs
or EXEs. Security evidence and versioning are based on the assembly. Assemblies
contain Microsoft Intermediate Language (MSIL) instructions, resource data, and
metadata. Since metadata describes the content of the assembly, the assembly does not
require any external description, such as in the system registry. .NET components are
much simpler and less error prone to install and uninstall than traditional COM
components, which had extensive registry entries.
A digital signature is required before an assembly can be deployed in the global assembly
cache. Digitally signed assemblies provide cryptographically generated verification
information that can be used by the CLR to enforce crucial dependency rules when
locating and loading assemblies. This is distinct from the security verification that is done
to make sure that code is type safe.
The identity of an unsigned assembly is defined simply as a human-readable name, along

with a version number. The identity of a digitally signed assembly is defined by a unique
cryptographic key pair. Optionally, an assembly's identity may also include a culture code
for supporting culturally specific character sets and string formats.
An assembly's version can be checked, so that the CLR can insure that the same assembly
version with which the client was built and tested is loaded. This eliminates the infamous
"DLL Hell" problem, where Windows applications could easily break when an older
version was replaced with a newer version (or vice versa). A digitally signed assembly
can be used to verify that the assembly contents were not altered after it was digitally
signed. Not only will you not accidentally use the wrong version, but you will not be
tricked into using a maliciously tampered component that could do serious harm.
Although there is often a one-to-one correspondence between namespace and assembly,

an assembly may contain multiple namespaces, and one namespace may be distributed
among multiple assemblies. While there is often a one-to-one correspondence between
assembly and binary code file (i.e., DLL or EXE), one assembly can span multiple binary
code files. An assembly is the unit of deployment; an application is the unit of
configuration.
Contents of an Assembly
For our next step of the case study, we split our Hotel Administrator's program into three
assemblies. The example CaseStudy directory for this chapter has an AcmeGui
application program (EXE), and two component (DLL) assemblies: Customer and Hotel.
The code associated with the customer and hotel classes has been moved to the
appropriate assemblies. When we discuss configuration later in the chapter, it is the
AcmeGui application that will be configured.
We will use the Customer and Hotel assemblies to understand the issues associated with
deployment. All public members of the Customer and Hotel assembly will be visible to
code outside of their respective assemblies. Members marked as internal can be used only
within the assembly.
If you look at Figure 7-1, you will see that the Solution Explorer shows that the AcmeGui
project has references to the Customer and Hotel dynamic link libraries. These references
enable the compiler to find the Hotel and Customer types used by AcmeGui and then
build the application. They do not dictate where the DLLs have to be when the project is
deployed; we will explain how this works when we discuss deployment. You will also
notice references made to system assemblies such as System.dll. Looking at the
properties for the reference will show you where the assembly is located. [1]
[1]
Select the assembly in the Solution Explorer, right-mouse click, select Properties in the
context menu.
Figure 7-1. AcmeGui's Solution Explorer showing references.

Creating a DLL is simple. Just select "Class Library" from the New Project Wizard in
Visual Studio.NET, specify a location and name, and then start coding. To setup a
reference to another DLL from your project you use the Add Reference menu item from
the Visual Studio.NET Project menu. Navigate to the DLL you want, select it with the
Select button, then click the OK button. [2] Every Assembly has a Manifest that describes
the metadata information associated with the Assembly. A manifest provides the
following information about an assembly:
[2]
It is straightforward to go from the monolithic program we had in the previous chapter
to the componentized one we have now.
Create two new Class Library projects in the AcmeGui Solution for Customer and Hotel.
In Visual Studio select File | New | Project. In the dialog box that comes up, select Visual
C# projects in the left top pane, then select Class Library in the right top pane. Enter the
name of the project (Customer or Hotel) and make sure the Add to Solution radio button
is selected.
Remove the appropriate files from the AcmeGui project and add them to the appropriate
project. In the Solution Explorer, select the file in the AcmeGui project, right-mouse
click, select exclude from project. Then in the Solution Explorer select the appropriate
project and right-mouse click, select Add, then Add Existing Item, navigate to the
appropriate file and select it, and hit the open button. You can select more than one file at
a time.
Build the two component projects by selecting their project name in the Solution
Explorer and select the build option for the assembly in the Build menu. Since we no
longer have a monolithic application, we have to indicate to the compiler how to resolve
references to the Customer and Hotel classes. Select the AcmeGui project in the Solution
Explorer, right-mouse click, then select Add Reference. Click on the Projects tab and you
should see the Customer and Hotel dlls there. Select them both and then hit the select
button. You should see both dynamic link libraries in the bottom list. Then click the OK
button. Now when you rebuild the solution, the AcmeGui project will compile and run.
You can click on the plus button next to References in any project to see what
dependencies it has.
• Assembly identity based on name, version, culture, and— optionally—a digital

signature
• Files that contribute to the assembly contents
• Other assemblies on which the assembly is dependent
• Permissions required by the assembly
Every assembly created by Visual Studio has a file, AssemblyInfo.cs, containing the
following attributes that can be used to set the information associated with an assembly:
[assembly: AssemblyTitle("")]
[assembly: AssemblyDescription("")]
[assembly: AssemblyConfiguration("")]
[assembly: AssemblyCompany("")]
[assembly: AssemblyProduct("")]
[assembly: AssemblyCopyright("")]
[assembly: AssemblyTrademark("")]
[assembly: AssemblyCulture("")]
[assembly: AssemblyVersion("1.0.*")]
[assembly: AssemblyDelaySign(false)]
[assembly: AssemblyKeyFile("")]
[assembly: AssemblyKeyName("")]
To explore how versioning, digital signing, and deployment work, we use the ILDASM
tool introduced in Chapter 2 to view the appropriate metadata. Visual Studio.NET installs
with ILDASM on the Tools menu. You can also find it in your \Program
Files\Microsoft.Net\FrameworkSDK\Bin directory.
Figure 7-2 shows the top level that you see when you open the Customer.dll assembly in
ILDASM and double-click on the OI.NetCs.Acme namespace. You see entries for the
MANIFEST, the Customers and Customer classes, the ICustomer interface, and the
CustomerListItem value type. Clicking on the plus (+) button will expand an entry.
Figure 7-2. Top-level ILDASM view of Customer component.
To view the manifest, double-click the MANIFEST node shown in Figure 7-2; the
resulting manifest information is displayed in Figure 7-3. Some of the numbers will vary
if you have rebuilt any of the samples, or you have a later version of .NET.
Figure 7-3. ILDASM showing manifest of Customer.dll.

The manifest contains information about the dependencies and contents of the assembly.
You can see that the manifest for Customer contains, among others, the following
external dependency. [3]
[3]
If you have rebuilt any of the components, you will, of course, see different build and
revision numbers.
.assembly extern mscorlib

{
.publickeytoken = (B7 7A 5C 56 19 34 E0 89 )
.ver 1:0:2411:0
}
The .assembly extern mscorlib metadata statement indicates that the Customer assembly
makes use of, and is therefore dependent on, the standard assembly mccorlib.dll, which is
required by all managed code. When an assembly makes a reference to another assembly,
you will see an .assembly extern metadata statement. If you open AcmeGui in ILDASM
and look at the manifest, you will see dependencies on the Customer and Hotel
assemblies as well as the System.Drawing assembly.
.assembly extern Customer

{
.ver 1:0:592:25677
}
.assembly extern Hotel
{
.ver 1:0:592:25677
}
.assembly extern System.Drawing
{
.publickeytoken = (B0 3F 5F 7F 11 D5 0A 3A)
.ver 1:0:2411:0
}
The System.Drawing assembly is a shared assembly, which can be seen in the
\WINNT\Assembly directory using Windows Explorer. Mscorlib, which is a shared
assembly, is not deployed in the assembly cache. Microsoft made a single exception here:
because mscorlib is so closely tied with the CLR engine (mscorwks [4]), it is installed in
the appropriate install directory (\WINNT\Microsoft.NET\Framework) for the current
.NET version.
[4]
Or mscorsvr.dll for servers.
In the System.Drawing shared assembly, the .publickeytoken = (B0 3F 5F 7F 11 D5 0A

3A) metadata statement provides a public key token, which is the lowest 8 bytes of a hash
of the public key that matches the corresponding private key owned by the
System.Drawing assembly's author. This public key token cannot actually be used
directly to authenticate the identity of the author of System.Drawing. However, the
original public key specified in the System.Drawing manifest can be used to
mathematically verify that the matching private key was actually used to digitally sign
the System.Drawing assembly. Since Microsoft authored System.Drawing.dll, the public
key token seen above is Microsoft specific. Of course, the matching private key is a
closely guarded corporate secret, and it is believed by most security experts that such a
private key is, in practice, virtually impossible to determine from the public key.
However, there is no guarantee that some mathematical genius will not find a back door
someday!
The .publickeytoken declaration

The .publickeytoken declaration provides only the least significant 8 bytes of the
SHA1 hash of the producer's public key (which is 128 bytes), which saves some
space but can still be used to verify at runtime that the assembly being loaded
comes from the same publisher as the one you compiled against. Alternatively,
the .publickey declaration could be used, which provides the full public key. This
would take up more space but makes it harder for villains to find a private key
that matches the full public key.
As we shall see shortly, the .publickeytoken statement is present in the client assembly's
manifest only if the referenced assembly has been digitally signed, and all assemblies
intended for shared deployment must be digitally signed. Microsoft has digitally signed
the standard .NET assemblies, such as mscorlib.dll, and System.Windows.Forms.dll with
private keys belonging to them. This is why the public key token for many of those
shared assemblies, seen in the \WINNT\Assembly directory using Windows Explorer, has
the same value repeated. Assemblies authored and digitally signed by other vendors are
signed with their own distinct private keys, and they will therefore result in a different
public key token in their client assembly's manifests. Later, we will look at how you can
create your own private and public key pair and digitally sign your own assemblies for
deployment into the global assembly cache.
Nonetheless, while unique, none of these digital keys can identify who the author of a
particular module is. A developer of assemblies can use the signcode utility to add a
digital certificate that will identify the publisher of the assembly.
The .ver 1:0:2411:0 metadata statement indicates the version of the System.Drawing
assembly. While these numbers have no intrinsic meaning, the Microsoft suggested
format of this version specification is Major:Minor:Build:Revision. Over time, as new
versions of this assembly are released, existing clients that were built to use this version
will continue using this version, assuming the conventional meaning of major and minor
values. Newer client programs will, of course, be able to access newer versions of this
assembly as they become available. The old and new versions can be deployed side-by-
side in the global assembly cache and be simultaneously available to old and new client
programs.
Note that the version 1:0:2411:0 appearing in the client manifest belongs to the current
version of the Acme.Gui assembly and is unrelated to the "1.0.*" version attribute
specified in the AssemblyInfo.cs file in the AcmeGui source code. We will soon look
more closely at the four fields that make up a version number, and how assembly
versioning works with the suggested format.
Now let us consider the information about the component itself in its manifest.
ILDASM shows the assembly metadata in the Customer manifest:
.assembly Customer
{
.custom instance void
[mscolib]System.Reflection.AssemblyKeyNameAttribute
::.ctor(string) = ( 01 00 00 00 00 )
...
// --- The following custom attribute is added
automatically, do not uncomment -------
// .custom instance void
[mscolib]System.Diagnostics.DebuggableAttribute
::.ctor(bool,
// bool) = ( 01 00 01 01 00 00 )
.hash algorithm 0x00008004
.ver 1:0:592:25677
}
The .assembly Directive

The .assembly directive declares the manifest and specifies to which assembly
the current module belongs. In this example, the .assembly directive specifies the
name of the assembly to be Customer. It is this name (combined with the version
number and optionally a public key) rather than the name of the DLL or EXE file
that is used at runtime to resolve the identity of the assembly. Also note that if the
assembly is signed, you will see the .publickey defined within the .assembly
directive. It also indicates what custom attributes have been added to the
metadata.
The .assembly Customer metadata statement indicates that the assembly name is
Customer. Note that this is not the name of a component class within the assembly, but
rather the assembly itself. This assembly is not digitally signed, and therefore it does not
contain a public key.
In multifile assemblies (discussed in a later section) the manifest stores a hash of each
file. The .hash algorithm 0x00008004 metadata statement indicates that SHA1 is the hash
algorithm that is to produce this hash-code value. Many hash-code algorithms exist.
Initially, however, only MD5 (0x000803) and SHA1 (0x000804) are supported by .NET.
Hash Algorithms
A hash algorithm is a mathematical function that takes the original data of
arbitrary size as input and generates a hash code, also known as a message digest,
which is a fixed-sized binary output. An effective hash function is a one-way
function that is highly collision free, with a result that is relatively small and
fixed in size. Ideally, a hash function is efficient to calculate as well. A one-way
function is a function that has no inverse, so that you cannot effectively
reproduce the original data from the hash-code value. [5] The phrase "highly
collision free" means that the probability that two distinct original input data
samples generate the same hash code is very small, and it is unlikely to calculate
two distinct input data samples that result in the same hash-code value. The well-
known MD5 and SHA1 hash algorithms are considered to be excellent choices
for use in digital signing, and they are both supported by .NET.
[5]
One-way encryption codes are used to store passwords in a passwords database. When
you log in, the password you enter is encrypted and compared with what is stored in the
database. If they match, you can log in. The password cannot be reconstructed from the
encrypted value stored in the passwords database.
Versioning an Assembly
An assembly manifest contains the version of the assembly as well as the version of each
of the assemblies that the assembly depends on. The version number of an assembly is
composed of four numerical fields: Major, Minor, Build, and Revision. There are no
semantics assigned to any of these fields by the CLR. Microsoft does suggest the
following convention:
• Major— a change to this field indicates major incompatible changes.
• Minor— a change to this field indicates minor, but incompatible changes.
• Build number— a change to this field indicates a new backward-compatible
release.
• Revision— a change to this field indicates a backward-compatible emergency bug
fix.
None of this is enforced by the CLR. You enforce this convention, or any other
convention you choose, by testing assemblies for compatibility and specifying the version
policy in a configuration file that we will discuss.
In the metadata for the Customer assembly, the .ver 1:0:592:25677 gives us the
assembly's version: Major Version 1, Minor Version 0, Build Number 592, Revision
25677.
The version information for the manifest can be defined in the source code using the
assembly attribute assembly::AssemblyVersion. This attribute (as with other global
attributes) can appear in a source file after a using statement but before any namespace or
class definitions. The AssemblyVersionAttribute class is defined in the
System::Reflection namespace. If this attribute is not used, a default version number of
0.0.0.0 is listed in the assembly manifest, which is generally not desirable.
In a project created with the VisualStudio.NET project wizard, the source file
AssemblyInfo.cs is automatically generated, with a version of 1.0.*, producing a major
version of 1, and a minor version of 0 and automatically generated build and revision
values. If you change the AssemblyVersionAttribute to, for example, "1.1.0.0", as shown
below, the version number displayed in the manifest will be modified accordingly to
1:1:0:0.
//AssemblyInfo.cs
...
[assembly: AssemblyVersion("1.1.0.0")];
If you specify any version number at all, you must at a minimum specify the major
number. If you specify only the major number, the remaining values will default to zero.
If you also specify the minor value, you can omit the remaining fields, which will then
default to zero, or you can specify an asterisk, which will provide automatically
generated values. The asterisk will cause the build value to equal the number of days
since January 1, 2000, and the revision value will be set to the number of seconds since
midnight, divided by 2. If you specify major, minor, and build values, and specify an
asterisk for the revision value, then only the revision is defaulted to the number of
seconds since midnight, divided by 2. If all four fields are explicitly specified, then all
four values will be reflected in the manifest. The following examples show valid version
specifications.
Specified in source Result in manifest

None 0:0:0:0
1 1:0:0:0
1.1 1:1:0:0
1.1.* 1:1:464:27461
1.1.43 1:1:43:0
1.1.43.* 1:1:43:29832
1.1.43.52 1:1:43:52
If you use the asterisk, then the revision and possibly the build number will automatically
change every time you rebuild the component. You must make an explicit change to the
major and minor numbers if you wish to have their values changed.
Strong Names
Before we can discuss version policy, we have to introduce the idea of a strong name. A
strong name is guaranteed to be globally unique for any version of any assembly. Strong
names are generated by digitally signing the assembly. This ensures that the strong name
not only is unique, but can be generated only by an individual that owns a secret private
key.
A strong name is made up of a simple text name, a public key, and a hash code that has
been encrypted with the matching private key. The hash code is known as a message
digest and the encrypted hash code is known as a digital signature. The digital signature
effectively identifies the assembly's author and ensures that the assembly has not been
altered. Two assemblies that have the same strong name and version are considered to be
identical assemblies. Two assemblies with different strong names are considered to be
different. A strong name is also known as a cryptographically strong name, since, unlike a
simple text name, a strong name is guaranteed to uniquely identify the assembly based on
its contents and its author's private key. A strong name has the following useful
properties:
• A strong name guarantees uniqueness based on encryption technology.

• A strong name establishes a unique namespace based on the use of a private key.
[6]
[6]
Do not confuse this namespace with the one used by the compiler to
disambiguate class names.
• A strong name prevents unauthorized personnel from versioning the assembly.

• A strong name allows the CLR to find the right version of a shared assembly.
Digital Signatures
Digital signatures are based on public key cryptographic techniques. In the world of
cryptography, the two main cryptographic techniques are symmetric ciphers (shared key)
and asymmetric ciphers (public key). Symmetric ciphers use one shared secret key for
encryption as well as decryption. DES, Triple DES, and RC2 are examples of symmetric-
cipher algorithms. Symmetric ciphers can be very efficient and powerful for message
privacy between two trusted cooperating individuals, but they are generally unsuitable for
digital signatures. Digital signatures are not used for privacy but for identification and
authentication. If you shared your symmetric key with everyone who would potentially
want to identify or authenticate you, you would inevitably share it with people who
would want to impersonate you.
Asymmetric ciphers are used in digital signatures. Asymmetric ciphers, also known as
public key ciphers, make use of a public/private key pair. The paired keys are
mathematically related and are generated together. It is, however, exceedingly difficult to
calculate one key from the other. The public key is typically exposed to everyone who
would like to authenticate its owner. On the other hand, the owners keep the matching
private signing key secret, so that no one can impersonate them. RSA is an example of a
public key cipher system.
Public key cryptography is based on a very interesting mathematical scheme that allows
plain text to be encrypted with one key and decrypted only with the matching key. For
example, if a public key is used to encrypt the original data (known as plain text), then
only the matching private key is capable of decrypting it. Not even the encrypting key
can decrypt it! This scenario is useful for sending secret messages to only the individual
who knows the private key.
The opposite scenario is where the individual who owns the private key uses that private
key to encrypt the plain text. The resulting cipher text is by no means a secret, since
everyone who is interested can obtain the public key to decrypt it. This scenario is useless
for secrecy but very effective for authentication purposes. To improve performance,
instead of encrypting the original data, a highly characteristic hash code is encrypted
instead.
If you use the matching public key to decrypt the encrypted hash code, you can
recalculate the hash code on the original data and compare the two values. If they match,
you can be certain that the owner of the private key was the digital signer. Of course, the
owner of the private key has to make sure to keep the private key secret, otherwise you
cannot prove that the data has not been tampered with from the time when it was digitally
signed. Figure 7-4 shows how a digital signature works.
Figure 7-4. How a digital signature works.

SHA1 and RSA
To sign the assembly, the producer calculates a SHA1 hash of the assembly (with
the bytes reserved for the signature preset to zero) and then encrypts the hash
value with a public key using RSA encryption. The public key and the encrypted
hash are then stored in the assembly's metadata.
Digitally Signing an Assembly
The process of digitally signing an assembly involves generating a public/private key

pair, calculating a hash code on the assembly, encrypting the hash code with the private
key, and writing the encrypted hash code along with the public key into the assembly for
all to see. The encrypted hash code and public key together comprise the entire digital
signature. The digital signature is written into a reserved area within the assembly that is
not included in the hash-code calculation. All these steps are performed with two simple
tools— the Strong Name utility (Sn.exe) and the Assembly Linker (Al.exe). To build and
digitally sign an assembly, the following steps are performed.
1. Develop and build the component.

2. Generate a public/private key pair.
3. Calculate a hash code on the contents of the assembly.
4. Encrypt the hash code using the private key.
5. Place the encrypted hash code into the manifest.
6. Place the public key into the manifest.
Step 1 is, of course, usually performed using Visual Studio.NET. Steps 2 through 6 are
known as digital signing. Step 2 is accomplished using the Strong Name utility Sn.exe.
Steps 3 through 6 are accomplished using either Visual Studio.NET or the Assembly
Linking utility Al.exe (that's "A-el", not "A-one").
To illustrate this process we will develop a version of our Customer and Hotel assemblies
that have strong names. They are located in the SignedCaseStudy directory. We generate
key pairs for the assemblies using Sn.exe, known as the Strong Name utility. This tool
generates a cryptographically strong name for the assembly. You generate a public/private
key pair and place them into a file named KeyPair.snk as shown in the following
command (which you can run from the source directory):
sn -k KeyPair.snk
The resulting KeyPair.snk file is a binary file and is not intended to be human readable. If
you are curious, you can write these keys into a comma-delimited text file with the
following command, then view it using Notepad.exe. This is not a required step.
sn -o KeyPair.snk KeyPair.txt
In the example you will finds these files in the Customer and Hotel subdirectories.
The next step is to apply the private key to the assembly. For developing and testing it is
convenient to do this at compilation time. When you release the assembly, however, you
have to use the official private key of the company. For security reasons this key is
probably known only to the corporate digital signing authority. The process of creating
the strong name cannot be postponed until after the assembly is built, because the public
key is part of the assembly's identity. Users of the assembly have to compile against the
full identity of the assembly. Delay signing, which splits the process of assigning the
strong name into two steps, is designed to solve this problem.
If you just want to apply the digital signature automatically at compile time without delay
signing, you simply use the AssemblyKeyFileAttribute— which, in the example, is in the
file AssemblyInfo.cs of the Customer project. The KeyPair.snk file generated previously
with the Sn.exe tool is specified in the attribute. The file path has to be relative to the
project output directory. Once the KeyPair.snk file has been added to the
AssemblyKeyFileAttribute the code must be recompiled.
[assembly: AssemblyKeyFile(".\\Customer\\KeyPair.snk")]
Delay signing requires a more complex procedure. When you build the assembly, the
public key is supplied to the compiler so that it can be put into the PublicKey field in the
assembly's manifest. Space is reserved in the file for the signature, but the signature is not
generated. When the actual signature is generated, it is placed in the file with the -R
option to the Strong Name utility (sn.exe).
To indicate to the compiler that you want to use delay signing, you include
AssemblyDelaySignAttribute in your source code. You also have to include the public
key using the AssemblyKeyFileAttribute.
Assuming you have generated the public/private key pair as described previously, you
then use the -p option of the Strong Name utility to obtain just the public key without
giving out the still secret private key.
sn -p KeyPair.snk PublicKey.snk
You then add the following two attributes to AssemblyInfo.cs:
[assembly: AssemblyDelaySign(true)]
[assembly: AssemblyKeyFile(".\\PublicKey.snk")]
The assembly still does not have a valid signature. You will not be able to install it into
the global assembly cache. You can disable signature verification of a particular assembly
by using the -Vr option on the Strong Name utility.
sn -Vr Customer.dll
Before you ship the assembly you must supply the valid signature. You use the -R option
on the Strong Name utility and supply the public/private key pair.
sn - R customer.dll KeyPair.snk
However you add the key, if you look at the manifest in ILDASM you will see that the
.publickey entry has been added to the assembly's metadata.
The .publickey attribute represents the originator's public key that resides in the
KeyPair.snk file. This is the public key that can be used to decrypt the message digest to
retrieve the original hash code. When the assembly is deployed into the global assembly
cache, this decrypted hash code is compared with a fresh recalculation of the hash code
from the actual assembly contents. This comparison is made to determine if the assembly
is legitimate (i.e., identical to the original) or illegitimate (i.e., corrupt or tampered). Of
course, when you use Sn.exe, it will produce a different key pair, and the public key
shown below will be different in your case accordingly.
If you use ILDASM to examine the manifest of the AcmeGui client program, you will
see the following:

{
.publickeytoken = (8B 0E 61 2D 60 BD E0 CA )
.ver 1:0:0:0
}
{
.publickeytoken = (CF 0B C2 2F 8E 2C 15 22 )
.ver 1:0:0:0
}
Now that Customer and Hotel have strong names, references to them have a public key
token, which is a hash of the public key that matches the corresponding private key for
the assembly. Note that we generated different keys for each assembly. Usually, each
company will use the same key pair for all its public components.
Now that we have discussed strong names, we can discuss the two methods of deploying
assemblies in .NET, and their associated default version policies. After this discussion we
will show how the default policy can be overridden in a configuration file.
Private Assembly Deployment
For private assembly deployment, the assembly is copied to the same directory as the
client program that references it. No registration is needed, and no fancy installation
program is required. When the component is removed, no registry cleanup is needed, and
no uninstall program is required. Just delete it from the hard drive. [7]
[7]
Of course this process does not put any icons on the desktop or entries on the Start
menu.
Of course, no self-respecting programmer would ever provide a commercial component

that required the end user to manually copy or delete any files in this way, even if it is
remarkably simple to do. Users have become accustomed to using a formal installation
program, so it should be provided, even if its work is trivial. However, for testing
purposes, manually copying and deleting an assembly is an ideal way to quickly and
painlessly manage deployment issues for developing, debugging, or testing purposes.
Recall that the deployment of COM components was never this simple, requiring at a
minimum, a registry script file. Gone are the days where you have to configure the
registry on installation, and then later carefully clean out the registry information when
you want to discard the component.
To privately deploy our componentized Hotel Administrator case study, create a directory
on your hard drive. Copy to that directory the files in the CaseStudy\bin\Debug directory,
AcmeGui.exe, Customer.dll, and Hotel.dll. Then run AcmeGui.exe. It will run. It is really
just that simple!
If you view the AcmeGui manifest in ILDASM, you will see the following dependency
entries:

{
.ver 1:0:593:19533
}
{
.ver 1:0:593:19532
}
Here are the corresponding assembly definitions in the components:
.assembly Customer
{
...
.ver 1:0:593:19533
}
.assembly Hotel
{
...
.ver 1:0:593:19532
}
From this you can see that the client program was built with Customer assembly version
1.0.593.19533 and Hotel assembly version 1.0.593.19532. Since neither assembly has a
strong name, however, the versions are not checked. If you were to build a Customer
assembly with a different version, and replace the one that AcmeGui was built with,
AcmeGui would still run. It does not matter whether you change the major build number
or the revision number.
If you were to use a version of the Customer component with a strong name (even if it
had the same version number), you would get the following runtime exception:
System.IO.FileLoadException: The located assembly's mani

fest definition with name 'Customer' does not match the
assembly reference.
If the Customer assembly has a strong name, even if the version numbers are the same,
the assembly names no longer match. If the AcmeGui client program was built with an
assembly that had a strong name, the CLR will bind only to an assembly that matches
exactly with the strong name and version. Even a different revision number will cause the
load to fail.
The details on binding failures can be seen in the Assembly Binding Log Viewer
(FUSLOGVW.exe). The sample log in Figure 7-5 resulted from an attempt to resolve
AcmeGui's reference to a Customer assembly that had a strong name when it was built
with a version of the assembly that did not have a strong name:
Figure 7-5. Assembly binding log for customer load failure.
Shared Assembly Deployment
The Assembly Cache is a known directory where the CLR looks for shared, side-by-side
assemblies. The term "side-by-side" means that multiple versions of the same component
may reside within the assembly cache alongside one another. The global assembly cache
contains shared assemblies that are globally accessible to all .NET applications on the
machine. A download assembly cache is accessible to applications such as Internet
Explorer that automatically download assemblies over the network.
Deploying a Shared Assembly
Before an assembly can be deployed into the global assembly cache, you must digitally
sign the assembly as discussed earlier. Developers can place the assembly into the global
assembly cache by using either using the Global Assembly Cache Utility Gacutil.exe
command-line utility, the Windows Explorer with the assembly cache viewer Windows
shell extension, or the .NET Admin Tool. Deploying shared assemblies on a user's
machine should be done with an installation program.
To illustrate this process we will deploy in the GAC the version of our Customer and
Hotel assemblies that are in the SignedCaseStudy directory. To deploy the components
into the GAC, you can use the command-line utility, Gacutil.exe.
Gacutil -i Customer.dll
Note that the -i option is case sensitive. You should then see the console message
"Assembly successfully added to the cache." The effect of this command is that a new
global assembly cache node named Customer is created in the \WINNT\Assembly
directory. As can be seen in Figure 7-6, the version number and originator (i.e., public
key token) are displayed for the assembly in Windows Explorer. We changed the version
of the component to 1.0.0.0 to distinguish it from the unsigned version.
Figure 7-6. Windows Explorer showing the global assembly cache.

You can also can drag and drop a component into the Assembly directory to install it in
the GAC. Alternatively, you can use the .NET Admin Tool to install an assembly into the
GAC. The .NET Admin Tool is an MMC snap-in located at
\WINNT\Microsoft.NET\Framework\v1.0.2914\ mscorcfg.msc. [8] The directory version
number will be different in a later release of the .NET Framework. While it may seem
overkill to introduce a third tool, this MMC snap-in is a very useful utility that simplifies
many tasks. Figure 7-7 shows the top-level window of this tool. To use the tool to add an
assembly to the GAC, just select Assembly Cache in the left pane, right-mouse click, and
select Add. Using the dialog box that pops up to navigate to the file, select the assembly
you want to add, and click the Open button.
[8]
To run a snap-in, you can just double-click on the .msc file in Windows Explorer. Since
we are going to use the .NET Admin Tool extensively, you may wish to add the tool to
the Visual Studio Tools menu, which you can do through Tools | External Tools... . For
the command enter mmc.exe and for the argument enter the complete path to
mscorcfg.msc.
Figure 7-7. .NET Admin Tool supports many .NET administrative functions.
After you have installed the assemblies in the GAC, copy just the AcmeGui client
program in the SignedCaseStudy directory to another directory. You can now run it
without any assemblies in the same directory.
What happens if we remove the version of Customer we installed in the GAC and place
in the GAC a Customer assembly signed with the same key, but a different version? A
FileNotfoundException is thrown by the CLR. We would get the same result if we
replaced it with a Customer assembly that had the same version, but signed with a
different key. The default binding policy for shared assemblies is an exact name match.
Versioning Shared Components
What happens if you install two versions of the same assembly in the GAC that were
signed with the same key? Place a Customer assembly with the version 1.1.0.0 in the
GAC. Figure 7-8 displays both versions of the Customer assembly installed in the Global
Assembly Cache with their respective version numbers and identical public key tokens.
Figure 7-8. .NET Admin Tool with side-by-side components in the global
assembly cache.
This is called by-side deployment. Both assemblies are available to client programs that
require them. Programs can bind to either of them without fear of getting the wrong
version.
Assembly Configuration
The CLR binds to an assembly when either a static or dynamic reference is made to it at
runtime. A static reference is defined permanently in the client assembly manifest when it
is compiled. A dynamic reference is produced programmatically at runtime, for example,
by calling the method System.Reflection.Assembly.Load.
You can use a strongly named assembly to force a client to bind to a specific version of
an assembly whether you have private or shared deployment. Suppose you want to allow
several backward-compatible assemblies to match? You can use XML configuration file
to specify some rules for the CLR to use when it tries to find an assembly that matches.
The .NET Admin Tool can be used to create and maintain these files through a graphical
interface.
The name of the configuration file client program's name is appended with a .config
extension. For our AcmeGui client the configuration file would be named
AcmeGui.exe.config. It is placed in the same directory as the client executable.
In addition to an application configuration file, there is an administration configuration

file called Machine.config. It is found in the Config subdirectory under the directory
where the .NET runtime is installed (\WINNT\Microsoft.NET\Framework\v1.0.2914\,
where the version number reflects the current build of .NET). An administration version
policy is defined with the same XML tags that an application configuration file uses.
However, the administrator configuration file overrides any settings in the application
configuration file.
Resolving an Assembly Reference at Runtime
If the reference has a strong name, the configuration files are examined first to determine
the correct assembly version(s) required. If the reference does not have a strong name,
any version will satisfy the reference. [9] If the assembly reference has been previously
resolved, that previously loaded assembly is used. The assembly cache is checked next
and, if the assembly is found there, that assembly is loaded. If the assembly is not found
in the assembly cache, the CLR probes for the assembly. We will discuss probing after we
discuss specifying version policy in the configuration files.
[9]
There is also a publishers configuration file that we do not discuss. If you are using
Internet Explorer, the configuration files might have to be downloaded from another
computer.
Specifying the Version Policy in a Configuration File
The <configuration> is the top-level tag for .NET configuration files. Assembly binding
information is found in the <runtime> section. A sample AcmeGui.exe.config file might
look like this:
<?xml version="1.0"?>
<configuration>
<runtime>
<assemblyBinding xmlns="urn:schemas-microsoft
com:asm.v1">
<dependentAssembly>
<assemblyIdentity name="Customer"
publicKeyToken="8b0e612d60bde0ca" />
<bindingRedirect oldVersion="1.0.0.0-1.1.0.0"
newVersion="1.1.0.0" />
</dependentAssembly>
</assemblyBinding>
</runtime>
</configuration>
Rules defining version policy are found in the <assemblyBinding> section. The XML
namespace specification is required. Each assembly whose version policy we want to set
is placed in its own <dependentAssembly> section. The assemblyIdentity element has
attributes that define the assembly this section refers to. The name attribute is required;
the publicKeyToken and culture attributes are optional. [10] The bindingRedirect element's
attributes define what versions can map to another version. The oldVersion attribute can
be a range, the newVersion attribute can be set only to one version. In the above example,
any references to versions 1.0.0.0 to 1.1.0.0 can be resolved by using version 1.1.0.0. In
other words, 1.1.0.0 is backward compatible with all those versions. You can specify
several bindingRedirect elements.
[10]
You may ask: Why is the publicKeyToken optional? After all, there is no version
resolution without it. As we shall see shortly, other policies can be defined that do not
require a public key.
You can use the .NET Admin Tool to specify this. To add an application to the tool first
select Applications in the left pane. Right-mouse click and select Add from the context
menu. Navigate to the application you want to configure. Select it and click the open
button. Figure 7-9 shows the AcmeGui application added to the admin tool.
Figure 7-9. AcmeGui added to the .NET admin tool.
To configure the Customer assembly, select Configured Assemblies in the left pane, right-
mouse click, and select Add from the context menu. In the dialog box that comes up,
select the radio button that has the text "Choose an assembly from the list of assemblies
this application uses." Then click the "Choose Assembly" button. Select Customer from
the list that pops up, and then click the Select button. The Assembly information for the
Customer assembly should be entered in the "Configure an Assembly" dialog. Click the
Finish button on that dialog. Select the "Binding Policy" tab. Figure 7-10 shows what you
should see after the binding policy that was in the sample configuration file was recorded.
Figure 7-10. Binding policy set for the Customer assembly.
After you select OK, you can navigate to the directory where the AcmeGui executable is,
and you will see a configuration file that the tool has created for you. It should resemble
our previous example.
Finding the Assembly's Physical Location
At this point the CLR knows what versions of the assembly will satisfy the reference. The
CLR does not yet know where the assembly resides on disk. If the assembly with the
right version has been previously loaded because of another reference to that assembly
earlier in the program, that assembly is used. If the assembly has a strong name, the
assembly cache is checked; if the correct version is found there, that assembly is used.
There are several elements you can specify in the configuration file to tell the CLR where
to try and find the assembly.
If the assembly has not yet been found, the runtime checks to see if a codebase has been
specified in the configuration file. Under the <dependentAssembly> section you can
specify a <codeBase> element. This element has two attributes, a version and a URI, to
check for the assembly. The Codebases tab on the .NET Admin Tool's assembly
properties dialog can be used to set them in the configuration file. Examples of this
element are:
<codeBase version="1.1.1.1"
href="http://www.abc.com/Customer.dll" />
<codeBase version="1.1.1.2"
href="file:///c:\AcmeGui\Customer.dll" />
To use a Codebase element outside the application's directory or subdirectories, a strong

name is required. At this point, whether or not the required assembly is found, the binding
process stops. If the assembly is not found, an exception is generated at this point.
If a CodeBase element was not found in the configuration file, the runtime continues to
probe for the assembly. At this point all searching is relative to the directory in which the
application runs, which is referred to as the application base.
The runtime first looks in the application base. It then looks in any subdirectories of the
application base that have the same name as the assembly. If a culture is specified in the
request, the runtime only looks for the assembly subdirectory under a subdirectory with
the name of the culture requested.
Finally, you can specify in the assemblyBinding section of the configuration file a
privatePath, which is a semicolon-delimited list of subdirectories of the application base
to look in.
<probing privatePath="\bin;\assemb" />
You can also set the privatePath on the properties tab for the application in the .NET
Admin Tool.
Assembly Location and Visual Studio.NET
Building an assembly within Visual Studio.NET requires a reference to a specific

assembly at a specific disk location. The rules just described apply when the application
is run, not built.
Within VS.NET you cannot browse to the GAC (\Winnt\Assembly) and add a reference.
The referenced component must be located somewhere else on disk. One of the properties
of a referenced component is the CopyLocal property. If set to true, the referenced
component is copied to the local project directory. While that copy would be used for the
compilation reference, whether it is the one linked to depends on the configuration file
settings.
Multimodule Assemblies
An assembly can be made up of multiple modules. A module is a DLL (or EXE) that
contains managed code plus metadata, but not necessarily a manifest. However, an
assembly must have one and only one manifest. Hence an assembly can contain multiple
modules, but only one of them can have a manifest that provides information on the
contents of all the modules in the assembly. The module with the manifest may have just
the manifest, or it can contain other code or resources.
The main advantage of breaking an assembly into multiple modules is that each module
is contained in a separate DLL file. This allows Web downloads to be performed on
demand, on a per-module basis. This can improve performance and memory
consumption. Even in a local scenario, the CLR loads classes on the local machine with
module granularity, which can improve efficiency. Another reason for constructing an
assembly with multiple modules is that you may have written each part of an assembly in
a different .NET language. To build an assembly that contains multiple modules, you
need to build each module separately, and then combine them with the Al.exe utility.
There are two ways to go about creating a multimodule assembly. One way is to create all
the modules without any manifest, and then create one additional module that contains
only a manifest for the entire assembly, but no actual code. The other technique is to have
just one module in the assembly that contains both code and a manifest for the entire
assembly, and to have all other modules in the assembly contain only code, with no
manifest. We will describe the first alternative, since it is more symmetric and easier to
visualize. The second alternative is not described here, however, it is done in a similar
way, with the same tools.
Visual Studio.NET does not allow you to do this for C# projects. The MultiModule
example illustrates the mechanics of how to create a multiple-module assembly. The
example directory contains three files. Add.cs and Sub.cs will be built into separate
modules and then combined together in an assembly. Compute.cs uses this assembly.
Add.cs has one class with one method; Add. Sub.cs has another class that has one
method, Sub.
public class MyCalc

{
public int Add(int a, int b)
{
return a + b;
}
}
public class MyCalcSub
{
public int Sub(int a, int b)
{
return a - b;
}
}
We create two modules with no assembly manifest by running build.bat, which has two
commands:
csc /target:module /out:add.dll add.cs

csc /target:module /out:sub.dll sub.cs
If you look at add.dll in ILDASM, you will see that there is a .module add.dll statement
but no .assembly statement. We now can build an assembly with a manifest using the
Assembly Linker tool Al.exe by running link.bat, which has one command:
Al add.dll, sub.dll /out:arith.dll
As Figure 7-11 shows, arith.dll contains only a manifest. The manifest shows that the
assembly is made up of two separate, distinct files, and the types in those files are listed
in the manifest.
Figure 7-11. Manifest for a multimodule assembly.
We have a simple client program, compute.cs, that uses the types in arith.dll.
public class Compute

{
{
MyCalc x = new MyCalc();
int y = x.Add(1, 3);
Console.WriteLine("y = " + y.ToString());
MyCalcSub z = new MyCalcSub();

y = z.Sub(1, 3);
Console.WriteLine("y = " + y.ToString());
return;
}
}
We can build it with the command:
csc /r:arith.dll compute.cs
This will produce compute.exe, which we can run.
Setup and Deployment Projects
Assemblies may be deployed as regular stand-alone binary code files (i.e., DLL or EXE
files), or they may be deployed using CAB, MSI, or MSM files. A CAB file is a cabinet
file with the .cab file-name extension. A CAB file is used to compress and combine other
files into one convenient manageable file. Although CAB files can be used for general
purposes, they have traditionally been used for CD-based and Web-based installation
purposes. MSI and MSM files are Microsoft Windows Installer files, with the .msi and
.msm file-name extensions. MSI files (and indirectly, MSM files) are used with the
Msiexec.exe Windows Installer program to deploy stand-alone applications and reusable
components.
MSI files are Microsoft Windows Installer installation packages that have the .msi file-
name extension. MSM files are merge modules that have the .msm file-name extension.
Windows Installer supports software installation, repair, upgrade, and removal. Windows
Installer packages are self-contained database files that provide installation information to
the Windows Installer service. An MSM file has an internal structure similar to that of an
MSI file, but it is somewhat simplified. Unfortunately, an MSM file cannot be used
directly by Windows Installer, since it lacks certain important database tables. Instead, the
MSM file must be merged into an MSI file to be used in an actual installation session.
However, MSM files are useful for separating out shared installation information into an
independent package that can then be merged into many other MSI packages.
Installation may be accomplished using the Windows Installer, Internet Explorer, or

simply by manually copying assemblies and associated files. To help the programmer
develop setup and deployment solutions, Visual Studio.NET provides several templates
and wizards for generating starter setup projects. These tools are available by way of the
New Project dialog box under the Setup and Deployment Projects node, as shown in
Figure 7-12. As you can see, the following templates are provided for generating starter
setup and deployment projects.
Figure 7-12. Setup and deployment project templates.

• Cab Project
• Setup Project
• Setup Wizard
• Merge Module Project
• Web Setup Project
CAB Project
A CAB project creates a cabinet file containing any number of other files that can be used
for traditional deployment purposes. CAB files have been used to package legacy COM
components deployed over the Internet, and they have also been used in traditional CD-
based installation programs. CAB files may now also be used for packaging managed
code. However, for .NET deployment, a CAB file can contain only one assembly, and the
CAB file must be assigned the same name as the contained assembly, but with the .cab
extension. For example, an assembly named SomeComponent.dll would have to be
contained in a cabinet file named SomeComponent.cab.
Setup Project
The Setup project template creates a Windows Installer .msi file for a desktop or
distributed application. A Setup project is not intended for deployment of Web-based
applications, since a specialized Web Setup project is used for that purpose. A Setup
project produces a program that installs an application onto a target machine. You may
create Setup projects within the same solution that contains the other projects to be
deployed. In a multitier solution, you can create one setup project for each project that is
to be deployed to a particular target computer. For example, in a simple three-tier
solution, you would probably have three deployment projects. Two simple deployment
projects would set up the client and server. A third deployment project would then look
after the more complex middle-tier business logic. Additional deployment projects may
come into play if the solution was highly complex, or if merge modules were
incorporated into the deployment strategy.
To create a Setup project, select File | New, then select Project. In the New Project dialog
box, select Setup and Deployment Projects as the Project Type. Finally, select Setup
Project as the Template, specify name and location, and then click OK. The result of this
is shown in Figure 7-13, showing Solution Explorer and the File System Editor.
Figure 7-13. Solution Explorer and the File System Editor in a Setup project.
Once the initial Setup project has been created with the Setup Project template, it can be
further developed using the File System Editor. The File System Editor allows you to
drag and drop, or copy and paste, files to be deployed by the Setup deployment project,
and control their destinations on the target machine. Initially, the File System Editor
shows an initial list of destination folders that you can deploy into, and you can also add
your own folders to this list.
Once you have an MSI file created, you can use Windows Installer, as shown in the
following command line:
Msiexec /i SomeSetup.msi
The Windows Installer program then starts up and displays a series of installation dialogs.
After Windows Installer has completed the deployment, you can try running the installed
application to verify that the installation was successful. If you run the same command
Msiexec /i SomeSetup.msi, it will detect that it already exists, so it gives you the choice
to either repair the installation or uninstall it.
Merge Module Project
A Merge Module project packages reusable setup information that can be independently
maintained and then merged as a shared installation package into other installation
packages. A Merge Module project produces a merge module .msm file that can be
merged into .msi files. This allows you to share common assemblies, associated files,
registry values, and setup functionality among multiple applications.
To start the Merge Module Project Wizard, select File | New, then select Project. In the
New Project dialog box, select Setup and Deployment Projects as the Project Type.
Finally, select Merge Module Project Wizard as the Template, specify name and location,
and then click OK.
Typically, an MSI file is intended for use by the end user for installing a complete
solution in one simple deployment session. In contrast, an MSM file is typically intended
for use by other developers who want to use components that you have developed in their
setup projects. Those other developers can merge your MSM file into their own MSI file
for deploying your components into their test and development environments, as well as
for their ultimate end user. End users should not be provided any MSM files, since they
are not directly installable using Windows Installer, and they are not very friendly to
work with.
To add an existing merge module project to a Setup project, create or open the Setup
project, select File | Add Project, and then select Existing Project. In the Add Existing
Project dialog box, browse to the location of the desired merge module project, select the
associated .vdp deployment project file, and then click Open.
We have just added the merge module project to the solution. We now have to add it to
the Setup project itself. Select the Setup project and invoke "Add:Project Output," then
select the merge module project in the dialog that appears.
Web-based Deployment
Web-based deployment uses Internet Explorer on the client to automatically download

assemblies packaged as EXE, DLL, or CAB files on demand from a .NET Web server.
HTML files can dynamically deploy assemblies as well as configuration files to control
the binding process. Web-based deployment results in assemblies being downloaded into
the client's assembly download cache on demand.
The Xml <object> tag is used to download and install assemblies, using either a relative
or an absolute URL. The following shows a relative URL example, where the assembly is
located relative to the directory of the containing HTML file on the Web server.
<object
id="SomeComponent"
classid="./SomeDirectory/MyComponent.dll#SomeClass">
</object>
The following shows an absolute URL example, where the assembly is located on a
specified Web server.
<object
id="SomeComponent"
classid="http://www.acme.com/MyComponent.dll#SomeClass">
</object>
By default, IE creates a separate application domain for each Web site that it encounters.
An application domain is a .NET feature that can be thought of as a scalable lightweight
process. An application domain efficiently provides fault isolation without the overhead
of running multiple processes. Each application domain may optionally have its own
configuration file to control binding and security. Also, a configuration file may specify
an isolated application domain for individual applications on the same Web server. Each
HTML file that specifies the same configuration file will be placed into the same
application domain. Application domains are discussed in the next chapter.
Summary
Deployment is very important, constituting one of the major phases in software

development. If it is not done properly, the entire development effort becomes a waste of
time and money. Fortunately, with assemblies, which are the fundamental units of .NET
deployment, many of the complexities and problems relating to deployment are
conveniently solved.
By simply building dynamic link libraries you can have reusable components without all
the difficulties associated with building and installing COM components.
Private assembly deployment can be simply copying the assembly to the same directory
as the client application. Public assembly deployment involves the creation of a strong
name for the assembly, followed by deployment into the global assembly cache. The
Strong Name Utility (Sn.exe) can be used to create the strong name for an assembly. The
Global Assembly Cache utility Gacutil.exe, or the .NET Admin Tool can then be used to
deploy the shared assembly into the global assembly cache. It is also possible to combine
multiple modules into a single assembly. The Visual Studio.NET CAB and setup wizards
are very useful for creating starter setup and deployment projects, which can save a great
deal of development time.
Chapter 8. .NET Framework Classes Top
It is impossible to cover in one chapter or one book all of the .NET Framework classes.
The .NET classes cover a large fraction of the Win32 API, as well as much else. While a
lot of attention has been focused on changes in Internet-related functionality, the
development model for Windows applications has changed as well.
This chapter focuses on classes illustrating the key concepts and patterns that appear
throughout the .NET Framework. Over the long run, experienced programmers will find
this approach more fruitful than attempting to explain a little about every class that you
might need without giving you much insight. Other chapters go into more depth about
other parts of the Framework such as Windows Forms, ASP.NET, ADO.NET, security,
and Web Services.
We start out by exploring the concept of reflection and metadata. Metadata appears
everywhere in .NET and is critical to understanding how the CLR can provide services
for your applications. Next we explore file input/output—for several reasons. First, it
introduces the important topic of serialization. Second, the Path class exemplifies how
some Framework classes provide some or all of their functionality through static
methods. Third, the formatter classes are used in several places in .NET.
Understanding serialization will give you a concrete idea of how the Framework can
handle objects transparently for you. Serialization also appears in a supporting role
wherever objects have to be moved or transported. Our discussion of the ISerializable
interface demonstrates how much easier it is to implement an interface in .NET than with
COM.
To develop an understanding of the .NET model for applications, we introduce

programming with threads under .NET and several .NET synchronization techniques to
handle multithreading issues. The various synchronization techniques illustrate the trade-
offs of using attributes supplied by the Framework versus doing it yourself.
To further your understanding of the .NET programming model, we introduce context

and the use of proxies and stubs. Application domains can achieve application isolation
with less performance penalty than a Win32 processes. [1]
[1]
Win32 process isolation uses the processor's MMU. .NET's application domain
isolation is done in software and requires verifiable code.
The asynchronous design pattern appears throughout .NET and is discussed in some
detail. We give some examples of remoting because it is a key technology and it
summarizes many of the concepts developed in this chapter. The chapter uses several
attributes provided by the .NET Framework, and we show how to implement and use
custom attributes. We discuss garbage collection, finalization, and the dispose pattern, so
that you can understand how to make sure resources are properly freed in your
applications.
Metadata and Reflection
The Serialization example in Chapter 2 demonstrates how metadata makes possible many
of the services of the Common Language Runtime. Many of the technologies we cover in
the rest of the book rely on metadata, although we will not always stop and point this out.
Metadata is information about the assemblies, modules and types that constitute .NET
programs. If you have ever had to create IDL to generate a type library so that your C++
COM objects could be called by Visual Basic, or to create proxies and stubs, you will
appreciate how useful metadata is and will be grateful that it comes "for free."
Compilers emit metadata, and the CLR, the .NET Framework, or your own programs can
use it. Since we want to give you an understanding of how metadata works, we will focus
on the use, not the creation, of metadata. Metadata is read using classes in the
System.Reflection namespace. [2]
[2]
There is a lower-level set of unmanaged COM interfaces for accessing metadata but we
will not discuss them here. See "Metadata in .NET" by Matt Pietrek in the October 2000
MSDN Magazine.
When you load an assembly and its associated modules and types, the metadata is loaded
along with it. You can then query the assembly to get those associated types. You can also
call GetType on any CLR type and get its metadata. GetType is a method on
System.Object from which every CLR type inherits. After you get the Type associated
with an object, you can use the reflection methods to get the related metadata.
The Reflection sample program takes the case study's Customer assembly and prints out
some of the metadata available. You should examine the output and source code as you
read the next sections. You should especially compare the output of the program with the
source code in the file customer.cs.
The program clearly shows that it is possible to retrieve all of the types in an assembly
and reconstruct the structures, interfaces, properties, events, and methods associated with
those types.
First we load the assembly into memory and write out its name.
Assembly a = Assembly.Load(assemblyName);
Console.WriteLine("Assembly {0} found.", a.FullName);
The output for this statement is appropriate for an unsigned assembly:
Assembly Customer, Version=1.0.583.29038, Culture=neutral,

PublicKeyToken=null found.
One of the properties of the Assembly class is the CodeBase, discussed in Chapter 7 on
Deployment. The security Evidence associated with this assembly is another property.
The Evidence class is discussed in Chapter 12 on Security.
The following code tries to get the entry point for the assembly:
MethodInfo entryMethodInfo = a.EntryPoint;
Since this is a dynamic linked library (DLL), there is no entry. If it were an executable
program we could use the Invoke method on the MethodInfo class to run the startup code
in the assembly. [3]
[3]
You can also load and execute the assembly from the AppDomain, as we discuss later
in this chapter.
The sample uses the Assembly's GetModules method to find the modules associated with
this assembly. In this case we have only one, "customer.dll." We could next find the types
associated with the module by using the GetTypes method on each Module instance
returned by GetModules. Since there is only one module, we use the Assembly's
GetTypes method to return an array of the assembly's types. Even if we had several
modules, we would use Assembly.GetTypes if we did not care about the association of
types and modules.
Type
The abstract class Type in the System namespace defines .NET types. Since there are no
functions outside of classes or global variables in C#, [4] getting all the types in an
assembly will allow us to get all the metadata about the code in that assembly. Type
represents all the types present in .NET: classes, interfaces, values, arrays, and
enumerations.
[4]
Although they are permitted by the CTS and are legal in managed C++.
The Type class is also returned by the GetType method on the System.Object class and
the static GetType method on the Type class itself. The latter method can be used with
types that can be resolved statically.
One of Type's properties is the Assembly to which it belongs. You can get all the types in
the containing assembly once you have the Type of one object. Type is an abstract class;
at runtime an instance of System.RuntimeType is returned.
If you examine the program's output you will see that each type in the assembly,
CustomerListItem, ICustomer, Customer, Customers is found and its metadata is printed
out. We can find out the standard attributes and the type from which the class derives for
each type through the Attributes and BaseType properties.
The methods associated with the Type class enable you to get the associated fields,
properties, interfaces, events, and methods. For example, the Customer type has no
interfaces, properties, or events, four fields, three constructors, and the methods inherited
from its BaseType System.Object:
Interfaces:
Fields:
CustomerId
FirstName
LastName
EmailAddress
Properties:
Events:
Constructors:
public .ctor(System.String first, System.String
last, System.String email)
public .ctor()
public .ctor(System.Int32 id)
Methods:
public Int32 GetHashCode()
public Boolean Equals(System.Object obj)
public String ToString()
public Type GetType()
The type Customers inherits from one interface and has one constructor and four of its
own methods in addition to the four it inherited from its BaseType System.Object:
Interfaces:
ICustomer
Fields:
Properties:
Events:
Constructors:
public .ctor()
Methods:
public Void ChangeEmailAddress(System.Int32 id,
System.String emailAddress)
public ArrayList GetCustomer(System.Int32 id)
public Void UnregisterCustomer(System.Int32 id)
public Int32 RegisterCustomer(System.String
firstName, System.String lastName, System.String
emailAddress)
These were obtained with the GetInterfaces, GetFields, GetProperties, GetEvents

GetConstructors, and GetMethods methods on the Type class. Since an interface is a type,
GetInterfaces returns an array of Types representing the interfaces inherited or
implemented by the Type queried. Since fields, properties, events, and methods are not
types, their accessor methods do not return Types. Each of their accessor methods returns
an appropriate class: FieldInfo, PropertyInfo, EventInfo, ConstructorInfo, and
MethodInfo. All these classes, as well as the Type class, inherit from the MemberInfo
class that is the abstract base class for member metadata.
Let us examine some of the metadata associated with a class method. Using the reflection
methods, we were able to reconstruct the signatures for all the classes and interfaces in
the Customer assembly. Here is the output for the methods of the Customers class:
public Void ChangeEmailAddress(System.Int32 id,

System.String emailAddress)
public ArrayList GetCustomer(System.Int32 id)
public Void UnregisterCustomer(System.Int32 id)
public Int32 RegisterCustomer(System.String firstName,
System.String lastName, System.String emailAddress)
Here is the code from the example that produced the output:
for (int j = 0; j < methodInfo.Length; j++)

{
if (methodInfo[j].IsStatic)
Console.Write(" static ");
if (methodInfo[j].IsPublic)
Console.Write(" public ");
if (methodInfo[j].IsFamily)
Console.Write(" protected ");
if (methodInfo[j].IsAssembly)
Console.Write(" internal ");
if (methodInfo[j].IsPrivate)
Console.Write(" private ");
Console.Write("{0} ", methodInfo[j].ReturnType.Name);
Console.Write("{0}(", methodInfo[j].Name);
ParameterInfo[] paramInfo =
methodInfo[j].GetParameters();
long last = paramInfo.Length - 1;
for (int k = 0; k < paramInfo.Length; k++)
{
Console.Write("{0} {1}", paramInfo[k].ParameterType,
paramInfo[k].Name);
if (k != last)
Console.Write(", ");
}
Console.WriteLine(")");
}
Except for the fact that a constructor does not have a return type, the exact same code
reconstitutes the calling sequences for the class's constructors.
The MethodInfo class has properties that help us determine if the method is static, public,
protected, internal, or private as well as the return type and method name. The method
parameters are stored in a property array of type ParameterInfo.
This example should also make clear that types are assembly relative. The same type
name and layout in two different assemblies is treated by the runtime as two separate
types. When versioning assemblies, one has to be careful when mixing versioned types,
or the same types in two different assemblies.
All this metadata allows the Common Language Runtime and the Framework to provide
services to your applications because it can understand the structure of your types.
Late Binding
Reflection can also be used to implement late binding. In late binding the method to be
called is determined during execution rather than compilation. It is one example of how
metadata can be used to provide functionality. As the previous example demonstrates,
you can extract the signature of a method associated with a type. The MethodInfo object
has all the needed metadata for a class method. The DynamicInvocation sample
demonstrates a very simple example of late binding.
We dynamically load an assembly and get the metadata for a method of a particular type:
Assembly a = Assembly.Load("Customer");
Type t = a.GetType("OI.NetCs.Acme.Customers");
MethodInfo mi = t.GetMethod("GetCustomer");
Using the reflection classes, we could have made this completely dynamic by arbitrarily
picking types, methods, and constructors from the Customer assembly using the
techniques of the last example, but we wanted to keep the DynamicInvocation example
simple.
The System namespace has an Activator class that has overloaded CreateInstance
methods to create an instance of any .NET type using the appropriate constructor. The
Activator class is discussed in this chapter's section on Remoting. We invoke a
constructor with no arguments to create an instance of the Customers object.
Object customerInstance = Activator.CreateInstance(t);
We then build an argument list and use the Invoke method of the MethodInfo instance to
call the GetCustomer method.
object[] arguments = new Object[1];

int customerId = -1;
arguments[0] = customerId;
object returnType = mi.Invoke(customerInstance,
arguments);
Using the reflection methods, we get the type information for each field in a return
structure. Note the GetValue method that gets the data for a particular field in a structure.
This is necessary because we cannot do pointer arithmetic to access an offset into a
structure.
if (returnType.GetType() ==
Type.GetType("System.Collections.ArrayList"))
{
System.Collections.ArrayList arrayList =
(System.Collections.ArrayList)returnType;
for (int i = 0; i < arrayList.Count; i++)
{
Type itemType = arrayList[i].GetType();
FieldInfo[] fi = itemType.GetFields();
for (int j = 0; j < fi.Length; j++)
{
object fieldValue = fi[j].GetValue(arrayList[i]);
Console.Write("{0, -10} = {1, -15}", fi[j].Name,
fieldValue);
}
}
}
This code did not use any specific objects or types from the Customer assembly. We did
use some knowledge about the assembly to keep the code simple in order to illustrate the
main points. It should be clear, however, how to make this completely general.
You can go one step further and use the classes that emit metadata (in
System.Reflection.Emit). You can dynamically create an assembly and then load and run
it.
Input and Output in .NET
To make a crude generalization, the input/output functions in the .NET Framework can be
divided into two broad categories, irrespective of the data storage (disk, memory, and so
on) that is being written to, or read from.
Data can be treated as a stream of bytes or characters. We can read a block of bytes from
a file to a memory buffer. Data can also be treated as a set of objects. Reading and writing
the objects is referred to as deserializing and serializing the objects. We can serialize
(write) the list of Customer objects to disk. We can then deserialize (read) the list of
Customer objects back into memory.
The System.IO namespace has several classes for reading and writing to various types of
storage while treating the data as bytes or characters. Serialization functionality can be
found in various places in the .NET framework. The System.Runtime.Serialization
namespace handles serialization of the Common Type System. The
System.Xml.Serialization namespace handles XML serialization.
Streams
Stream is an abstract class that is the basis for reading from and writing bytes to some
storage such as a file. It supports both synchronous and asynchronous reading and
writing. Asynchronous methods are discussed later in this chapter. The Stream class has
the typical methods that you would expect: Read, Write, Seek, Flush, and Close.
The FileStream class is derived from Stream to represent the reading and writing of files
as a series of bytes. The FileStream constructor builds the actual stream instance. The
overridden Stream methods implement the reading and writing to the file.
Other classes derived from Stream include MemoryStream, BufferedStream, and

NetworkStream (in System.Net.Sockets).
The FileStream example (in the FileIO directory, as are all the IO examples) illustrates
how to use the Stream classes. If the file does not exist, a new file is created and the
numbers from 0 to 9 are written to it. If the file already exists, the code starts reading 5
bytes from the end of the file and then writes them out. (You should run the example
twice. The first time creates and writes the file, and the second time reads the file.)
byte[] data = new Byte[10];

FileStream fs = new FileStream("FileStreamTest.txt",
FileMode.OpenOrCreate);
if (fs.Length == 0)
{
Console.WriteLine("Writing Data...");
for (short i = 0; i < 10; i++)
data[i] = (byte)i;
fs.Write(data, 0, 10);
}
else
{
fs.Seek(-5, SeekOrigin.End);
int count = fs.Read(data, 0, 10);
for (int i = 0; i < count; i++)
{
Console.WriteLine(data[i]);
}
}
fs.Close();
Primitive Datatypes and Streams
The stream derived classes will work if you are reading and writing bytes of data as a
block. If you need to read and write the primitive common types (Boolean, String, Int32,
and so on) in and out of a stream, use the BinaryReader and the BinaryWriter classes. The
Binary example shows how to use these classes. You create the appropriate stream
(FileStream in the example) and pass it to the BinaryReader or BinaryWriter constructor.
You can then use one of the overloaded Read or Write methods to read or write a datatype
to or from the stream. (Again, you should run the example twice.)
FileStream fs = new FileStream("BinaryTest.bin",

if (fs.Length == 0)
{
BinaryWriter w = new BinaryWriter(fs);
for (short i = 0; i < 10; i++)
w.Write(i);
w.Close();
}
else
{
BinaryReader r = new BinaryReader(fs);
for (int i = 0; i < 10; i++)
Console.WriteLine(r.ReadInt16());
r.Close();
}
fs.Close();
TextReader and TextWriter
The TextReader and TextWriter abstract classes treat the data as a sequential stream of
characters (i.e., as text). TextReader has methods such as Close, Peek, Read, ReadBlock,
ReadLine, and ReadToEnd. TextWriter has methods such as Close, Flush, Write, and
WriteLine. The overloaded Read methods read characters from the stream. The
overloaded Write and WriteLine methods write various types to the stream. If an object is
written to the stream, the object's ToString method is used.
StringReader and StringWriter are derived from TextReader and TextWriter. These
classes read and write characters from a string. The StringWriter's constructor uses a
StringBuilder object. The StringBuilder class was discussed in Chapter 3. StreamReader
and StreamWriter are also derived from TextReader and TextWriter. They read and write
text to and from a Stream object. As with the BinaryReader and BinaryWriter class you
create a stream and pass it to the constructor. Hence, these classes can use any Stream
derived class data storage. The Text example uses the StreamWriter and StreamReader
classes.
FileStream fs = new FileStream("TextTest.txt",

if (fs.Length == 0)
{
StreamWriter sw = new StreamWriter(fs);
sw.Write(100);
sw.WriteLine(" One Hundred");
sw.WriteLine("End of File");
sw.Close();
}
else
{
string text;
StreamReader sr = new StreamReader(fs);
text = sr.ReadLine();
while (text != null)
{
Console.WriteLine(text);
}
sr.Close();
}
fs.Close();
File Manipulation
The framework has two classes that are very useful for working with files.
If you need to manipulate the file in addition to reading and writing to it, the File class
provides the basic functionality. Since the File class has only static members, you have to
provide the name of the file as an argument. The FileInfo class has a constructor that
creates an object that represents a file. You then use the methods to manipulate that
particular file.
The File class methods always perform a security check. If you are going to continually
access a particular file, you may want to use the FileInfo class, because the security check
is made only once in the constructor. Security is discussed in more detail in Chapter 12.
File Class
The File class has methods for creating and opening files that return FileStream,
StreamWriter, or StreamReader objects that do the actual reading and writing. The
overloaded Create methods return a FileStream object. The CreateText method returns a
StreamWriter. The overloaded Open method can either create a new file or open an
existing one for reading or writing, depending on the method parameters. The object
returned is a FileStream object. The OpenText method returns a StreamReader. The
OpenRead method returns a FileStream object. The OpenWrite method returns a
FileStream.
The File class also has methods for copying, deleting, and moving files. You can test for
the existence of a file. File attributes can be read or modified, such as:
• creation time
• last access time
• last write time
• archive, hidden, normal, system, or temporary
• compressed, encrypted
• read-only
• whether the file is a directory
Path Class
Many of the file names needed for input arguments have to be full paths. Or you might
only want to manipulate parts of the path. The Path class has static methods that make
this easier. The Path class has static fields that indicate various platform-specific aspects
of pathnames such as the separator characters for directories, paths, and volumes, and the
illegal characters for pathnames.
Its static methods let you change the extension of a file, or find the directory where
temporary files reside. The GetFullPath method is particularly useful. You can pass it a
relative path such as ".\foo.txt" and it will return the full path of the file. This is very
useful for the File or security classes that require the full file path.
FileInfo Class
The FileInfo constructor creates an object that represents a disk file. The constructor
takes one argument, a string representing the name of the file. The class has properties
that represent file properties such as the creation time, full pathname, and size of the file.
It has creation and open methods that are analogous to the File class methods but operate
on this file instance and therefore do not need a file-name parameter. The FileInfo class
also has methods to move and copy the file.
File Example
The File example illustrates the use of the File and FileInfo classes.
The static Delete method of the File class is used to remove a previous version of a file if
it is present. The static CreateText method creates a new file and returns a StreamWriter
instance which is used to write some text to the file. The stream is then closed. The static
Move method then renames the file.
A FileInfo instance is constructed to represent this renamed file. The complete file name,
size, and creation date for the file are written to the console. The file is opened as text and
a StreamReader instance is used to read and write out the contents of the file.
File.Delete("file2.txt");
StreamWriter sw = File.CreateText("file.txt");
sw.WriteLine("The time has come the Walrus said, to talk
of many things");
sw.WriteLine("Of shoes, and ships, and sealing wax, of
cabbages and kings");
sw.WriteLine("And why the sea is boiling hot, and whether
pigs have wings.");
sw.Close();
File.Move("file.txt", "file2.txt");
FileInfo fileInfo = new FileInfo("file2.txt");
Console.WriteLine("File {0} is {1} bytes in length and was
created on {2}", fileInfo.FullName, fileInfo.Length,
fileInfo.CreationTime);
Console.WriteLine("");
StreamReader sr = fileInfo.OpenText();
String s = sr.ReadLine();
while (s != null)
{
Console.WriteLine(s);
s = sr.ReadLine();
}
sr.Close();
Serialization
Using the File and Stream classes can be quite cumbersome if you have to save a
complicated data structure with linked objects. You have to save the individual fields to
disk, remembering which field belongs to which object, and which object instance was
linked to another object instance. When restoring the data structure you have to
reconstitute that arrangement of fields and object references.
The serialization technology provided by the .NET Framework does this for you.
Serialization converts managed data structures to a byte stream. Deserialization converts
the byte stream back to managed data structures. Serializing and deserializing can be
done on different machines so long as they both host the CLR.
Objects can be serialized without writing special code because, as we have seen, the
runtime can query the object's metadata and the serialized stream tags each value with the
name of the field to which it applies.
To inform the framework that a class can be serialized, mark the class with the
System.Serializable attribute. Any field or property that should not be serialized can be
marked with the System.NonSerialized attribute. For example, fields that represent
calculated values need not be serialized. All you have to do is mark the class with the
attribute; you need write no other code to save the class's fields. Only instance fields are
saved; the static fields of a class are never saved.
The Serialization example shows how to apply serialization to the case study's
HotelBroker class in the Hotel assembly. The Serializable attribute has been applied to
the HotelBroker class definition. The Serializable attribute has also been applied to all the
classes that are used by HotelBroker or that HotelBroker derives from—Broker, Hotel,
HotelReservation, Reservable, and Reservation—because in order for HotelBroker to be
serializable, those classes must be as well. If any of those classes were not marked, a
runtime exception would be thrown when the framework tried to serialize an object of
that type.
[Serializable]
public class HotelBroker : Broker, IHotelInfo,
IHotelAdmin, IHotelReservation
{
[NonSerialized] private ArrayList cities;
. . .
[Serializable] public class Hotel : Reservable
. . .
[Serializable] public class HotelReservation : Reservation
. . .
[Serializable] public abstract class Reservable
. . .
[Serializable] public abstract class Reservation
. . .
[Serializable] public abstract class Broker
. . .
The cities field has been marked as NonSerialized, since the hotel's city is saved with the
serialized hotels and therefore can be restored, as the modified AddCity method
demonstrates. The cities field would be null if the HotelBroker class had been
deserialized, because the cities field was not saved. [5]
[5]
Of course we could have serialized the cities field and not have to deal with the case
where cities could be null, but we wanted to demonstrate the NonSerialized attribute.
private void AddCity(string city)

{
if (cities == null)
{
cities = new ArrayList();
foreach(Hotel h in units)
AddCity(h.City);
}
if (!cities.Contains(city))
cities.Add(city);
}
Serialization Objects
Although the framework knows how to save an object marked with the Serializable
attribute, you still have to specify the format in which the object is saved (i.e., store the
data as binary, XML, or some custom format), and the storage medium. To specify the
format in which an object is saved, you use an instance of an object that supports the
IFormatter interface. [6]
[6]
How does the runtime know whether a class supports the IFormatter interface? Query
the metadata!
The Framework ships with two such classes, System.Runtime.

Serialization.Formatters.Binary.BinaryFormatter and System.Runtime.
Serialization.Formatters.Soap.SoapFormatter. The BinaryFormatter uses a binary,
compact format for serializing and deserializing on platforms that support the Common
Language Runtime. The SoapFormatter uses the industry-standard SOAP protocol that is
discussed in Chapter 11 on Web Services. Since it is an XML-, and therefore text-based
protocol, it can be used to communicate with a non-CLR-based platform. The binary
format is faster when serializing and deserializing data.
You can, of course, implement your own formatter classes. You might do this if you had
to talk to a system with its own legacy byte format.
The Serialization example has code to demonstrate saving and restoring both binary and
SOAP formats using a FileStream. Of course, you could use any Stream-based class
representing some data medium.
private static void Save(HotelBroker broker,

string formatter)
{
FileStream s;
if (formatter == "b")
{
s = new FileStream("hotels.bin", FileMode.Create);
BinaryFormatter b = new BinaryFormatter();
b.Serialize(s, broker);
}
else
{
s = new FileStream("hotels.txt", FileMode.Create);
SoapFormatter sf = new SoapFormatter();
sf.Serialize(s, broker);
}
s.Close();
}
private static void Load(ref HotelBroker broker,
string formatter)
{
FileStream s;
if (formatter == "b")
{
s = new FileStream("hotels.bin", FileMode.Open);
BinaryFormatter b = new BinaryFormatter();
broker = (HotelBroker) b.Deserialize(s);
}
else
{
s = new FileStream("hotels.txt", FileMode.Open);
SoapFormatter sf = new SoapFormatter();
broker = (HotelBroker)sf.Deserialize(s);
}
s.Close();
Here is some sample output from the Serialization example. First we add a hotel and save
it with the SOAP formatter. We then exit the program.
Enter command: cities

Atlanta
Boston
Commands: quit, cities, list, add, fetch, save
Enter command: list

City Name Rooms Rate
Atlanta Dixie 100 115
Atlanta Marriott 500 70
Boston Sheraton 250 95
Enter command: add

Hotel City: Philadelphia
Hotel Name: Franklin
Number Rooms: 100
Room Rate: 200
Enter command: save

Formatter: b(inary), s(oap)s

Atlanta
Boston
Philadelphia
Enter command: list

Philadelphia Franklin 100 200
Enter command: quit
We then run the program again and restore what we saved [7] in the first run.
[7]
If you look at the hotels.txt file you will see a huge file, with a lot of "empty" entries.
This stems from the simplistic array data structure we used for reservations, which is a
very sparse matrix.

Atlanta
Boston
Enter command: list
Enter command: fetch

Formatter: b(inary), s(oap)s
Philadelphia Franklin 100 200

Atlanta
Boston
Philadelphia
ISerializable
Sometimes the serialization provided by the Framework is not satisfactory. You can
provide custom serialization for a class by implementing the ISerializable interface and
adding a constructor to the class. The ISerializable interface has one member:
GetObjectData. This method is used when data is serialized.
The ISerializable example demonstrates how this is done. As before, the class has to be
marked as Serializable.
[Serializable]
public class HotelBroker : Broker, IHotelInfo,
IHotelAdmin, IHotelReservation, ISerializable
{
private ArrayList cities;
...
The SerializationInfo class is used to store all the data that needs to be saved. This class's
AddValue method is overloaded to handle the saving of various types, including object. [8]
When you save the type you provide a name that is used to recall the object when
deserializing the data. The StreamingContext class gives you information about the
stream being used in the serialization. For example, you can find out if the stream being
used is a file or is being remoted to another computer.
[8]
Some of the AddValue overloads are not CLS compliant when the types being saved
are not CLS-compliant types such as unsigned integers. Be careful not to use those types
where .NET language interoperability is required. You have to watch for this in other
places in the Framework such as the Convert class or the Parse methods of the various
CTS types, or any other place where data is formatted, converted, read, or written out
(such as the TextWriter classes).
public void GetObjectData(SerializationInfo info,

StreamingContext context)
{
long numberHotels = units.Count;
info.AddValue("NumberHotels", numberHotels);
info.AddValue("Hotels", units);
}
You also have to implement a special constructor that is used by the framework to
recreate the object when it is deserialized. It has the same arguments as does
GetObjectData. Here you use the various Get methods on SerializationInfo to restore the
data. Note that since we did not save the cities field, we had to manually restore it. The
constructor is private because only the Framework uses it. If you forget to add the
constructor, you will get a SerializationException when you try to restore the object.
private HotelBroker(SerializationInfo info,

StreamingContext context) : base(MAXDAY, MAXUNIT)
{
long numberHotels = info.GetInt32("NumberHotels");
units = (ArrayList)info.GetValue("Hotels",
typeof(ArrayList));
if (numberHotels == units.Count)
Console.WriteLine("All hotels deserialized.");
else
Console.WriteLine("Error in deserialization.");
cities = new ArrayList();

foreach(Hotel h in units)
AddCity(h.City);
In this example we only did custom serialization for the HotelBroker object. For all the
other objects we still relied on the Framework's serialization. This example works the
same way that the Serialization example did. The sample output would look the same.
.NET Application Model
Serialization gave you a concrete example of the flexible environment the .NET
Framework provides for writing code. Now let us take a look at the model in which .NET
applications run. The Win32 environment in which a program runs is called its process.
This environment consists of
• the address space in which the code and data of the program reside
• one or more threads
• a set of environmental variables that is associated with the program
• a current drive and directory
Threads
A thread is the actual execution path of a program's code. One or more threads run inside
a process to allow for multiple execution paths inside a process. With multiple threads,
for example, a program can update the user interface with partial results on one thread as
a calculation proceeds on another thread. All threads in the same process share the
process environment, so that they can all access process memory.
Threads are scheduled by the operating system; processes and application domains [9] are
not scheduled. Threads are given a limited timeslice in which to run, so that they can
share the processor with other threads. Higher-priority threads will get to run more often
than lower-priority threads. After some time elapses, a thread will get another chance to
run. When a thread is restarted, it resumes running from where it was stopped.
[9]
Application domains are discussed later in this chapter.
Threads maintain a context, which has to be saved and restored when the operating
system's scheduler switches from one thread to another. A thread's context includes the
machine registers and stack that contain the state of the executing code.
The System.Threading.Thread class models an executing thread. The Thread object that
represents the current executing thread can be found from the static property
System.Threading.Thread.CurrentThread.
Unless your code runs on a multiprocessor machine, or you are trying to use time while a
uniprocessor waits for some event such as an I/O event, using multiple threads does not
save any time on your computing tasks. It does, however, allow making the system seem
more responsive to tasks requiring user interaction. Using too many threads can decrease
performance as contention between the threads for the CPU increases.
To help you understand threads we provide a four-part Threading example that uses the
Customer and Hotel assemblies from the case study to make reservations. Let us look
first at Step 0. The code is found in the file Threading.cs.
.NET threads run as delegates defined by the System.Threading. ThreadStart class. The
delegate returns void and takes no parameters.
public delegate void ThreadStart();
The NewReservation class has a public member function MakeReservation that will
define the thread function. Since the thread function takes no parameters, any data that
this function uses is assigned to fields in the NewReservation instance.
ThreadStart threadStart1 = new

ThreadStart(reserve1.MakeReservation);
The thread delegate is created and passed as a parameter to the constructor that creates
the System.Threading.Thread instance. The Start method on the Thread instance is
invoked to begin the thread's execution. When we discuss the asynchronous programming
model, we will show you how to pass parameters to a thread delegate. The program now
has two threads—the original one that executed the code to start the program, and the
thread we have just created that attempts to make a hotel reservation.
public class NewReservation

{
. . .
public void MakeReservation()
{
. . .
Console.WriteLine("Thread {0} starting.",
Thread.CurrentThread.GetHashCode());
. . .
ReservationResult result =
hotelBroker.MakeReservation(customerId, city, hotel,
date, numberDays);
. . .
}
}
. . .
NewReservation reserve1 = new NewReservation(customers,
hotelBroker);
reserve1.customerId = 1;
reserve1.city = "Boston";
reserve1.hotel = "Presidential";
reserve1.sdate = "12/12/2001";
reserve1.numberDays = 3;

Thread thread1 = new Thread(threadStart1);
Console.WriteLine("Thread {0} starting a new thread.",
thread1.Start();
To cause the original thread to wait until the second thread is done, the Join method on
the System.Threading.Thread instance is called. The original thread now blocks (waits)
until the reservation thread is complete. The results of the reservation request are written
to the console by the reservation thread.
thread1.Join();
Console.WriteLine("Done!");
Thread Synchronization
An application can create multiple threads. Look at the code in Step 1 of the Threading
example. Now multiple reservation requests are being made simultaneously.

hotelBroker);
. . .
hotelBroker);
. . .

thread1.Start();
thread2.Start();
thread1.Join();
thread2.Join();
The problem with our reservation systems is that there is no guarantee that one thread
will not interfere with the work being done with the other thread. Threads run only for a
brief period before they are interrupted and another thread is scheduled to run on the
processor. They may not be finished with whatever operation they were working on when
their timeslice is up.
For example, they might be in the middle of updating a data structure. If another thread
tries to use the information in that data structure, or update the data structure, the results
of operations will be at best inconsistent and incorrect, and at worst a system crash (i.e., if
references to obsolete structures were not yet updated).
Let us look at one of several places in the customer and reservation code where we could
have a problem. Examine the code for the Reserve method in the file broker.cs. First a
check is made of the existing bookings for a given hotel for a given date to see if rooms
are available. If there are, the booking is made.
. . .
{
if (numCust[i, unitid] >= unit.capacity)
{
return result;
}
}
. . .
. . .
This code can produce inconsistent results! One thread could be rescheduled after it finds
that the last room is available, but before it gets a chance to make the booking. The other
thread could run, find the same available room, and make the booking. When the second
thread runs again, starting from where it left off, it will also book the last room at the
hotel.
To simulate this occurrence, this step of the threading example puts a

System.Threading.Thread.Sleep call between the code that checks for availability and the
code that makes the booking. The Sleep(0) call will cause the thread to stop executing
and give up the remainder of its timeslice. We then setup our program so that the two
threads try to reserve the only room at a hotel for the same time. Examine the code in the
Main routine that sets this up:
hotelBroker.AddHotel("Boston", "Presidential", 1,
(decimal) 10000);
. . .
hotelBroker);

hotelBroker);
Running the program will give results something like this:

Added Boston Presidential Hotel with one room.
Thread 3 starting new threads.
Thread 5 starting.
Reserving for Customer 1 at the Boston Presidential Hotel
on 12/12/2001 12:00:00 AM for 3 days
Thread 5 entered Broker::Reserve
Thread 5 sleeping in Broker::Reserve
Thread 6 starting.
on 12/13/2001 12:00:00 AM for 1 days
Thread 6 sleeping in Broker::Reserve
Thread 5 left Broker::Reserve
Reservation for Customer 1 has been booked
ReservationId = 1
ReservationRate = 10000
ReservationCost = 30000
Comment = OK
ReservationId = 2
Comment = OK
Done!
Both customers get to reserve the last room on December 13! Note how Thread 5 enters
the Reserve method and finds the room is available before it gets rescheduled. Thread 6
enters Reserve and also finds the room is available before it gets rescheduled. Thread 5
then books the room, and Thread 6 does as well.
Operating systems provide means for synchronizing the operation of multiple threads, or
multiple processes accessing shared resources. The .NET Framework provides several
mechanisms to prevent threading conflicts.
Every object in the .NET framework can be used to provide a synchronized section of
code (critical section). Only one thread at a time can execute within such a section. If one
thread is already executing inside that synchronized code section, any threads that
attempt to access that section will block (wait) until the executing thread leaves it.
Synchronization with Monitors
The System.Threading.Monitor class allows you to create a critical section by

synchronizing on an object to avoid race conditions or incorrect answers. Step 2 of the
Threading example demonstrates the use of the Monitor class with the this pointer of the
HotelBroker instance.
public ReservationResult Reserve(Reservation res)

{
. . .
Monitor.Enter(this);
. . .
Monitor.Exit(this);
return result;
}
The thread that first calls the Monitor.Enter(this) method will be allowed to execute the
code of the Reserve method because it will acquire the Monitor lock based on the this
pointer. Subsequent threads that try to execute will be forced to wait until the first thread
releases the lock with Monitor.Exit(this). At that point they will be able to call
Monitor.Enter(this) and acquire the lock.
A thread can call Monitor.Enter several times, but each call must be balanced by a call to
Monitor.Exit. If a thread wants to try to acquire a lock, but does not want to block so that
it can do some work and try again, it can use the Monitor.TryEnter method.
In C# you can use the lock keyword in place of Monitor.Enter/Exit. With the lock
keyword, the above fragment would be:
public ReservationResult Reserve(Reservation res)

{
lock(this);
{
. . .
}
return result;
}
Now that we have provided synchronization, the identical case tried in Step 1 does not
result in one reservation too many for the hotel. Notice how the second thread cannot
enter the Reserve method until the first thread that entered has left.

Thread 3 launching 2 threads.
Thread 5 starting.
on 12/12/2001 12:00:00 AM for 3 days
Thread 5 trying to enter Broker::Reserve
Thread 6 starting.
on 12/13/2001 12:00:00 AM for 1 days
Thread 6 trying to enter Broker::Reserve
Reservation for Customer 2 could not be booked
Room not available
ReservationId = 1
Comment = OK
Done!
Notification with Monitors
A thread that has acquired a Monitor lock can wait for a signal from another thread that is
synchronizing on that same object without leaving the synchronization block. The thread
invokes the Monitor.Wait method and relinquishes the lock. When notified by another
thread, it reacquires the synchronization lock.
A thread that has acquired a Monitor lock can send notification to another thread waiting
on the same object with the Pulse or the PulseAll methods. It is important that the
receiving thread be waiting when the pulse is sent; otherwise, if the pulse is sent before
the wait, the other thread will wait forever and will never see the notification. This is
unlike the reset events discussed later in this chapter. If multiple threads are waiting, the
Pulse method will put only one thread on the ready queue to run. The PulseAll will put all
of them on the ready queue.
The pulsing thread no longer has the monitor lock but is not blocked from running. Since
it is no longer blocked, but does not have the lock, to avoid a deadlock or race condition
this thread should try to reacquire the lock (through a Monitor.Enter or Wait) before
doing any potentially damaging work.
The PulseAll example illustrates the Pulse and PulseAll methods. Running the example
produces the following output:
First thread: 2 started.

Thread: 5 started.
Thread: 6 started.
Thread: 6 waiting.
Thread: 5 waiting.
Thread 6 sleeping.
Done.
Thread 6 awake.
Thread: 6 exited.
Thread 5 sleeping.
Thread 5 awake.
Thread: 5 exited.
The class X has a field "o" of type object that will be used for a synchronization lock.
The class also has a method Test that will be used as a thread delegate. The method
acquires the synchronization lock and then waits for a notification. When it gets the
notification, it sleeps for half a second and then relinquishes the lock.
The main method creates two threads that use Test method of class X as their thread
delegate and share the same object to use for synchronization. It then sleeps for 2 seconds
to allow the threads to issue their wait requests and relinquish their locks. Next it calls
PulseAll to notify both waiting threads and relinquishes its hold on the locks. Eventually
each thread will reacquire the lock, write a message to the console, and relinquish the
lock for the last time.
class X
{
object o;
public X(object o)
{
this.o = o;
}
public void Test()
{
try
{
long threadId =Thread.CurrentThread.GetHashCode();
Console.WriteLine("Thread:{0} started.",threadId);
Monitor.Enter(o);
Console.WriteLine("Thread:{0} waiting.",threadId);
Monitor.Wait(o);
Console.WriteLine("Thread {0} sleeping",threadId);
Thread.Sleep(500);
Console.WriteLine("Thread {0} awake.", threadId);
Monitor.Exit(o);
Console.WriteLine("Thread: {0} exited.",threadId);
}
catch(Exception e)
{
long threadId =Thread.CurrentThread.GetHashCode();
Console.WriteLine("Thread: {0} Exception: {1}",
threadId, e.Message);
Monitor.Exit(o);
}
}
}
class Class1
{
static public object o = new object();

{
Console.WriteLine("First thread: {0} started.",
X a = new X(o);
X b = new X(o);
ThreadStart ats = new ThreadStart(a.Test);

ThreadStart bts = new ThreadStart(b.Test);
Thread at = new Thread(ats);

Thread bt = new Thread(bts);
at.Start();
bt.Start();
Thread.Sleep(2000);
Monitor.Enter(o);
Monitor.PulseAll(o);
// Monitor.Pulse(o);
Monitor.Exit(o);
Console.WriteLine("Done.");
}
}
Comment out the PulseAll call, uncomment the Pulse call, and only one thread completes
because the other thread is never put on the ready queue. Remove the Sleep(2000) from
the main routine and the other threads block forever, because the pulse occurs before the
threads get a chance to call the Wait method and hence they will never be notified.
These methods can be used to coordinate several threads' use of synchronization locks.
The Thread.Sleep method causes the current thread to stop execution (block) for a given
time period. Calling Thread.Suspend will cause the thread to block until Thread.Resume
is called on that same thread. Threads can also block because they are waiting for another
thread to finish (Thread.Join). This method was used in the Threading examples so that
the main thread could wait until the reservation requests were completed. Threads can
also block because they are waiting on a synchronization lock.
A blocked thread can be awakened by calling Thread.Interrupt on the blocked thread. The
thread will receive a ThreadInterruptedException. If it does not catch this exception, the
runtime will catch it and kill the thread.
If, as a last resort, you have to kill a thread outright, call the Thread.Abort method on the
thread. Thread.Abort causes the ThreadAbortException to be thrown. This exception
cannot be caught, but it will cause all the finally blocks to be executed. In addition,
Thread.Abort does not cause the thread to wake up from a wait.
Since finally blocks may take a while to execute, or the thread might be waiting, aborted
threads may not terminate immediately. If you need to be sure that the thread has
finished, you should wait on the thread's termination using Thread.Join.
Synchronization Classes
The .NET Framework has classes that represent the standard Win32 synchronization
objects. These classes all derive from the abstract WaitHandle class. This class has static
methods, WaitAll and WaitAny, that allow you to wait for all of a set of synchronization
objects being signaled or on just one of a set of synchronization objects being signaled. It
also has an instance method, WaitOne, that allows you to wait for this instance to be
signaled. How the object gets signaled depends on the particular type of synchronization
object that is derived from WaitHandle.
A Mutex object is used for interprocess synchronization. Monitors and synchronized code
sections work only within one process. An AutoResetEvent and ManualResetEvent are
used to signal whether an event has occurred. An AutoResetEvent remains signaled until
a waiting thread is released. A ManualResetEvent remains signaled until its state is set to
unsignaled with the Reset method. Hence many threads could be signaled by this event.
Unlike Monitors, code does not have to be waiting for the signal before the pulse is set
for the reset events to signal a thread.
The Framework has provided classes to solve some standard threading problems. The
Interlocked class methods allow atomic operations on shared values such as increment,
decrement, comparison, and exchange. ReaderWriterLock is used to allow single-writer,
multiple-reader access to data structures. The ThreadPool class can be used to manage a
pool of worker threads.
Automatic Synchronization
You can use attributes to synchronize the access to instance methods and fields of a class.
Access to static fields and methods is not synchronized. To do this, you derive the class
from the class System.ContextBoundObject and apply a Synchronization attribute to the
class. This attribute cannot be applied to an individual method or field.
The attribute is found in the System.Runtime.Remoting.Contexts namespace. It describes

the synchronization requirements of an instance of the class to which it is applied. You
can pass one of four values which are static fields of the SynchronizationAttribute class
to the SynchronizatonAttribute constructor: NOT_SUPPORTED, SUPPORTED,
REQUIRED, REQUIRES_NEW. The Threading example Step 3 illustrates how to do
this.
Synchronization(SynchronizationAttribute.REQUIRED)]
public abstract class Broker : ContextBoundObject
{
. . .
In order for the CLR to make sure that the thread in which this object runs on is
synchronized properly, the CLR has to track the threading requirements of this object.
This state is referred to as the context of the object. If one object needs to be
synchronized, and another does not, they are in two separate contexts. The CLR has to
acquire a synchronization lock on behalf of the code when a thread that is executing a
method on the object that does not need to be synchronized starts executing a method on
an object that does. The CLR knows that this has to be done because it can compare the
threading requirements of the first object with the threading requirements of the second
object by comparing their contexts.
Objects that share the same state are said to live in the same context. For example, two
objects that do not need to be synchronized can share the same context.
ContextBoundObject and Contexts are discussed in more detail in the section on
Contexts.
With this intuitive understanding of contexts we can now explain the meaning of the
various Synchronization attributes. NOT_SUPPORTED means that the class cannot
support synchronization of its instance methods and fields and therefore must not be
created in a synchronized context. REQUIRED means that the class requires
synchronization of access to its instance methods and fields. If a thread is already being
synchronized, however, it can use the same synchronization lock and live in an existing
synchronization context. REQUIRES_NEW means that not only is synchronization
required, but access to its instance methods and fields must be with a unique
synchronization lock and context. SUPPORTED means that the class does not require
synchronization of access to its instance methods and fields, but a new context does not
have to be created for it.
You can also pass a Boolean flag to the constructor to indicate if reentrancy is required. If
required, call-outs from methods are synchronized. Otherwise, only calls into methods
are synchronized.
With this attribute there is no need for Monitor.Enter and Monitor.Exit in the
Broker::Reserve method.
Just as in Step 2, this example attempts to make two reservations for the last room in a
Hotel. In addition, a third thread attempts to cancel a reservation. Here is the output from
running this example:

Thread 13 launching 3 threads.
MakeReservation: Thread 28 starting.
on 12/12/2001 12:00:00 AM for 3 days
Thread 28 entered Reserve.
MakeReservation: Thread 29 starting.
on 12/13/2001 12:00:00 AM for 1 days
CancelReservation: Thread 30 starting.
Cancelling Reservation 10
Thread 28 left Reserve.
Thread 29 entered Reserve.
Thread 29 left Reserve.
Room not available
Thread 30 entered CancelReservation.
Thread 30 left CancelReservation.
ReservationId = 1
Comment = OK
Done!
As in the previous case the second thread could not enter the Reserve method until the
thread that entered first finished. Only one reservation is made.
What is different about using the automatic approach is that you get the synchronization
in all the methods of the class whether you need it or not. Accessing any data in the class
is also singly threaded.
Note how only one thread can be in any method of the class; a thread using
CancelReservation blocks threads from using MakeReservation. With a reservation
system this is the behavior you want, since you do not want the MakeReservation to
attempt to use a data structure that might be in the middle of being modified. In situations
where a method on the object does not require synchronization, however, you will be
synchronized anyway and the interactivity of the application will be reduced.
The other drawback to this approach is that it can increase contention and interfere with
scalability since you are not just locking around the specific areas that need
synchronizing.
The attribute approach is simpler than using critical sections. You do not have to worry
about the details of the getting the synchronization correct. On the other hand, you get
behavior that reduces interactivity and scalability. Different applications, or different
parts of the same application, will choose the approach that makes the most sense.
Thread Isolation
An exception generated by one thread will not cause another thread to fail. The
ThreadIsolation example demonstrates this.
class tm
{
public void m()
{
Console.WriteLine("Thread {0} started",
Thread.Sleep(1000);
for(int i = 0; i < 10; i++)
Console.WriteLine(i);
}
class te
{
public void tue()
{
Console.WriteLine("Thread {0} started",
Exception e = new Exception("Thread Exception");
throw e;
}
}
class ThreadIsolation
{
{
tm tt = new tm();
te tex = new te();
ThreadStart ts1 = new ThreadStart(tt.m);
ThreadStart ts2 = new ThreadStart(tex.tue);
Thread thread1 = new Thread(ts1);
Thread thread2 = new Thread(ts2);
Console.WriteLine("Thread {0} starting new threads.",
thread1.Start();
thread2.Start();
Console.WriteLine("Thread {0} done.",
}
}
The following output is generated. Note how the second thread can continue to write out
the numbers even though the first thread has aborted from the unhandled exception. Note
also how the "main" thread that spawned the other two threads can finish without causing
the others to terminate.
Thread 2 starting new threads.

Thread 2 done.
Thread 5 started
Thread 6 started
Unhandled Exception: System.Exception: Thread Exception

at te.tue() in F:\Work\ThreadIsolation.cs:line 23
0
1
2
3
4
5
6
7
8
9
The AppDomain class (discussed later in the chapter) allows you to set up a handler to
catch an UnhandledException event.
Synchronization of Collections
Some lists, such as TraceListeners, are thread safe. When this collection is modified, a
copy is modified and the reference is set to the copy. Most collections, like ArrayList, are
not thread safe. Making them automatically thread safe would decrease the performance
of the collection even when thread safety was not an issue.
An ArrayList has a static Synchronized method to return a thread-safe version of the

ArrayList. The IsSynchronized property allows you to test whether the ArrayList you are
using is the thread-safe version. The SyncRoot property can return an object that can be
used to synchronize access to a collection. This allows other threads that might be using
the ArrayList to be synchronized with the same object.
Context
In order for us to understand how the runtime is able to enforce a threading requirement
based on an attribute, we have to introduce the concept of context. Step 4 of the
Threading example is the same code as Step 3, but with some additional output:
Is the customer object a proxy? False

Is the bookings object a proxy? True
Thread 13 ContextId 0 launching 3 threads.
MakeReservation: Thread 28 ContextId 0 starting.
on 12/12/2001 12:00:00 AM for 3 days
Thread 28 ContextId 1 entered Reserve.
on 12/13/2001 12:00:00 AM for 1 days
CancelReservation: Thread 30 ContextId 0 starting.
Cancelling Reservation 10
Thread 28 ContextId 1 left Reserve.
Thread 29 ContextId 1 left Reserve.
Room not available
Thread 30 ContextId 1 entered CancelReservation.
Thread 30 ContextId 1 left CancelReservation.
ReservationId = 1
Comment = OK
In the last step of the Threading example we see that when a thread enters a method of
the Broker class, it has a different ContextId than when it runs outside of the Broker
class. It runs in a different context.

...
Objects derived from Broker have different runtime requirements (such as different
synchronization requirements) than the other objects in the program, since access to
Broker objects must be synchronized and access to other objects need not be
synchronized. The environment that represents an object's runtime requirements that the
CLR needs to be aware of is called a context. There are two contexts in the Threading
Step 3 example—Context 1 where the Broker object lives and Context 0 where all other
objects live. Every thread in the program runs in Context 1 when executing inside a
Broker object, Context 0 everywhere else. Contexts are independent of threads.
A context is a collection of one or more objects that have identical runtime requirements.
The .NET concept of a context is identical to the COM+ concept of a context. [10] In
general you cannot say what the runtime must do to in a given context because it depends
on exactly what the runtime requirements are. A context that has transactional
requirements requires different action than one that does not. Or a context that has to
maintain a REQUIRED synchronization requirement is different from one that has to
maintain a REQUIRES_NEW synchronization requirement.
[10]
At this point in time, though, COM+ contexts and .NET contexts are different. For a
discussion of contexts in COM+ see Understanding and Programming COM+ by Robert
J. Oberg.
You can get the Context class instance that represents the current context from the static
property Thread.CurrentContext. ContextId is a property of that class.
Proxies and Stubs
How does the runtime enforce the different requirements of different contexts? When an
object resides in another context (such as the HotelBroker object in the NewReservation
instance), an object reference to a proxy object is returned instead of a reference to the
object itself. The actual object resides in its original, or home, context. The proxy is an
object that represents the original object in a different context. The static method
RemotingServices.IsTransparentProxy determines whether an object reference points to a
real object instance or a proxy. Look at the code in the Threading Step 4 example main
routine:

customers = new Customers();
bool bTrans;
bTrans = RemotingServices.IsTransparentProxy(customers);
Console.WriteLine("Is the customer object a proxy? {0}",
bTrans);
bTrans = RemotingServices.IsTransparentProxy(hotelBroker);
Console.WriteLine("Is the bookings object a proxy? {0}",
bTrans);
which causes the following output:
Is the customer object a proxy? False

Is the bookings object a proxy? True
When a program starts up it is given a default context. [11] All objects, like the Customers
object, that do not have any special requirements are created inside that context (context
0). An object, such as the HotelBroker object, that has a different set of requirements
(synchronization) is created in a different context (context 1), and a proxy is returned to
the creating context (context 0).
[11]
As will be clear in the next section, the sentence should really read, "When a new
application domain starts up, it is given a new default context." Contexts are application-
domain relative. Two different application domains will have two separate default
contexts, each with id 0.
Now when you access the MakeReservation method in the HotelBroker object, you are
actually accessing a method on the proxy. The proxy method can apply the
synchronization lock and then delegate to the actual HotelBroker object's method. When
the actual object's method returns, it returns to the proxy. The proxy can then remove the
synchronization lock and return to the caller. This technique, where the runtime uses a
proxy to intercept method calls to the actual object, is called interception.
ContextBoundObject
The Broker class has to derive from the class ContextBoundObject so that the runtime
knows to setup a different context if one is required. If you remove the derivation of
Broker from ContextBoundObject, you will once again get the unsynchronized access,
and both customers will be able to reserve the last room at the hotel, even though the
class is still marked with the Synchronization attribute. Objects that do not derive from
ContextBoundObject can run in any context (agile objects).
Since other contexts work with a proxy, or a reference to the actual object, the runtime
must translate (marshal) the call from one context to another. Hence,
ContextBoundObject inherits from MarshalByRefObject. MarshalByRefObject is the
base class for objects that want to be able to be marshaled by reference. Otherwise, as we
will discuss in the section on application domains, by default, objects are marshaled by
value (i.e., copied).
One advantage of using synchronization techniques such as a Monitor is that a Monitor

can be called from any context. Another potential disadvantage of using automatic
synchronization is the performance hit from marshaling and using proxies rather than the
actual object.
As will be clear when we discuss Application Domains, since the customer object has no
dependency on context, it is the actual object, not a proxy. It can be copied to any context
within the same application domain.
Application Isolation
When writing applications it is often necessary to isolate parts of them so that a failure of
one part does not cause a failure in another. In Windows, application isolation has been at
the process level. In other words, if a process is stopped or crashes, other processes will
be unaffected. Unless shared memory is used, one process cannot directly address
memory in another process's address space.
For an application to use separate processes to achieve isolation is expensive. To switch

from one process to another, the contents of the machine registers must be saved and
restored. This includes a thread and process switch. A thread switch requires saving
registers, such as the instruction pointer, and loading the information for a new thread, as
well as updating the scheduling information for the threads. A process switch includes,
accounting information and processor rights that have to be saved for the old process and
restored for the new one.
Application Domain
The .NET Application Domain is a more lightweight unit for application isolation, fault
tolerance, and security. Multiple application domains can run in one process. Since the
.NET code can be checked for type safety and security, the CLR can guarantee that one
App Domain can run without interference from another App Domain in the same process.
No process switch is required to achieve application isolation.
Application Domains can have multiple contexts, but a context exists in only one
AppDomain. Although a thread runs in one context of one application domain at a time,
the Threading example Step 3 demonstrates that a thread can execute in more than one
context. One or more threads can run in an App Domain at the same time. An object lives
in only one context.
Each AppDomain starts with a single thread and one context. Additional threads and
contexts are added as needed.
There is no relationship between the number of application domains and threads. A Web
server might require an application domain for each hosted application that runs in its
process. The number of threads in that process would be far fewer, depending on how
much actual concurrency the process can support.
To enforce application isolation, code in one application domain cannot make direct calls
into the code (or even reference resources) in another application domain. They must use
proxies.
Application Domains and Assemblies
Applications are built from one or more assemblies, but each assembly is loaded into a
particular application domain. Each application domain can be unloaded independently of
the others, but you cannot unload an individual assembly from an App Domain. The
assembly will be unloaded when the App Domain is unloaded. Unloading an App
Domain also frees all resources associated with that App Domain.
Each process has a default application domain that is created when the process is started.
This default domain can be unloaded only when the process shuts down.
Applications such as ASP.NET or Internet Explorer critically depend on preventing the

various applications that run under it from interfering with each other. By never loading
application code into the default domain, they can ensure that a crashing program will not
bring down a host.
AppDomain Class
The AppDomain class abstracts application domains. The AppDomain sample illustrates
the use of application domains.
This class has static methods for creating and unloading application domains:
AppDomain domain = AppDmain.CreateDomain("CreatedDomain2",

null, null);
. . .
AppDomain.Unload(domain);
While the CreateDomain method is overloaded, one signature illustrates application-

domain isolation:
AppDomain CreateDomain(string Name, Evidence securityInfo,

AppDomainSetup info)
The Evidence parameter is a collection of the security constraints on the application

domain. While we will discuss this in greater detail in the Security chapter, the domain's
creator can modify this collection to control the permissions that the executing app
domain can have. The AppDomainSetup parameter specifies setup information about the
domain. Among the information specified is the location of the App Domain's
configuration file and where private assemblies are loaded. Hence, each App Domain can
be configured independently of every other. Code isolation, setup isolation, and control
over security combine to ensure that application domains are independent of each other.
App Domain Events
To help in maintaining application isolation, the AppDomain class allows you to set up
event handlers for:
• when a domain unloads

• when the process exits
• when an unhandled exception occurs
• when attempts to resolve assemblies, types, and resources fail
AppDomain Example
The AppDomain example lets us examine various aspects of application domains. If you
run the example you will get the output in Figure 8-1.
Figure 8-1. Output of AppDomain example.
First, the name, thread, and context of the default domain are written out.
AppDomain currentDomain = AppDomain.CurrentDomain;

Console.WriteLine("At startup, Default AppDomain is {0}
ThreadId: {1} ContextId {2}\n",
currentDomain.FriendlyName,
Thread.CurrentThread.GetHashCode(),
Thread.CurrentContext.ContextID);
We then load and execute an assembly. The code in this assembly just prints out a string
and its domain's name, thread, and context. Notice that it executes in the default domain.
int val = currentDomain.ExecuteAssembly("TestApp.exe");
We then create an instance of the Customers type from the Customer assembly in the
default domain. The CreateInstance method of the AppDomain class returns an
ObjectHandle instance. You can pass this ObjectHandle between application domains
without loading the metadata associated with the wrapped type. When you want to use
the object as its actual type instead of as an opaque object instance, you must unwrap it
by calling the Unwrap method on the ObjectHandle instance.
ObjectHandle oh = currentDomain.CreateInstance("Customer",
"OI.NetCs.Acme.Customers");
. . .
Customers custs = (Customers)oh.Unwrap();
We add a new customer and then list all the existing customers. Notice that both the
constructor of this type and the methods execute in the same thread and context that the
default domain does.
We then create a new domain and create an instance of the same type as before in that
new domain.

null, null);
. . .
oh = domain.CreateInstance("Customer",
"OI.NetCs.Acme.Customers");
. . .
Customers custs2 = (Customers)oh.Unwrap();
Note that the constructor call that results from the CreateInstance method executes in the
new domain and is therefore in a different context from where the CreateInstance call
was made, but it is executing on the same thread that made the CreateInstance call.
When we list the customers in this new object, we get a different list of customers. This is
not surprising, since it is a different Customers object. Nonetheless, the customer list
method executes in the default domain!
Using RemotingServices.IsTransparentProxy, we see that the ObjectHandle is a proxy to

the Customers object that lives in the newly created AppDomain. However, when you
unwrap the object to get an instance handle, you do not get a proxy, but you get an actual
object reference. By default, objects are marshaled by value (copied) from one
AppDomain to another.
If the Customers object is not serializable, you will get an exception when you try to copy
it because the runtime will not know how to make the copy. This exception would be
thrown when you do the Unwrap, not the CreateInstance. The latter returns a reference;
the copy is made only when the ObjectHandle is unwrapped. If the object cannot be
serialized, it cannot be copied from one AppDomain to another.
Next we create a new thread, and that thread creates a new application domain, loads, and
executes an assembly. The assembly starts executing at its entry point, the Main routine
of the AppDomainTest class.

null, null);
. . .
string[] args = new String[1];
args[0] = "MakeReservation";
int val = domain.ExecuteAssembly("TestApp.exe", null,
args);
. . .
AppDomain.Unload(domain);
The Main routine loads the Hotel assembly into the newly created App Domain. It then
queries the metadata of the assembly for the HotelBroker type information. It uses that
type information to create a HotelBroker object. The HotelBroker class is marked with a
synchronization attribute. As a result, the HotelBroker constructor and the
MakeReservation method run in a different context than the default context.
Assembly a = AppDomain.CurrentDomain.Load("Hotel");
Type typeHotelBroker =
a.GetType("OI.NetCs.Acme.HotelBroker");
HotelBroker hotelBroker =
(HotelBroker)Activator.CreateInstance(typeHotelBroker);
DateTime date = DateTime.Parse("12/2/2001");
ReservationResult rr = hotelBroker.MakeReservation(1,
"Boston", "Sheraton", date, 3);
Console.WriteLine("\tReservation Id: {0}",
rr.ReservationId);
Marshaling, AppDomains, and Contexts
By default, objects are copied from one App Domain to another (marshal by value). The
Remoting section will show how to marshal by reference between App Domains. This
ensures that code in one application domain is isolated from another.
Objects are marshaled by reference between contexts. This allows the CLR to enforce the
requirements (such as synchronization or transactions) of different objects. This is true
whether the client of the object is in the same application domain or not.
Since most objects do not derive from ContextBoundObject, they can reside or move
from one context to another as required. Threads can cross application domain and
context boundaries within the same Win32 process.
Asynchronous Programming
.NET supports a design pattern for asynchronous programming. This pattern is present in
many places in .NET (including I/O operations, as noted earlier, and as we will see in
Chapter 11 for Web services). Asynchronous programming provides a way for you to
provide a method call without blocking the method caller. From the perspective of the
client, the asynchronous model is easier to use than threading. It offers much less control
over the synchronization than using synchronization objects, however, and the class
designer would probably find threading much easier to use.
The Asynchronous Design Pattern
This design pattern is composed of two parts, a set of methods and an interface
IAsyncResult. The methods of the pattern are:
IAsyncResult BeginXXX(inputParams, AsyncCallback cb,

Object AsyncObject)
ReturnValue EndXXX(outputParams, IAsyncResult ar);
As a design pattern, the XXX represents the actual method being called asynchronously
(i.e., BeginRead/EndRead for the System.IO.FileStream class). The BeginXXX should
pass all input parameters of the synchronous version (in, in/out, and ref) as well as the
AsyncCallback and AsyncObject parameters. The EndXXX should have all the output
parameters of the synchronous version (ref ,out, and in/out) parameters in its signature. It
should return whatever object or value the synchronous version of the method would
return. It should also have an IAsyncResult parameter. A CancelXXX can be provided if
it makes sense.
The AsyncCallback is a delegate that represents a callback function.
public delegate void AsyncCallback(IAsyncResult ar);
The AsyncObject is available from IAsyncResult. It is provided so that in the callback

function you can distinguish which asynchronous read the callback was generated by.
The Framework uses this pattern so that the FileStream synchronous Read
int Read(byte[] array, int offset, int count);
becomes in the asynchronous version:
IAsyncResult BeginRead(byte[] array, int offset,

int numBytes, AsyncCallback userCallback,
object stateObject);
int EndRead(IAsyncResult asyncResult);
Any exception thrown from BeginXXX should be thrown before the asynchronous
operation starts. Any exceptions from the asynchronous operation should be thrown from
the EndXXX method.
IAsyncResult
IAsyncResult is returned by a BeginXXX method (such as BeginRead). This interface

has four elements:
interface IAsyncResult
{
public boolean IsCompleted();
public boolean CompletedSynchronously;
public WaitHandle AsyncWaitHandle;
public Object AsyncState;
}
IsCompleted is set to true after the server has processed the call. The client can then
destroy all resources. If BeginXXX completed synchronously, CompletedSynchronously
is set to true. Most of the time this is ignored, and CompletedSynchronously is set to the
default value of false. In general, a client never knows whether the BeginXXX method
executed asynchronously or asynchronously. If the asynchronous operation is not yet
finished, the EndXXX method will block until it is.
The AsyncWaitHandle returns a WaitHandle that can be used for synchronization. As

discussed previously, this handle can be signaled, so that the client can wait on it. Since
you can specify a wait time period, you do not have to block forever if the operation is
not yet complete.
The AsyncState is the object provided as the last argument in the BeginXXX call. Since it
is contained in the IAsyncResult passed to the callback function, examining its value
would allow you to determine which BeginXXX caused this particular instance of the
callback.
Using Delegates for Asynchronous Programming
Any developer of .NET objects who wants to provide an asynchronous interface should
follow this pattern. Nonetheless, there is no need for most developers to develop a custom
asynchronous solution for their objects. Delegates provide a very easy way to support
asynchronous operations on any method without any action on the class developer's part.
Of course, this has to be done with care, because the object was written with certain
assumptions about the thread it is running on and its synchronization requirements.
The two Asynch examples use the Customers object from our case study Customer
assembly. The first example registers new customers asynchronously and does some
processing while waiting for each registration to finish. The second example uses a
callback function with the asynchronous processing. In addition to allowing the program
to do processing while waiting for the registrations to finish, the callback allows the
system to take some asynchronous action for each individual registration.
In the examples, we just print out to the console to show where work could be done. To
increase the waiting time to simulate longer processing times we have put calls to
Thread.Sleep() in Customers:: RegisterCustomer as well as in the sample programs. Now
let us look at the code within the examples.
Suppose the client wants to call the RegisterCustomer method asynchronously. The caller
simply declares a delegate with the same signature as the method.
public delegate int RegisterCustomerCbk(string firstName,

string LastName, string EmailAddress);
You then make the actual method the callback function:
RegisterCustomerCbk rcc = new

RegisterCustomerCbk(customers.RegisterCustomer);
Begin/End Invoke
When you declare a delegate, the compiler generates a class with a constructor and three
methods: BeginInvoke, EndInvoke, and Invoke. The BeginInvoke and EndInvoke are
type-safe methods that correspond to the BeginXXX and EndXXX methods and allow
you to call the delegate asynchronously. The Invoke method is what the compiler
implicitly uses when you call a delegate. [12] To call RegisterCustomer asynchronously
just use the BeginInvoke and EndInvoke methods.
[12]
If you open the executable from the DelegateAccount example in Chapter 5 in
ILDASM, you can observe this. The NotifyCallback class has the BeginInvoke,
EndInvoke, and Invoke methods defined. If you look at the Withdraw method for
Account, you will notice that the C# line notifyDlg(balance) has been transformed to
instance void NotifyCallback::Invoke(valuetype [mscorlib]System.Decimal).

for(int i = 1; i < 5; i++)

{
firstName = "FirstName" + i.ToString();
lastName = "SecondName" + (i * 2).ToString();
emailAddress = i.ToString() + ".biz";
IAsyncResult ar = rcc.BeginInvoke(firstName, lastName,

emailAddress, null, null);
while(!ar.IsCompleted)
{
Console.WriteLine("Could do some work here while wait
ing for customer registration to complete.");
ar.AsyncWaitHandle.WaitOne(1, false);
}
customerId = rcc.EndInvoke(ar);
Console.WriteLine(" Added CustomerId: " +
customerId.ToString());
}
The program waits on the AsyncWaitHandle periodically to see if the registration has
finished. If it has not, some work could be done in the interim. If EndInvoke is called
before RegisterCustomer is complete, EndInvoke will block until RegisterCustomer is
finished.
Asynchronous Callback
Instead of waiting on a handle, you could pass a callback function to BeginInvoke (or a
BeginXXX method).

AsyncCallback cb = new AsyncCallback(CustomerCallback);
object objectState;
IAsyncResult ar;
for(int i = 5; i < 10; i++)

{
firstName = "FirstName" + i.ToString();
lastName = "SecondName" + (i * 2).ToString();
emailAddress = i.ToString() + ".biz";
objectState = i;
ar = rcc.BeginInvoke(firstName, lastName,
emailAddress, cb, objectState);
}
Console.WriteLine
("Finished registrations...could some do work here.");
Thread.Sleep(25);
Console.WriteLine(
"Finished work..waiting to let registrations complete.");
Thread.Sleep(1000);
You then get the results in the callback function:
public void CustomerCallback(IAsyncResult ar)

{
int customerId;
AsyncResult asyncResult = (AsyncResult)ar;
RegisterCustomerCbk rcc =
(RegisterCustomerCbk)asyncResult.AsyncDelegate;
customerId = rcc.EndInvoke(ar);
Console.WriteLine(" AsyncState: {0} CustomerId {1}
added.", ar.AsyncState, customerId);
Console.WriteLine(" Could do processing here.");
return;
}
You could do some work when each customer registration was finished.
Threading with Parameters
The asynchronous callback runs on a different thread from the one on which BeginInvoke
was called. If your threading needs are simple and you want to pass parameters to your
thread functions, you can use asynchronous delegates to do this. You do not need any
reference to the Threading namespace. The reference to that namespace in the
AsynchThreading example is just for the Thread.Sleep method needed for demo
purposes.
PrintNumbers sums the numbers from the starting integer passed to it as an argument to
10 greater than the starting integer. It returns that sum to the caller. PrintNumbers can be
used for the delegate defined by Print.
public delegate int Print(int i);
public class Numbers

{
public int PrintNumbers(int start)
{
int threadId = Thread.CurrentThread.GetHashCode();
Console.WriteLine("PrintNumbers Id: " +
threadId.ToString());
int sum = 0;
for (int i = start; i < start + 10; i++)
{
Console.WriteLine(i.ToString());
Thread.Sleep(500);
sum += i;
}
return sum;
}
}
The Main routine then defines two callbacks and invokes them explicitly with different
starting integers. It waits until the both of the synchronization handles are signaled.
EndInvoke is called on both, and the results are written to the console.
Numbers n = new Numbers();
Print pfn1 = new Print(n.PrintNumbers);

Print pfn2 = new Print(n.PrintNumbers);
IAsyncResult ar1 = pfn1.BeginInvoke(0, null, null);

IAsyncResult ar2 = pfn2.BeginInvoke(100, null, null);
WaitHandle[] wh = new WaitHandle[2];

wh[0] = ar1.AsyncWaitHandle;
wh[1] = ar2.AsyncWaitHandle;
// make sure everything is done before ending

WaitHandle.WaitAll(wh);
int sum1 = pfn1.EndInvoke(ar1);

int sum2 = pfn2.EndInvoke(ar2);
Console.WriteLine("Sum1 = " + sum1.ToString() +

" Sum2 = " + sum2.ToString());
Here is the program's output:
MainThread Id: 2
PrintNumbers Id: 13
0
PrintNumbers Id: 17
100
1
101
2
102
3
103
4
104
5
105
6
106
7
107
8
108
9
109
Sum1 = 45 Sum2 = 1045
Remoting
While a complete discussion of remoting is beyond the scope of this book, a brief
introduction provides a powerful example of how metadata and marshal by reference
(MBR) work. Remoting also provides a mechanism for having executable servers.
Unlike remoting in Microsoft's COM technology, it requires a minimal amount of

infrastructure programming. The small amount that is required allows programmers either
a degree of flexibility or the ability to customize remoting for their particular
applications.
The .NET framework provides two ways to provide connections between two
applications on different computers. Web Services, discussed in Chapter 11, enable
computers that do not host the Common Language Runtime to communicate with
computers that do. The remoting technology discussed here builds distributed
applications between computers that host the CLR.
Remoting Overview
The key parts of Remoting are:
• Communication channels for transport of messages.

• Interception to allow for message generation for communication over the
channels.
• Formatters to put the messages into a byte stream that is sent over the channel.
These are the same formatters that were discussed in the section on serialization.
Interception
Proxies and stubs (referred to in .NET as dispatchers) transform the function calls on the
client or server side into messages that are sent over the network. This is called
interception, because the proxies and dispatchers intercept a method call to send it to its
remote destination. Unlike COM, metadata provides the information so the CLR can
generate the proxies and stubs for you.
A proxy takes the function call off the stackframe of the caller and transforms it into a
message. The message is then sent to its destination. A dispatcher takes the message and
transforms it into a stackframe so that a call can be made to the object.
For example, assume the UnregisterCustomer method from the Customer assembly runs
in one App Domain and is called from another. It makes no difference whether the App
Domains are in the same process or on the same machine.
The proxy would take the integer id argument on the stackframe of the client making the
call and put it in a message that encoded the call and its argument. On the server side, the
dispatcher would take that message and create a function call on the server's stack for the
call UnregisterCustomer (int id) and make that call into the object. The client and server
codes do not know that they are being remoted.
Channels And Formatters
The formatter converts the message into a byte stream. The .NET framework comes with
two formatters, binary and SOAP (text-based XML discussed in Chapter 11 on Web
Services). The byte stream is then sent over a communication channel.
The .NET framework comes with two channels, although you can write your own. The
HTTP channel uses the HTTP protocol and is good for communicating over the Internet
or through firewalls. The TCP channel uses the TCP (sockets) protocol and is designed
for high-speed communication. You have four permutations of formatters and transport:
binary over TCP, binary over HTTP, SOAP over HTTP, and SOAP over TCP.
Remote Objects
Clients obtain a proxy by activating a remote object. Remote objects must derive from
MarshalByRefObject, because you work with a proxy to the object reference, not with
the object reference itself. This is the same concept discussed in the section on contexts,
where marshal by reference is also used to access context bound objects.
Local objects passed as method parameters from one application domain to another can
be passed by value (copied) or by reference.
To be passed by value, they must be serializable. The object is serialized, sent across the
transport layer, and recreated on the other side. We have already seen this in the
AppDomain example.
To be passed by reference, the class must derive from MarshalByRefObject. The

Remoting example illustrates pass by reference.
Remote objects can be either server or client activated. Server-activated objects are not
created until the first method call on the object. Server-activated objects come in two
flavors. SingleCall objects are stateless. Each method cause a new object to be created.
Singleton objects can be used by multiple client activation requests. Singleton objects can
maintain state. SingleCall objects will scale better than Singleton objects because they do
not retain state and can be load balanced.
Client-activated objects are activated when the client requests them. While they can last
for multiple calls and hold state, they cannot store information from different client
activations. This is similar to calling CoCreateInstanceEx in DCOM.
Activation
Objects are activated on the client side in one of three ways by using the Activator class.
• Activator.GetObject is used to get a reference to a server-activated object.

• Activator.CreateInstance is used to create a client-activated object. You can pass
parameters to the object's constructor using the overloaded CreateInstance method
that takes an array of objects to be passed to the constructor.
• The C# new syntax can be used to create a server- or client-activated object. A
configuration file is used to describe how new should be used.
Sample Remotable Object
For our Remoting example, we remote our Customers object from the Customer
assembly.
In the remoting example directory there are two solutions. One represents the client
program, the other the server program. Each can be built independently of the other. Start
the server program first. Notice that it waits for a client request. You can then run the
client program, which will run against objects that live inside the server. We will discuss
the details of the client and server code and output in the next few sections.
Notice that we had to make only two simple changes to our object. The Customers class
in the server project had to be made remotable by inheriting from MarshalByRefObject.
public class Customers : MarshalByRefObject, ICustomer
The CustomerListItem that was going to be transferred by value had to be made

serializable.
[Serializable]
{
}
Sample Remoting Program
In the Remoting example the client accesses a server-activated object. The server is the
TcpServerChannel class that uses using a binary format with the TCP protocol. The
channel will use port 8085. The server registers the type being remoted, the endpoint
name to refer to this object, and the type of activation. The server then waits for client
requests.
TcpServerChannel chan = new TcpServerChannel(8085);

ChannelServices.RegisterChannel(chan);
RemotingConfiguration.RegisterWellKnownServiceType(
typeof(Customers), "AcmeCustomer",
WellKnownObjectMode.Singleton);
. . .
The server has to be started before the client program can access the object.
The client sets up a TcpClientChannel object and then connects to the object. In the
Activator.GetObject method call it specifies the type of the object it wants, and the
endpoint where the server is listening to for object requests. If you want to run the client
and server on separate machines, substitute the server machine name for localhost in the
endpoint. Unlike COM location transparency, the client has to specify a specific
endpoint; there is no redirection through an opaque registry entry.
TcpClientChannel chan = new TcpClientChannel();

ChannelServices.RegisterChannel(chan);
Customers obj = (Customers)Activator.GetObject(
typeof(Customers),
"tcp://localhost:8085/AcmeCustomer");
if (obj == null) System.Console.WriteLine(
"Could not locate server");
else
. . .
The client then uses the proxy to make calls on the object as if it were a local instance.
bool bRet = RemotingServices.IsTransparentProxy(obj);

. . .
ArrayList ar;
ar = obj.GetCustomer(-1);
ShowCustomerArray(ar);
obj.RegisterCustomer("Boris", "Badenough",
ar = obj.GetCustomer(-1);
ShowCustomerArray(ar);
To run the program, start the server program in one console window, then run the client
program from another console window.
The output depends on what kind of server-activated object is being activated. [13] If the
server activation type is Singleton, which supports the maintaining state, you get the
behavior you would expect from the nonremoted case. A new customer is added, and you
find that new customer in the list when you ask for all the existing customers. As you
would expect, the initial activate call results in the Customers constructor being called
once for each server invocation, no matter how many times the client program is run.
[13]
In the example, you can try out both Singleton and SingleCall activation by
commenting out the appropriate line in the code in server.cs.
Object reference is a proxy?: True

Client: AppDomain Client.exe Thread 19 Context 0

3 Boris Badenough [email protected]
If the activation type is SingleCall, which creates a new object instance for every method
call, the results are quite different. Four different objects are created. The first is created
by the initial activate request. The second is created by the initial call to GetCustomer.
The third is created by the RegisterCustomer call. The fourth is created by the second call
to GetCustomer. The last object created never sees the new customer, because no state is
saved. Note that the static nextCustId member of the Customer class is treated as a static
with respect to the new object instances of the Customer class, just as you would expect.
Same client code, different results! Since the object is already activated, if you run the
client program a second time for the same server invocation, the Customers constructor
will be called only three times.
Object reference a proxy?: True

Client: AppDomain Client.exe Thread 19 Context 0

Since the client uses a proxy, the object executes inside the server's application domain,
but on a different thread than the main server thread. The object's constructor is not called
until the first method call on the object. Notice how in both cases we have remoted an
ArrayList of types without any special work aside from making the type serializable. The
presence of metadata makes the programmer's work much easier.
Metadata and Remoting
In order for the client to request an object of a specific type, metadata about the type has
to be available to the client. For some applications, a reference can be made to the actual
assembly where the object is stored.
For many applications, however, you do not want to give the client access to your source
code. For the metadata that the client needs, a reference need only be made to an object
without the implementation details.
One way to do this is to build a version of the object that has methods with no
implementation. This interface class can then be built into an assembly that can be given
to the client. You can throw the System. NotSupportedException in the methods if you
wish to make sure that it is never used by mistake for the real object.
[System.Serializable]
{
}
. . .
public class Customers : MarshalByRefObject, ICustomer
{
{
throw new NotSupportedException();
}
public void UnregisterCustomer(int id)
{
}
{
}
public ArrayList GetCustomer(int id)
{
}
}
For Web Services you use the SOAPSUDS tool to extract the metadata from the service,
and then generate an assembly that has the required metadata. You can then build a proxy
DLL and have the client program refer to it. This is conceptually equivalent to the first
approach.
The server, of course, has to reference the real object's assembly.
Unlike the COM model, there is no reference counting, interface negotiation, building
and registering separate proxies and stubs, worrying about global identifiers, or use of the
registry. Because of metadata, all you have to do is inherit from MarshalByRefObject to
make an object remotable.
Remoting Configuration Files
You use configuration files to define where the object is activated. The client can then use
the new operator to create the object. The big advantage here is that as the object location
changes (such as a URL or TCP channel), or the formatter you want to use changes, the
client does not have to be rebuilt.
Multiple classes can be configured on the client. Configuration files are loaded into the
client using the RemotingConfiguration. Configure method.
Custom Attributes
Chapter 5 introduced the concept of attributes, which have already appeared in several
examples. In this chapter we used the Serializable and Synchronization attributes, which
are provided by .NET Framework classes. The .NET Framework makes the attribute
mechanism entirely extensible, allowing you to define custom attributes, which be added
to the class's metadata. This custom metadata is available through reflection and can be
used at runtime. To simplify the use of custom attributes, you may declare a base class to
do the work of invoking the reflection API to obtain the metadata information.
The example AttributeCustom illustrates the custom attribute InitialDirectory.

InitialDirectory controls the initial current directory where the program runs. By default
the current directory is the directory containing the program's executable. In the case of a
Visual Studio C# project, built in Debug mode, this directory is bin\Debug, relative to the
project source code directory.
Using a Custom Attribute
Before we discuss implementing the custom attribute, let us look at how the
InitialDirectory attribute is used. To be able to control the initial directory for a class, we
derive the class from the base class DirectoryContext. We may then apply to the class the
attribute InitialDirectory, which takes a string parameter giving a path to what the initial
directory should be. The property DirectoryPath extracts the path from the metadata. If
our class does not have the attribute applied, this path will be the default. Here is the code
for our test program.
When you run this sample on your system, change the directory in the attribute to one
that exists on your machine.
// AttributeCustom.cs
using System;
using System.IO;
class Normal : DirectoryContext

{
}
[InitialDirectory(@"\OI\NetCs\Chap08")]
class Special : DirectoryContext
{
}
public class AttributeCustom

{
{
Normal objNormal = new Normal();
Console.WriteLine("path = {0}",
objNormal.DirectoryPath);
ShowDirectoryContents(objNormal.DirectoryPath);
Special objSpecial = new Special();
Console.WriteLine("path = {0}",
objSpecial.DirectoryPath);
ShowDirectoryContents(objSpecial.DirectoryPath);
}
private static void ShowDirectoryContents(string path)
{
DirectoryInfo dir = new DirectoryInfo(path);
FileInfo[] files = dir.GetFiles();
Console.WriteLine("Files:");
foreach (FileInfo f in files)
Console.WriteLine(" {0}", f.Name);
DirectoryInfo[] dirs = dir.GetDirectories();
Console.WriteLine("Directories:");
foreach (DirectoryInfo d in dirs)
Console.WriteLine(" {0}", d.Name); }
}
Here is the output:
path = C:\OI\NetCs\Chap08\AttributeCustom\bin\Debug
Files:
AttributeDemo.exe
AttributeDemo.pdb
Directories:
path = c:\OI\NetCs\Chap8
Files:
Directories:
AppDomain
Asynch
AsynchThreading
AttributeCustom
DynamicInvocation
FileIO
ISerializable
MarshalByReference
PulseAll
Reflection
Remoting
Serialization
Threading
ThreadIsolation
Defining an Attribute Class
To create a custom attribute, you must define an attribute class, derived from the base
class Attribute. By convention give your class a name ending in "Attribute." The name of
your class without the "Attribute" suffix will be the name of the custom attribute. In our
example the class name is InitialDirectoryAttribute, so the attribute's name is
InitialDirectory.
You may provide one or more constructors for your attribute class. The constructors
define how to pass positional parameters to the attribute (provide a parameter list,
separated by commas). It is also possible to provide "named parameters" for a custom
attribute, where the parameter information will be passed using syntax name = value.
You may also provide properties to read the parameter information. In our example, we
have a property Path, which is initialized in the constructor.
// DirectoryAttribute.cs
using System;
public class InitialDirectoryAttribute : Attribute

{
private string path;
public InitialDirectoryAttribute(string path)
{
this.path = path;
}
public string Path
{
get
{
return path;
}
}
}
Defining a Base Class
The last step in working with custom attributes is to provide a means to extract the
custom attribute information from the metadata using the reflection classes. You can
obtain the Type of any object by calling the method GetType, which is provided in the
root class object. Using the class's method GetCustomAttributes you can read the custom
attribute information.
To make the coding of the client program as simple as possible, it is often useful to
provide a base class that does the work of reading the custom attribute information. [14]
We provide a base class DirectoryContext, which is used by a class wishing to take
advantage of the InitialDirectory attribute. This base class provides the property
DirectoryPath to return the path information stored in the metadata. Here is the code for
the base class:
[14]
With single implementation inheritance there is a cost to providing a base class. If you
need to derive from another class such as ContextBoundObject, the base class has to
derive from that class.
// DirectoryContext.cs
using System;
using System.Reflection;
using System.IO;
public class DirectoryContext

{
virtual public string DirectoryPath
{
get
{
Type t = this.GetType();
foreach (Attribute a
in t.GetCustomAttributes(true))
{
InitialDirectoryAttribute da =
a as InitialDirectoryAttribute;
if (da != null)
{
return da.Path;
}
}
return Directory.GetCurrentDirectory();
}
}
}
We must import the System.Reflection namespace because GetType returns the current
Type of the object. GetCustomAttributes method can then obtain a collection of Attribute
objects from the metadata. Since this collection is heterogeneous, consisting of different
types, the C# as operator is used to test whether a given collection element is of the type
InitialDirectoryAttribute. If we find such an element, we return the Path property.
Otherwise, we return the default current directory, obtained from GetCurrentDirectory.
Garbage Collection and Finalization
Memory management is a critical aspect of programming and can be the source of many
errors. Whenever a resource is created, memory must be provided for it. And when the
resource is no longer needed, the memory should be reclaimed. If the memory is not
reclaimed, the amount of memory available is reduced. If such "memory leaks" recur
often enough (which can happen in long-running server programs), the program can
crash. Another potential bug is to reclaim memory while it is still required by another part
of the program.
.NET greatly simplifies the programming of memory management through an automatic

garbage collection facility. The CLR tracks the use of memory that is allocated on the
managed heap, and any memory that is no longer referenced is marked as "garbage."
When memory is low, the CLR traverses its data structure of tracked memory and
reclaims all the memory marked as garbage. Thus the programmer is relieved of this
responsibility.
Although a good foundation for resource management, garbage collection by itself does
not address all issues. Memory allocated from the managed heap is not the only kind of
resource needed in programs. Other resources, such as file handles and database
connections, are not automatically deallocated, and the programmer may need to write
explicit code to perform cleanup. The .NET Framework provides a Finalize method in the
Object base class for this purpose. The CLR calls Finalize when the memory allocated for
an object is reclaimed.
Another concern with garbage collection is performance. Is there a big penalty from the
automated garbage collection? The CLR provides a very efficient multigenerational
garbage collection algorithm. In this section we examine garbage collection and
finalization in the .NET Framework, and we provide several code examples.
Finalize
System.Object has a protected method Finalize, which is automatically called by the CLR
after an object becomes inaccessible. (As we shall see, finalization for an object may be
suppressed by a call to the method SuppressFinalize of the System.GC class.) Since
Finalize is protected, it can only be called through the class or a derived class. The default
implementation of Finalize does nothing. For any cleanup to be performed, a class must
override Finalize. Also, a class's Finalize implementation should call the Finalize of its
base class.
C# Destructor Notation
The C# language provides a special tilde notation ~SomeClass to represent the
overridden Finalize method, and this special method is called a destructor. The C#
destructor automatically calls the base class Finalize. Thus the following C# code
~SomeClass()
{
// perform cleanup
}
generates code that could be expressed
protected override void Finalize()

{
// perform cleanup
base.Finalize();
}
The second code fragment is actually not legal C# syntax, and you must use the
destructor notation.
Although C# uses the same notation and terminology for destructor as C++, the two are
very different. The C++ destructor is called deterministically when a C++ object goes out
of scope or is deleted. The C# destructor is called during the process of garbage
collection, a process which is not deterministic, as discussed below.
Limitations of Finalization
Finalization is nondeterministic. Finalize for a particular object may run at any time
during the garbage collection process, and the order of running finalizers for different
objects cannot be predicted. Moreover, under exceptional circumstances a finalizer may
not run at all (for example one finalizer goes into an infinite loop, or a process aborts
without giving the runtime a chance to clean up).
Also, the thread on which a finalizer runs is not specified.
Another issue with finalization is its effect on performance. There is significantly more
overhead associated with managing memory for objects with finalizers, both on the
allocation side and on the deallocation side. [15]
[15]
Finalization internals and other details of garbage collection are discussed in depth in
the two-part article "Garbage Collection" by Jeffrey Richter, MSDN Magazine,
November and December 2000.
Thus you should not implement a finalizer for a class unless you have very good reason
for doing do. And if you do provide a finalizer, you should probably provide an alternate,
deterministic mechanism for a class to perform necessary cleanup. The .NET Framework
provides a Dispose design pattern for deterministic cleanup.
Unmanaged Resources and Dispose
The classic case for a finalizer is a class that contains some unmanaged resource, such as
a file handle or a database connection. If they are not released when no longer need, the
scalability of your application can be affected. As a simple illustration, consider a class
that wraps a file object. We want to make sure that a file that is opened will eventually be
closed. The object itself will be destroyed by garbage collection, but the unmanaged file
will remain open, unless explicitly closed. Hence we provide a finalizer to close the
wrapped file.
But as we discussed, finalization is nondeterministic, so a file for a deleted object might

hang around open for a long time. We would like to have a deterministic mechanism for a
client program to clean up the wrapper object when it is done with it. The .NET
Framework provides the generic IDisposable interface for this purpose.
public interface IDisposable

{
void Dispose();
};
The design pattern specifies that a client program should call Dispose on the object when
it is done with it. In the Dispose method implementation, the class does the appropriate
cleanup. As backup assurance, the class should also implement a finalizer, in case
Dispose never gets called, perhaps due to an exception being thrown. [16] Since both
Dispose and Finalize perform the cleanup, cleanup code can be placed in Dispose, and
Finalize can be implemented by calling Dispose. One detail is that once Dispose has been
called, the object should not be finalized, because that would involve cleanup being
performed twice. The object can be removed from the finalization queue by calling
GC.SuppressFinalize. Also, it is a good idea for the class to maintain a boolean flag such
as disposedCalled, so that if Dispose is called twice, cleanup will not be performed a
second time.
[16]
One of the virtues of the exception handling mechanism is that as the call stack is
unwound in handling the exception, local objects go out of scope and so can get marked
for finalization. We provide a small demo later in this section.
The example program DisposeDemo provides an illustration of finalization and the

dispose pattern. The class SimpleLog implements logging to a file, making use of the
StreamWriter class (discussed earlier in this chapter).
// SimpleLog.cs
using System;
using System.IO;
public class SimpleLog : IDisposable
{
private StreamWriter writer;
private bool disposeCalled = false;
public SimpleLog(string fileName)
{
name = fileName;
writer = new StreamWriter(fileName, false);
writer.AutoFlush = true;
Console.WriteLine("logfile " + name + " created");
}
public void WriteLine(string str)
{
writer.WriteLine(str);
}
public void Dispose()
{
if(disposeCalled)
return;
writer.Close();
GC.SuppressFinalize(this);
Console.WriteLine("logfile " + name + " disposed");
disposeCalled = true;
}
~SimpleLog()
{
Console.WriteLine("logfile " + name + " finalized");
Dispose();
}
}
The class SimpleLog supports the IDisposable interface, and thus implements Dispose.
The cleanup code simply closes the StreamWriter object. To make sure that a disposed
object will not also be finalized, GC.SuppressFinalize is called. The finalizer simply
delegates to Dispose. To help monitor object lifetime, a message is written to the console
in the constructor, in Dispose, and in the finalizer. [17]
[17]
The Console.WriteLine in the finalizer is provided purely for didactic purposes and
should not be done in production code, for reasons we shall discuss shortly.
Here is the code for the test program:
// DisposeDemo.cs
using System;
using System.Threading;
public class DisposeDemo

{
{
SimpleLog log = new SimpleLog(@"log1.txt");
log.WriteLine("First line");
Pause();
log.Dispose();
log.Dispose();
log = new SimpleLog(@"log2.txt");
log.WriteLine("Second line");
Pause();
log = new SimpleLog(@"log3.txt");
log.WriteLine("Third line");
Pause();
log = null;
GC.Collect();
Thread.Sleep(100);
}
private static void Pause()
{
Console.Write("Press enter to continue");
string str = Console.ReadLine();
}
}
The SimpleLog object reference log is assigned in turn to three different object instances.
The first time, it is properly disposed. The second time, log is reassigned to refer to a
third object, before the second object is disposed, resulting in the second object becoming
"garbage." The Pause method provides an easy way to pause the execution of this console
application, allowing us to investigate the condition of the files log1.txt, log2.txt, and
log3.txt at various points in the execution of the program.
Running the program results in the following output:
logfile log1.txt created

First line
Press enter to continue
logfile log1.txt disposed
Second line
Third line
logfile log3.txt finalized
logfile log2.txt finalized
After the first pause, the file log1.txt has been created, and you can examine its contents
in Notepad. If you try to delete the file, you will get a sharing violation, as illustrated in
Figure 8-2.
Figure 8-2. Trying to delete an open file results in a sharing violation.
At the second pause point, log1.txt has been disposed, and you will be allowed to delete
it. log2.txt has been created (and is open). At the third pause point, log3.txt has been
created. But the object reference to log2.txt has been reassigned, and so there is now no
way for the client program to dispose of the second object. [18] If Dispose were the only
mechanism to cleanup the second object, we would be out of luck. Fortunately, the
SimpleObject class has implemented a finalizer, so the next time garbage is collected, the
second object will be disposed of properly. We can see the effect of finalization by
running the program through to completion. The second object is indeed finalized, and
thence disposed. In fact, as the app domain shuts down, Finalize is called on all objects
not exempt from finalization, even on objects that are still accessible.
[18]
This example illustrates that it is the client's responsibility to help the scalability of the
server by cleaning up objects (using Dispose) before reassigning them. Once an object
has been reassigned, there is no way to call Dispose, and the object will hang around for
an indeterminate period of time until garbage is collected. Effective memory management
involves both the server and client.
In our code we explicitly make the third object inaccessible by the assignment log = null,
and we then force a garbage collection by a call to GC.Collect. Finally we sleep briefly,
to give the garbage collector a chance to run through to completion, before the
application domain shuts down. Coding our test program in this way is a workaround for
the fact that the order of finalization is nondeterministic. The garbage collector will be
called automatically when the program exits and the application domain is shut down.
However, at that point, system objects, such as Console, are also being closed. Since you
cannot rely on the order of finalizations, you may get an exception from the WriteLine
statement within the finalizer. The explicit call to GC.Collect forces a garbage collection
while the system objects are still open. If we omitted the last three lines of the Main
method, we might well get identical output, but we might also take an exception.
We provide similar code at the end of the Main methods of our other test programs, so
that our print statements in finalizers work properly without randomly throwing
exceptions.
Alternate Name for Dispose

The standard name for the method that performs cleanup is Dispose. The convention is
that once an object is disposed, it is finished. In some cases, the same object instance may
be reused, as in the case of a file. A file may be opened, closed, and then opened again. In
such a case, an additional cleanup method should be called Close. In other cases some
other natural name may be used.
Our SimpleLog class could plausibly have provided an Open method, and then it would
have made sense to name our cleanup method Close. For simplicity, we did not provide
an Open method, and so we stuck to the name Dispose.
Garbage Collection and Generations
Using the dispose pattern we can mitigate the issue of nondeterministic finalization, but
what about the performance of the garbage collector? It turns out that the overall memory
management efficiency of .NET is quite good, thanks to two main points:
• Allocation is very fast. Space on the managed heap is always contiguous, so

allocating a new object is equivalent to incrementing a pointer. (By contrast, an
allocation on an unmanaged heap is relatively slow, because a list of data
structures must be walked to find a block that is large enough.)
• The CLR uses generations during garbage collecting, reducing the number of
objects that are typically checked for being garbage.
Generations
As an optimization, every object on the managed heap is assigned to a generation. A new

object is in generation 0 and is considered a prime candidate for garbage collection. Older
objects are in generation 1. Since such an older object has survived for a while, the odds
favor its having a longer lifetime than a generation 0 object. Still older objects are
assigned to generation 2 and are considered even more likely to survive a garbage
collection. The maximum generation number in the current implementation of .NET is 2,
as can be confirmed from the GC.MaxGeneration property.
In a normal sweep of the garbage collector, only generation 0 will be examined. It is here
that the most likely candidates are for memory to be reclaimed. All surviving generation
0 objects are promoted to generation 1. If not enough memory is reclaimed, a sweep will
next be performed on generation 1 objects, and the survivors will be promoted. Then, if
necessary, a sweep of generation 2 will be performed, and so on up until MaxGeneration.
Finalization and Stack Unwinding
As mentioned earlier, one of the virtues of the exception handling mechanism is that as
the call stack is unwound in handling the exception, local objects go out of scope and so
can get marked for finalization. The program FinalizeStackUnwind provides a simple
illustration. It uses the SimpleLog class discussed previously, which implements
finalization.
// FinalizeStackUnwind.cs
using System;
public class FinalizeStackUnwind

{
{
try
{
SomeMethod();
}
catch(Exception e)
{
}
GC.Collect();
}
private static void SomeMethod()
{
// local variable
SimpleLog alpha = new SimpleLog("alpha.txt");
// force an exception
throw new Exception("error!!");
}
}
A local variable alpha of type SimpleLog is allocated in SomeMethod. Before the method
exits normally, an exception is thrown. The stack unwinding mechanism of exception
handling detects that alpha is no longer accessible, and so is marked for garbage
collection. The call to GC.Collect forces a garbage collection, and we see from the output
of the program that finalize is indeed called.
logfile alpha.txt created

error!!
logfile alpha.txt finalized
logfile alpha.txt disposed
Controlling Garbage Collection with the GC Class
Normally it is the best practice simply to let the garbage collector perform its work
behind the scenes. Sometimes, however, it may be advantageous for the program to
intervene. The System namespace contains the class GC, which enables a program to
affect the behavior of the garbage collector. We summarize a few of the important
methods of the class.
SuppressFinalize
This method requests the system to not call Finalize for the specified object. As we saw
previously, you should call this method in your implementation of Dispose, to prevent a
disposed object from also being finalized. [19]
[19]
You should be careful in the case of an object that might be "closed" (like a file) and
later reopened again. In such a case it might be better not to suppress finalization. Once
finalization is suppressed, it can be made eligible for finalization again by calling GC.
ReRegisterForFinalize. For a discussion of advanced issues in garbage collection and
finalization, refer to the Jeffrey Richter article previously cited.
Collect
You can force a garbage collection by calling the Collect method. An optional parameter
lets you specify which generations should be collected. Use this method sparingly, since
normally the CLR has better information on the current state of memory. A possible use
would be a case when your program has just released a number of large objects, and you
would like to see all this memory reclaimed right away. Another example was provided in
the previous section, where a call to Collect forced a collection while system objects were
still valid.
MaxGeneration
This property returns the maximum number of generations that are supported.
GetGeneration
This method returns the current generation number of an object.
GetTotalMemory
This method returns the number of bytes currently allocated. A parameter lets you specify
whether the system should perform a garbage collection before returning. If no garbage
collection is done, the indicated number of bytes is probably larger than the actual
number of bytes actually being used by live objects.
Sample Program
The program GarbageCollection illustrates using these methods of the GC class. The
example is artificial, simply illustrating object lifetime, and the effect of the various GC
methods. The class of objects that are allocated is called Member. This class has a string
property called Name. Write statements are provided in the constructor, Dispose, and in
the destructor. A Committee class maintains an array list of Member instances. The
RemoveMember method simply removes the member from the array list. The
DisposeMember method also calls Dispose on the member being expunged from the
committee. The ShowGenerations method displays the generation number of each
Member object. GarbageCollection.cs is a test program to exercise these classes, showing
the results of various allocations and deallocations and the use of methods of the GC
class. The code and output should be quite easy to understand.
All the memory is allocated locally in a method Demonstrate Generations. After this
method returns and its local memory has become inaccessible, we make an explicit call to
GC.Collect. This forces the finalizers to be called before the app domain shuts down, and
so we avoid a possible random exception of a stream being closed when a WriteLine
method is called in a finalizer. This is the same point mentioned previously for the earlier
examples.
Summary
This chapter introduced the .NET application model. Through metadata, the framework
can understand enough about your application to provide many services that you do not
have to implement. On the other hand, we have seen how the framework is structured so
that you can substitute your own objects and implementations where needed.
Type safety enables application domains to provide a robust, yet cheap, form of
application isolation. Contexts, proxies, and interception allow the runtime to
transparently provide services to parts of applications that require them.
Another aspect of the .NET application model is the pervasive use of attributes, which
can be easily added to source code and are stored with the metadata. We saw examples of
the use of attributes for serialization and for synchronization, and we demonstrated how
to implement and use custom attributes.
NET simplifies the programming of memory management through an efficient

generational automatic garbage collection facility. Finalization is nondeterministic, but
you can support deterministic cleanup by implementing the dispose pattern.
Chapter 9. Programming with ADO.NET Top
The framework database programming classes are referred to as ADO.NET. ADO.NET

introduces the DataSet class that works with relational data in a relational manner while
you are disconnected from any data source. You need not connect and update or query the
database unless you have a specific reason for doing so. You can, of course, work in the
traditional connected manner if you choose.
ADO.NET data providers [1] allow you to execute commands directly against the data
source. Functionality is exposed directly without intermediary objects such as OLEDB,
which stands between ADO and the data source. The .NET DataAdapter class models a
data source as a set of database commands and a connection to that data source.
Differences between data sources are not hidden by generic interfaces. The OLEDB data
provider allows for nested transactions with data sources that support that functionality;
the SqlServer data provider does not. [2]
[1]
.NET Data Providers are what used to be called in the beta literature managed
providers. You may still see them referred to by that term.
[2]
There is a Begin method on the OleDbTransaction class; the SqlTransaction class does
not have such a method.
.NET Data Providers supply data to a dataset or a data reader. A dataset is a memory-
resident, lightweight relational database that is not connected to any database. You can
also obtain a dataset from an XML document or create an XML document from a dataset.
This allows you to work, if it makes sense, with your data as relational data, or as
hierarchical XML data. [3] Data readers model the traditional method of working with a
database.
[3]
The many-to-many relations that you can have in a relational database do not
automatically map to XML hierarchies. But this is no different from working with the
classic object-relational model clash.
The data access classes that currently ship with the framework are found in the
namespaces: System.Data, System.Data.SqlClient, System.Data.OleDb,
System.Data.Common, and System.Data.SqlTypes. The Sql and OleDb namespaces
reflect the SqlServer and OleDb .NET Data Providers. An ODBC .NET Data Provider
has been written, and additional ones will be written in the future.
This chapter changes the implementation of the Customer and Hotel assemblies of the
Case Study to use SQL Server. An air travel service that the Acme Travel Agency can use
to make air travel reservations is added to illustrate the use of XML.
To make our examples concrete we use SQL Server 2000 and the SQL Server data
provider. [4] Nonetheless, much of the basic functionality discussed in this chapter applies
to the OleDb data provider as well.
[4]
If you do not have an SQL Server available you can go to the Microsoft site and
download the MSDE, which is a scaled down version of SQL Server. As of this writing
MSDE is available for free. Microsoft suggests using MSDE in the future instead of
Access. Since we use vanilla functionality you should be able to use the OleDb data
provider against the Access version of the Northwind traders by changing the Sql classes
to the corresponding OleDb classes. We have not yet tested this scenario, however.
This chapter assumes you have some understanding of database concepts.
Setting Up the Example Databases

This chapter assumes that SQL Server 2000 has been installed using the Local
System account, with authentication mode set to Mixed Mode. The user is
assumed to be sa, with a blank password.
Several examples in this chapter make use of the Northwind Traders sample
database, which is installed along with SQL Server. In addition, there are other
example programs that use the HotelBroker and AirlineBroker databases, which
are supplied specifically for use with this book.
Some of the example programs make changes to these databases, and other
examples assume a freshly installed database. This means that some of the
examples will not always work as expected unless you reinstall them again. You
can reinstall each of these databases by running the SQL scripts that are
provided.
Please refer to the readme.txt file in the sample code directory for this chapter for
more information about database setup.
.NET Data Providers
The prefix on the database classes and methods indicates the data provider used to access
the data source. For example, the OleDb prefix applies to the OleDb data provider. The
Sql prefix applies to the SqlServer data provider.
The SqlServer data provider uses the native SQL Server wire protocol. The OleDb data
provider goes through the COM interop layer to talk to the various OleDb providers. For
example, you could talk to SqlServer through the OleDb data provider to the OLEDB
provider for SQL Server. Nonetheless, the performance of going through the SqlServer
data provider will be superior. The advantage of the OleDb and the ODBC data providers
is that you can work with ADO.NET against most data sources that you work with today.
There are some interfaces that define common functionality, and some base classes that
can be used to provide common functionality, but there is no requirement for a data
provider to fit a specification that does not correspond to the way the underlying data
source works.
For example, the SqlDataAdapter class and the OleDbDataAdapter class both use the
abstract base classes DbDataAdapter and DataAdapter that are found in the
System.Data.Common namespace. SqlTransaction and the OleDbTransaction classes
both implement the IDbTransaction interface. The OleDbError class and the SqlError
class do not resemble each other at all. Server-side cursors are not in the ADO.NET
model because some databases (such as Oracle and DB2) do not have native support for
them. Any support for them in the SQL Server data provider would be as an extension. [5]
[5]
Besides, server-side cursors are rarely appropriate so it is not surprising that databases
do not support them. Scrolling through the output is usually the result of a user
interaction. Holding state on the server while the user interacts with the data is not the
way to build a scalable application.
As Table 9-1 shows, the Connection, Command, DataReader, DataAdapter, and

DataParameter classes of the data providers do have some parallels that are defined by
the IDbConnection, IDbCommand, IDataReader, IDbDataAdapter, and IDataParameter
interfaces. Nothing, of course, prevents an implementation of these classes from having
additional methods beyond those specified in the interfaces.
Table 9-1. Comparison of Parallel Classes in the OleDb and SQL Server Data
Providers
Interface OleDb SQL Server
IDbConnection OleDbConnection SqlConnection
IDbCommand OleDbCommand SqlCommand
IDataReader OleDbDataReader SqlDataReader
IDbDataAdatpter OleDbDataAdapter SqlDataAdapter
IDbTransaction OleDbTransaction SqlTransaction
IDataParameter OleDbDataParameter SqlDataParameter
Classes such as the DataSet or the DataTable, which are independent of any data
provider, do not have any prefix.
If database scalability is important, it is important to suppress finalization on any of your

classes that do not require it. You will get vastly improved performance because you will
reduce the amount of time the finalizer thread runs.
The Visual Studio.NET Server Explorer
Visual Studio.NET Server Explorer is a very useful tool for working with databases.
While not as powerful as the SQL Server Enterprise Manager, it can give you the basic
functionality you need when writing or debugging database applications. It will be very
useful when we work with the examples in this chapter.
To access the Server Explorer, use the View | Server Explorer menu item. The Server
Explorer is a dockable window that can be moved around as required. Figure 9-1
illustrates the Server Explorer.
Figure 9-1. Visual Studio.NET Server Explorer window.

You can find information about all the fields in a table, or look at and edit the data in the
tables. You can create or edit stored procedures and design tables. We will use the Server
Explorer in the first few examples to show a little bit of how it can be used.
Data Readers
To make these concepts concrete, let us use some of these classes to access a database.
This example is found in this chapter's Connected subdirectory.
We will need a connection to the database, a command to issue against the database, and
a reader to retrieve the data, so we declare three objects: [6]
[6]
The objects are declared outside of the try/catch block so that they can be used in the
finally block. They are set to null because their use in the finally block could theoretically
occur before they are initialized inside the try block (and so the compiler indicates with a
warning).
SqlConnection conn = null;

SqlCommand command = null;
SqlDataReader reader = null;
The connection string that is used to connect to the database is set next. You will have to
replace the server element with the appropriate value for your machine. You will also
have to specify the appropriate user and password for the database. [7] You can also set the
connect string as an SqlConnection property. A simple select statement will be the
command run against the database.
[7]
For Northwind the usual uid=sa;pwd=; will probably work.
String ConnString =
"server=localhost;uid=sa;pwd=;database=Northwind";
string cmd = "select CustomerId, CompanyName from
Customers";[8]
[8]
Here and several other places long SQL statement strings are broken up and placed on
multiple lines to allow for book formatting.
Figure 9-2 shows the tables and stored procedures for the Northwind database.
Figure 9-2. Server Explorer showing Northwind database tables and stored
procedures.
We create an SqlConnection object and then a create an SqlCommand object that is
attached to the connection just created. A connection to the database must be opened
before any command can be executed.
conn = new SqlConnection(ConnString);

command = new SqlCommand(cmd, conn);
conn.Open();
When the command is executed using the ExecuteReader method on the SqlCommand
object, an SqlDataReader instance is returned. This reader can be used to iterate through
the result set. The column names can be used to fetch the data elements from the current
result set row.
reader = command.ExecuteReader();
if (reader != null)
{
Console.WriteLine("CustomerId\tCompanyName");
while (reader.Read())
Console.WriteLine("{0}\t\t{1}", reader["CustomerId"],
reader["CompanyName"]);
}
The reader and the connection are closed in the finally block.
if (reader != null) reader.Close();

if (conn.State == ConnectionState.Open) conn.Close();
If the connection is not closed explicitly, the finalizer on the SqlConnection object will
eventually get called and the connection will be closed. Since the garbage collector is not
deterministic, there is no way to know when this will happen. Therefore always close
your connections. If you do not, you will use more connections than you need (even with
connection pooling), and this could interfere with your applications scalability. You could
also run out of connections.
Here is the output the program produces:
CustomerId CompanyName
ALFKI Alfreds Futterkiste
ANATR Ana Trujillo Emparedados y helados
ANTON Antonio Moreno Taquería
AROUT Around the Horn
BERGS Berglunds snabbköp
BLAUS Blauer See Delikatessen
BLONP Blondesddsl père et fils
BOLID Bólido Comidas preparadas
BONAP Bon app'
BOTTM Bottom-Dollar Markets
BSBEV B's Beverages
...
You use the Visual Studio.NET Server Explorer to check the results of the program.
Select the Customers table under the Northwind database explorer and right-click to get a
context menu. Select "Retrieve Data from Table," and you can retrieve the data associated
with the table and compare it with the results of the program. You will see that they are
the same. Figure 9-3 shows this.
Figure 9-3. Server Explorer showing Customers table, fields, and data.
The Connected Database Scenario
This scenario of working with a database is referred to as connected. The program

connects to the database, does the work it needs to do, and then disconnects. You can run
through the returned data only in the forward direction. This corresponds to the classic
ADO forward-only cursor/recordset. In the connected mode you must open and close the
database connection explicitly.
Keeping a connection continually open is not the best way to work in an environment
where you want to minimize the resources consumed (connections are expensive) to
allow for scalability. A DataSet allows you to work disconnected from a data source.
Nonetheless, as will be discussed later, depending on your concurrency assumptions,
using a DataReader instead of a DataSet might still be the right approach.
As will be discussed later, the SqlConnection is used with the DataAdapter to establish
connections with the database in the same way as illustrated here with the SqlCommand.
SqlConnection also controls database properties such as transactions and isolation levels.
A root transaction is issued by invoking the BeginTransaction method on the
SqlConnection class. [9] If in the previous example we connected SQL Server through the
OleDbConnection class, the connection string would be:
[9]
Since OLEDB allows for nested transactions, nested transactions can be started by
invoking the Begin method on the OleDbTransaction class.
"Provider=SQLOLEDB.1;server=localhost;uid=sa;pwd=;
database=Northwind";
You would have to provide the correct server, user, and password. While the
SqlCommand executes a command against a database in the same way whether you use a
DataAdapter or an SqlDataReader, the mechanics of doing so is different. This will
become clearer when we discuss the SqlDataAdapter class.
You specify the type of SqlCommand with the CommandType property. For the Sql data
provider this can be either Text (the default) or StoredProcedure. The CommandText can
also be specified as a property. We will soon show how parameters can be applied to
database commands.
An SqlDataReader instance is returned by the ExecuteReader method on an

SqlCommand instance. If you wanted to program in a way that was independent of a data
provider, you could use the IDataReader interface instead. You could then invoke
methods on the interface instead of an object.
IDataReader idr = command.ExecuteReader();
Similar techniques can be used with the other data-provider classes that implement
interfaces used by multiple data providers. Until the SqlDataReader instance is closed,
the SqlCommand object cannot be used for any purpose other than for executing its Close
method.
Executing SQL Statements
The ExecuteReader method on the SqlCommand returns a DataReader instance. Data is

returned when the command is a select statement or a stored procedure that returns
results. When you know there will be no results returned it is more efficient to use the
ExecuteNonQuery method. The SqlCommand.ExecuteReader method uses the stored
procedure sp_executesql. Some commands that use "SET" statements may not work
properly. Other providers might have different restrictions when their ExecuteReader
method is used.
In general, for commands that do not return data, use the

SqlCommand.ExecuteNonQuery method. The NonQuery example shows how this works.
For illustrative purposes this example connects to SQL Server through the OleDb data
provider.
string cmd = "update Customers set CompanyName =

'Maria Inc' where ContactName = 'Maria Anders'";
...
command = new OleDbCommand(cmd, conn);
int NumberRows = command.ExecuteNonQuery();
The number of rows returned should be 1. Figure 9-4 shows the results of the change to
the first row.
Figure 9-4. Rows in Customers table in the Server Explorer showing the
changed ContactName. Compare with Figure 9-3 to see the original value of the
first row.
For insert, update, and delete statements, the number of rows affected is returned. SQL
Server returns -1 for all other statements (Native or OLEDB provider). Other providers
might return 0 or -1.
To fetch a single value (such as an aggregate computation) use the ExecuteScalar method.
Against SQL Server 2000 you can use the SqlCommand.ExecuteXmlReader to retrieve
XML results directly from the server.
DataReader
When created, the SqlDataReader is positioned before the first record returned of the first
result set. You must invoke the Read method before accessing any data. As the
DataReader example demonstrates, the item property can be used to access the individual
fields or column values in the current row:
All the fields in a row can be accessed with the GetValues method.
object[] fields = new object[NumberFields];

...
int NumberFields = reader.GetValues(fields);
GetValue returns the column value in its native format. You can also access the column
values as particular datatypes: GetBoolean, GetDecimal, GetString, etc. The GetName
method returns the column name of a particular column.
To reinforce what was mentioned earlier, only one record at a time is accessible with a
DataReader. Make sure you close the DataReader when you are done with it.
Multiple Result Sets
The SqlDataReader class can handle multiple result sets, as the DataReader example
demonstrates. Two queries separated by a semicolon represent two SQL statements that
will cause two results sets to be generated, one for each statement.
string ConnString =
"server=localhost;uid=sa;pwd=;database=Northwind";
string cmd = "select CustomerId, CompanyName from
Customers where CustomerId like 'T%';select
CustomerId, CompanyName from Customers
where CustomerId like 'W%'";
...
int ResultSetCounter = -1;
int NumberFields = 0;
...
reader = command.ExecuteReader();
if (reader != null)
{
NumberFields = reader.FieldCount;
object[] fields = new object[NumberFields];
Console.WriteLine("Result Set\tCustomerId\tCompanyName");
do
{
ResultSetCounter++;
while(reader.Read())
{
NumberFields = reader.GetValues(fields);
Console.Write("\t{0}", ResultSetCounter);
for (int i = 0; i < NumberFields; i++)
{
Console.Write("\t\t{0}", fields[i]);
}
Console.Write("\n");
}
}while(reader.NextResult());
}
...
The FieldCount method returns the number of columns in the result set. Since the
GetValues method returns the native format of the data, an array of objects is passed to it.
The NextResult method navigates to the next result set.
Parameters Collection
Sometimes you have to parameterize a SQL statement. You also might have to associate
the input and output arguments of a stored procedure with variables in your program.
To do this you build the SqlCommand class's Parameters property, which is a collection
of SqlParameter instances. The installation procedure added the get_customers stored
procedure to the Northwind database to illustrate the use of a simple stored procedure,
which takes one input argument that is the company name and returns the customer id for
that customer.
CREATE PROCEDURE get_customers

(@companyname nvarchar(40), @customerid nchar(5) OUTPUT)
AS
select @customerid = CustomerID from Customers where
CompanyName = @companyname
RETURN
GO
The StoredProcedure example shows how to do this.
command = new SqlCommand("get_customers", conn);

command.CommandType = CommandType.StoredProcedure;
SqlParameter p = null;
p = new SqlParameter("@companyname", SqlDbType.NVarChar,
40);
p.Direction = ParameterDirection.Input;
p.Value = "Ernst Handel";
command.Parameters.Add(p);
p = new SqlParameter("@customerid", SqlDbType.NChar, 5);
p.Direction = ParameterDirection.Output;
command.Parameters.Add(p);
command.ExecuteNonQuery();
Console.WriteLine("{0} CustomerId = {1}",
command.Parameters["@companyname"].Value,
command.Parameters["@customerid"].Value);
Each individual SqlParameter member of the Parameters collection represents one

parameter of an SQL statement or stored procedure. As this example illustrates, the
parameters need not have any relationship to any particular table or column in the
database.
At a minimum you have to specify—either through the constructor or by setting

properties—the name and database type of the parameter. If the parameter is of variable
length, you have to specify the size.
In this example two parameters are added to the parameters collection. The first
represents the input argument to the stored procedure, the second the return value from
the stored procedure.
The name of the parameter corresponds to the name of the argument in the stored
procedure get_customers. The other values to the SqlParameter constructor define the
data type of the parameter. The first is a variable Unicode string up to 40 characters in
length. The second variable is a 5-character fixed-length Unicode string. If this was an
OLEDB .NET Data Provider you would bind to the parameters by position since only the
SQL Server .NET Data Provider binds parameters by name.
The Value property is used to set or get the value of the parameter. It is used to initialize
the @companyname parameter for input to the stored procedure. It is also used to obtain
the value that the stored procedure set for the @customerid parameter.
Output parameters must be specified as such with the Direction property. In this example
the @companyname parameter is set as an input parameter with the value
ParameterDirection.Input. The @customerid parameter is set as an output parameter with
the value ParameterDirection. Output. Output parameters must be specified, since input
parameters are the default. To bind to the return value of a stored procedure use
ParameterDirection.ReturnValue. For bidirectional parameters use
ParameterDirection.InputOutput.
You can use the parameter names to access individual parameters in the SqlCommand
parameters collection.
Parameterized commands work with both SqlDataReader and DataAdapter classes. When
the DataSet class is discussed, you will see how to specify the Source property of the
parameter, which indicates which column in the DataSet the parameter represents.
SqlDataAdapter and the DataSet Class
The DataSet class is a memory-resident, lightweight relational database class. It has

properties that reflect the tables (Tables) and relationships between tables (Relations)
within the data set. You can control whether corresponding constraints are enforced with
the EnforceConstraints property. You can name the data set with the DataSetName
property. You can also set the name of the data set in the DataSet constructor.
The SqlDataAdapter class is used to get data from the database into the DataSet. The
constructor of the HotelBroker class shows how to use a data adapter class to populate a
data set. The code is found in the Hotel subdirectory of the HotelBrokerAdmin directory
of the case study for this chapter.
conn = new SqlConnection(connString);

citiesAdapter = new SqlDataAdapter();
citiesAdapter.SelectCommand = new SqlCommand(
"select distinct City from Hotels", conn);
citiesDataset = new DataSet();
citiesAdapter.Fill(citiesDataset, "Cities");
The SqlDataAdapter class has properties associated with it for selecting, inserting,
updating, and deleting data from a data source. Here the SqlCommand instance is
associated with the SelectCommand property of the SqlDataAdapter instead of being
executed independently through one of its own execute methods.
The Fill method of the SqlDataAdapter is then used to execute the select command and
fill the DataSet with information to be put in a table whose name is supplied as an
argument. If the database connection was closed when the Fill method was executed, it
will be opened. When finished, the Fill method will leave the connection in the same state
as it was when it was first called.
At this point the connection to the database could be closed. You now can work with the
DataSet and its contained data independently of the connection to the database.
SqlDataAdapter is implemented with the SqlDataReader class, so you can expect better
performance with the latter. The SqlDataReader might also be more memory efficient
depending on how your application is structured. If you do not need the features of the
DataSet, there is no point incurring the overhead. If you are doing expensive processing
you can free up the database connection by using a DataSet. You may get better
scalability by loading the data into the DataSet, freeing the associated database resources,
and doing the processing against the DataSet.
Disconnected Mode
This scenario of working with a database is referred to as disconnected. Connected mode

represents a tightly coupled, connected environment where state and connections can be
maintained. Client-server environments are examples where this is true. ADO and
OLEDB were designed for this world. In a connected-mode environment data readers can
be used. If necessary, ADO can be used through the COM interop facility. In fact ADO
was not rewritten for .NET so that absolute backward compatibility could be maintained,
bugs and all.
Connections, however, are expensive to maintain in environments where you want to be

able to scale to a large number of users. In this environment there is often no need to hold
locks on database tables. This aids scalability because it reduces contention on database
tables. The DataTable objects in the DataSet's Tables collection with their associated
constraints can mimic the tables and relationships in the original database. For
applications that are implemented completely with .NET, DataSet instances can be passed
around or remoted to the various parts of an application. For applications that can make
optimistic assumptions about concurrency this can produce large gains in scalability and
performance. This is true of many types of Internet- or intranet-based applications.
In the disconnected mode, a connection is made in the same way as with the connected
mode of operation. Data is retrieved using the data provider's data adapter class. The
SelectCommand property specifies the SQL statement used to place data into the data set.
Unlike the data reader, which is related to a particular database connection, the data set
has no relationship to any database, including the one from which the data originally
came.
DataSet Collections
When data is placed into a DataSet, the related tables and columns are also retrieved.
Each data set has collections that represent all the tables, columns and data rows
associated with it.
The HotelBroker class in the Case Study has a method called PrintHotels that illustrates
how to retrieve this information and write it to a Console. The hotelsDataset is a data set
that has already been filled with the data from the HotelBroker database.
DataTable t = hotelsDataset.Tables["Hotels"];
if (t == null)
return;
foreach(DataColumn c in t.Columns)
Console.Write("{0, -20}", c.ColumnName);
foreach (DataRow r in t.Rows)

{
for (int i = 0; i < t.Columns.Count; i++)
{
Type type = r[i].GetType();
if (type.FullName == "System.Int32")
Console.Write("{0, -20}", r[i]);
else
{
string s = r[i].ToString();
s = s.Trim();
Console.Write("{0, -20}", s);
}
}
}
The Tables collection includes all the DataTable instances in the DataSet. In this
particular case there is only one, so there is no need to iterate through that collection. The
program then iterates through all the columns in the table and sets them up as headers for
the data that will be printed out. After the headers have been set up, all the rows in the
table are iterated through. For each column in the row, we ascertain its type and print out
the value appropriately. The program checks only for the types that are in the Hotels
database table. Checking for types instead of printing out the row values as object enables
us to format the data appropriately.
As we will show later, you can populate the dataset through these collections without
having to obtain it from a data source. You can just add tables, columns, and rows to the
appropriate collections.
DataSet Fundamentals
You can also fetch a subset of the data in the DataSet. The Select method on a DataTable
uses the same syntax as an SQL statement where clause. Column names are used to
access the data for a particular row. This example comes for the HotelBroker class, where
it is used to get the hotels for a particular city.
public ArrayList GetHotels(string city)

...
DataRow[] rows = t.Select("City = '" + city + "'");
HotelListItem hl;
ArrayList hotels = new ArrayList();
for (int i = 0; i < rows.Length; i++)
{
hl.HotelName = rows[i]["HotelName"].ToString();
hl.City = rows[i]["City"].ToString().Trim();
hl.NumberRooms = (int) rows[i]["NumberRooms"];
hl.Rate = (decimal) rows[i]["RoomRate"];
hotels.Add(hl);
}
return hotels;
The AddHotel method of the HotelBroker class demonstrates how to add a new row to a
DataSet. A new DataRow instance is created, and the column names are used to add the
data to the columns in the row.
To propagate your new row back to a database, you have to add it to the row collection of
the table, and then use the Update method on the SqlDataAdapter class to do so. It is the
data adapter that mediates between the DataSet and the database. We will discuss later
how to do perform edits on the dataset in order to accept or reject changes before
propagating them back to the database.
public string AddHotel(string city, string name,

int number, decimal rate)
...
DataRow r = t.NewRow();
r["HotelName"] = name;
r["City"] = city;
r["NumberRooms"] = number;
r["RoomRate"] = rate;
t.Rows.Add(r);
hotelsAdapter.Update(hotelsDataset, "Hotels");
To delete rows from the DataSet, you first find the particular row or rows you want to
delete and then invoke the Delete method on each DataRow instance. When the Update
method on the data adapter is called, it will be deleted from the database.
The Remove method removes the DataRow from the collection. It is not marked as
deleted, since it is no longer in the DataSet. When the Update method on the data adapter
is called, it will not be deleted from the database.
The DeleteHotel method in the HotelBroker class illustrates deleting rows from a
DataSet.
public string DeleteHotel(string city, string name)

...
t = hotelsDataset.Tables["Hotels"];
r = t.Select("City = '" + city + "' and HotelName = '"
+ name + "'");
...
for (i = 0; i < r.Length; i++)
r[i].Delete();
...
To update a row in a dataset, you just find it and modify the appropriate columns. This
example comes from the ChangeRooms method in the HotelBroker class. When the
Update method on the data adapter is called, the modification will be propagated back to
the database.
public string ChangeRooms(string city, string name,

int numberRooms, decimal rate)
...
DataRow[] r = t.Select(
"City = '" + city + "' and HotelName = '" + name + "'");
...
for (int i = 0; i < r.Length; i++)
{
r[i]["NumberRooms"] = numberRooms;
r[i]["RoomRate"] = rate;
}
...
Updating the Data Source
How does the SqlDataAdapter.Update method propagate changes back to the data
source? Changes to the DataSet are placed back based on the InsertCommand,
UpdateCommand, and DeleteCommand properties of the SqlDataAdapter class. Each of
these properties takes an SqlCommand instance that can be parameterized to relate the
variables in the program to the parts of the related SQL statement. The code fragment we
use to show this comes from the HotelBroker constructor.
A SqlCommand instance is created to represent the parameterized SQL statement that

will be used when the SqlDataAdapter.Update command is invoked to add a new row to
the database. At that point, the actual values will be substituted for the parameters.
SqlCommand cmd = new SqlCommand("insert Hotels(City,

HotelName, NumberRooms, RoomRate)
values(@City, @Name, @NumRooms, @RoomRate)", conn);
The parameters have to be associated with the appropriate columns in a DataRow. In the
AddHotel method code fragment discussed previously, columns were referenced by the
column names: "HotelName," "City," "NumberRooms," and "RoomRate." Notice how
they are related to the SQL statement parameters @Name, @City, @NumRooms,
@RoomRate in the SqlParameter constructor This last argument sets the Source property
of the SqlParameter. The Source property sets the DataSet column to which the parameter
corresponds. The Add method places the parameter in the Parameters collection
associated with the SqlCommand instance.
SqlParameter param = new SqlParameter("@City",

SqlDbType.Char, 20, "City");
cmd.Parameters.Add(param);
cmd.Parameters.Add(new SqlParameter("@Name",
SqlDbType.Char, 20, "HotelName"));
cmd.Parameters.Add(new SqlParameter("@NumRooms",
SqlDbType.Int, 4, "NumberRooms"));
cmd.Parameters.Add(new SqlParameter("@RoomRate",
SqlDbType.Money, 8, "RoomRate"));
Finally the SqlDataAdapters' InsertCommand property is set to the SqlCommand

instance. Now this command will be used whenever the adapter has to insert a new row
in the database.
hotelsAdapter.InsertCommand = cmd;
Similar code appears in the HotelBroker constructor for the UpdateCommand and
DeleteCommand properties to be used whenever a row has to be updated or deleted.
hotelsAdapter.UpdateCommand = new SqlCommand(

"update Hotels set NumberRooms = @NumRooms, RoomRate =
@RoomRate where City = @City and HotelName =
@Name", conn);
hotelsAdapter.UpdateCommand.Parameters.Add(new
SqlParameter("@City", SqlDbType.Char, 20, "City"));
SqlParameter("@Name", SqlDbType.Char, 20,
"HotelName"));
SqlParameter("@NumRooms", SqlDbType.Int, 4,
"NumberRooms"));
SqlParameter("@RoomRate", SqlDbType.Money, 8,
"RoomRate"));
hotelsAdapter.DeleteCommand = new SqlCommand(

"delete from Hotels where City = @City and HotelName =
@Name", conn);
hotelsAdapter.DeleteCommand.Parameters.Add(new
SqlParameter("@City", SqlDbType.Char, 20, "City"));
hotelsAdapter.DeleteCommand.Parameters.Add(new
SqlParameter("@Name", SqlDbType.Char, 20,
"HotelName"));
Whatever changes you have made to the rows in the DataSet will be propagated to the
database when SqlDataAdapter.Update is executed. How to accept and reject changes
made to the rows before issuing the SqlDataAdapter.Update command is discussed in a
later section.
Auto Generated Command Properties
The SqlCommandBuilder class can be used to automatically generate any

InsertCommand, UpdateCommand, and DeleteCommand properties that have not been
defined. Since the SqlCommandBuilder needs to derive the necessary information to
build those properties dynamically, it requires an extra round trip to the database and
more processing at runtime. Therefore, if you know your database layout in the design
phase, you should explicitly set the InsertCommand, UpdateCommand, and
DeleteCommand properties to avoid the performance hit. If the database layout is not
known in advance, and a query is specified by the user, the SqlCommandBuilder can be
used if the user subsequently wants to update the results.
This technique works for DataTable instances that correspond to single tables. If the data
in the DataTable is generated by a query that uses a join, then the autogeneration
mechanism cannot generate the logic to update multiple tables. The SqlCommandBuilder
uses the SelectCommand property to generate the command properties.
A primary key or unique column must exist on the table in the DataSet. This column must
be returned by the SQL statement set in the SelectCommand property. The unique
columns are used in a where clause for update and delete.
Column names cannot contain special characters such as spaces, commas, periods,
quotation marks, or nonalphanumeric characters. This is true even if the name is
delimited by brackets. You can specify a fully qualified table name such as
SchemaName.OwnerName.TableName.
A simple way to use the SqlCommandBuilder class is to pass the SqlDataAdapter
instance to its constructor. The SqlCommandBuilder then registers itself as a listener for
RowUpdating events. It can then generate the needed InsertCommand, UpdateCommand,
or DeleteCommand properties before the row update occurs.
The CommandBuilder example demonstrates how to use the SqlCommandBuilder class.
Database Transactions and Updates
When the data adapter updates the data source, it is NOT done as a single transaction. If
you want all the inserts, updates, and deletes done in one transaction, you must handle the
transaction programmatically.
The SqlConnection object has a BeginTransaction method that returns a SqlTransaction

object. When you invoke the BeginTransaction method, you can optionally specify the
isolation level. If you know what you are doing, and understand the trade-offs, you can
improve the performance and scalability of your application by setting the appropriate
isolation level. If you set the isolation level incorrectly or inappropriately, you can have
inconsistent or incorrect data results. [10]
[10]
Discussing isolation levels in detail would remove our focus from .NET to database
programming. Any good intermediate to advanced book on database programming would
discuss the concept of isolation levels and locking. For specific information about the
SQL Server locking mechanism you can read the Microsoft Press Inside SQL Server
books, among others. Tim Ewald's book Transactional COM+ has a good chapter on the
issue of isolation and its relation to building scalable applications.
The SqlTransaction class has Commit and Rollback methods to commit or abort the
transaction. You open the SqlConnection, invoke the BeginTransaction method, use the
SqlDataAdapter as normal, and then call SqlTransaction.Commit or
SqlTransaction.Rollback as appropriate. Then close the connection. The Save method on
SqlTransaction can be used to set a savepoint in the transaction.
In order to minimize the database resources you hold, and therefore increase the
scalability of your application, you want to minimize the time between calling
BeginTransaction and the call to Commit or Rollback.
Here is some code from the Transactions example. It uses the AirlineBroker database
introduced later in the chapter. Note that we only open the connection right before the
Fill, and the transaction statements bracket the Update.
conn = new SqlConnection(ConnString);

da = new SqlDataAdapter();
ds = new DataSet();
da.SelectCommand = new SqlCommand(selectCmd, conn);
da.InsertCommand = new SqlCommand(insertCmd, conn);
...
conn.Open();
da.Fill(ds, "Airlines");
...
trans = conn.BeginTransaction();
da.InsertCommand.Transaction = trans;
da.Update(ds, "Airlines");
trans.Commit();
...
To ensure that the SQL Server data provider operates properly, you should use the
Commit and Rollback methods on the SqlTransaction object to commit or roll back the
transactions started with SqlConnection.BeginTransaction. Do not use the SQL Server
transaction statements.
If you use stored procedures for your database work, you can certainly issue SQL Server
transaction statements inside the stored procedures instead of using the SqlTransaction
object. Stored procedures can be used to encapsulate transactional changes. The
MakeReservation stored procedure in the HotelBroker database does just that.
Optimistic vs. Pessimistic Locking and the DataSet
Transactions help preserve database consistency. When you move money from your
savings to your checking account to pay your phone bill, transaction processing ensures
that the credit and withdrawal will both happen, or neither will happen. You will not wind
up with a situation where the money goes into your checking account but is not
withdrawn from the savings (good for you, and bad for the bank) or the reverse (bad for
you, but good for the bank). Nothing about that transaction prevents your spouse from
using that same money to eat out at a fancy restaurant. [11]
[11]
The failure to distinguish between these two leads to the apparently common problem
(as related to me by a bank vice president) of people wondering why their checks bounce
when their ATM balance said they had enough money to withdraw some cash.
Under an optimistic locking strategy, you assume this will not happen, but you have to be
prepared to deal with it when does. [12] A pessimistic locking strategy requires
coordination among all the users of a database table so that this never happens. Of course,
the fewer locks you hold on database rows to prevent use by more than one user, the
more scalable your application will be.
[12]
This is the database equivalent of overdraft protection.
An understanding of how this affects your application applies to both reads and actual
updates. For example, suppose your spouse sees that money is available in the checking
account and makes plans based on that fact. This could be as much of a problem as the
actual withdrawal of money from the joint checking account.
While a discussion of how to solve these problems is beyond the scope of this chapter, it
is important to realize that the issue arises because no locks are held on the database
records held within a DataSet. Just using the DataSet with SqlDataAdapter.Update
assumes an optimistic locking strategy.
Why does this matter? It matters because the performance and scalability of your
application depend on it. Why is it so complicated? Because there is no answer that
applies to all applications in all situations. If users do not share the same set of data,
optimistic concurrency is an excellent assumption. If you have to lock records for a long
period of time, this increases the wait to use these resources, thus decreasing performance
and scalability.
You have to understand transaction isolation levels, the database's Lock Manager, the
probability of contention for particular rows, and the probability that this contention
results in deadlock in your application. You have to understand how much time and
resources you can spend reconciling divergent operations, and how much tolerance for
inconsistent or incorrect results your application can stand, in order to decide under what
circumstances you want to avoid deadlock at all costs, or can deal with the consequences
of conflicting operations. [13]
[13]
Tim Ewald's book is worth reading to understand this topic. Philip Bernstein and Eric
Newcomer's Principles of Transaction Processing is another good reference.
You might have to use the DataSet with additional logic to test whether the records in the
DataSet have been changed since the last time they were fetched or modified. Or you
might just decide to use the SqlDataReader and refetch the data. It all depends.
For example, when making a reservation in our HotelBroker case study you cannot make
an optimistic assumption about the availability of rooms. It is not acceptable to assume an
infinite supply of rooms at a hotel and let the reservations clerk deal with what happens
when more people show up then there are rooms for. [14] We use the MakeReservation
stored procedure to check on the availability of a room before we make the reservation.
[15]
[14]
Of course, airlines and hotels overbook. This is a conscious strategy to deal with
passengers or guests not making explicit cancellations, not a database concurrency
strategy.
[15]
In fact, the transaction in MakeReservation includes the checking of the availability of
the room as well as the actual making of the reservation in order to maintain consistency.
It also breaks up what could be one multiple table join into several queries in order to
return better error information.
Sometimes, even without concurrency issues, the DataSet cannot be used to add new
rows in isolation from the database. Sometimes, as in our HotelBroker application, an
arbitrary primary key cannot be used. [16] Many users will be making reservations at the
same time. Reservation ids cannot be assigned locally; some central logic on the database
has to be employed to issue them. [17] The MakeReservation stored procedure does this as
well.
[16]
For instance, a GUID. Well, theoretically GUIDs could be used in our case, but when
was the last time you got a reservation number from a hotel or airline that was composed
of 32 identifiers? Many times a primary key has meaning to an organization—for
example, a part number whose subsections indicate various categories.
[17]
Of course, if performance were critical, instances of the HotelBroker could be
preassigned ranges of reservation ids to give out. But this would have to be done by some
central authority as well (the database, some singleton object?). But then this raises the
issue of state management in the middle tier. This just reinforces my previous point about
the dependency of any solution on the specific requirements of your program. It also
reinforces the maxim that any programming problem can be solved either by trading
memory against time, or adding another level of indirection.
The degree of disconnected operation that your application can tolerate has to be
understood before you can decide how to use SqlDataReader or the DataSet in your
applications.
Why bother to use the DataSet at all in our HotelBroker application? In fact, the code for
the Customer object does not use the DataSet at all. The HotelBroker object does—for
two reasons. The first is pedagogical. We wanted to show you how a complete application
might use the features of the DataSet, rather than just isolated sample programs. Second,
in the Web version of the application which is developed in subsequent chapters, it is
convenient to cache certain pieces of information. For example, it is probably reasonable
to assume that a user can work with their own local copy of reservations. On the other
hand, the information about a customer such as their email address can be obtained just
once when they log in. There is no need for an elaborate mechanism to cache customer
information, so the Customer object uses methods on the SqlCommand object.
Working with DataSets
Figure 9-5 depicts the hierarchy of classes that exist within the DataSet class. It will be
helpful to glance at this diagram over the next few sections that discuss these classes.
Figure 9-5. DataSet class hierarchy.

Multiple Tables in a DataSet
Each DataSet has a collection of one or more DataTable objects. Each DataTable object
represents one table.
With a SelectCommand that contains a join you can place data from multiple database
tables into one DataTable. If you want to update the multiple tables, you will have to
specify the update commands, because all relationship of the data to the original tables is
lost. The SqlDataAdapter for the HotelBroker object has the following SelectCommand
property:
string cmd = "select CustomerId, HotelName, City,

ArrivalDate, DepartureDate, ReservationId
from Reservations, Hotels where
Reservations.HotelId = Hotels.HotelId";
adapter.SelectCommand = new SqlCommand(cmd, conn);
dataset = new DataSet();
adapter.Fill(dataset, "Reservations");
The DataSet will only have one DataTable called Reservations. The fact that some of the
data came from the Hotels table is lost.
You can also load more than one table into a dataset. The DataSchema example does just
this:
adapter.SelectCommand = new SqlCommand(

"select * from [Order Details] where ProductId = 1",
conn);
adapter.FillSchema(dataset, SchemaType.Source,
"Order Details");
adapter.Fill(dataset, "Order Details");
adapter.SelectCommand = new SqlCommand("select * from

Shippers", conn);
adapter.FillSchema(dataset, SchemaType.Source,
"Shippers");
adapter.Fill(dataset, "Shippers");
There will be two tables, OrderDetails and Shippers, in the DataSet. The method
SqlDataAdapter.FillSchema fills the DataSet with the primary key information associated
with the tables. The code can now iterate through the tables and print out both the data
and the primary keys of the tables. The Columns collection on the DataTable enables you
to find the DataColumns for the DataTable.
foreach(DataTable t in dataset.Tables)
{
Console.WriteLine(t.TableName);
DataColumn[] dc = t.PrimaryKey;
for (int i = 0; i < dc.Length; i++)
{
Console.WriteLine("\tPrimary Key Field {0} = {1}", i,
dc[i].ColumnName);
}
Console.Write("\t");
Console.Write("{0, -20}", c.ColumnName);
foreach(DataRow r in t.Rows)
{
Console.Write("\t");
Console.Write("{0, -20}", r[c].ToString().Trim());
}
}
The example output shows the tables, primary keys, columns, and data:
Order Details
Primary Key Field 0 = OrderID
Primary Key Field 1 = ProductID
OrderID ProductID UnitPrice Quantity Discount
10285 1 14.4 45 0.2
10294 1 14.4 18 0
...
Shippers
Primary Key Field 0 = ShipperID
ShipperID CompanyName Phone
1 Speedy Express (503) 555-9831
2 United Package (503) 555-3199
3 Federal Shipping (503) 555-9931
Table Creation without a Data Source
One can use a DataSet as a memory-resident relational database not based on any
database. In fact, we will explore various features of the DataSet in the DataEditing
example by adding the data and relationships directly to the data set without extracting
them from a database.
First we create a new DataSet and turn on constraint checking. We then add four
DataTables to the DataSet: Books, Categories, Authors, and BookCategories. Even
though it is set in the example code for pedagogical reasons, EnforceConstraints by
default is true.
DataSet ds = new DataSet();

ds.EnforceConstraints = true;
DataTable categories = ds.Tables.Add("Categories");

DataTable bookcategories = ds.Tables.Add("BookCategories");
DataTable authors = ds.Tables.Add("Authors");
DataTable books = ds.Tables.Add("Books");
Each DataTable object has a collection of DataColumn objects. Each object represents
one column of the table. We then add columns to the table definition.
Type stringType = System.Type.GetType("System.String");

Type intType = System.Type.GetType("System.Int32");
DataColumn categoryname =
categories.Columns.Add("Category",stringType);
DataColumn cn = bookcategories.Columns.Add
("CategoryName", stringType);
DataColumn loc =
bookcategories.Columns.Add(
"LibraryofCongressNumber", stringType);
DataColumn auid = authors.Columns.Add("AuthorId",

intType);
authors.Columns.Add("AuthorLastName", stringType);
authors.Columns.Add("AuthorFirstName", stringType);
DataColumn ISBN = books.Columns.Add("ISBN", stringType);
DataColumn booksauid = books.Columns.Add("AuthorId",
intType);
books.Columns.Add("Title", stringType);
DataColumn bloc =
books.Columns.Add("LibraryofCongressNumber", stringType);
Constraints and Relations
Each DataTable object has a collection of DataRow objects. Each object represents one
row of the table. When you add a DataRow, it is subject to the constraints on the
DataTable objects (assuming the DataSet's EnforceConstraints property has been set to
true).
Primary Keys
There are several constraints on a table. The Primary Key constraint is the unique
identifier for the table. Other unique constraints force the values in various column(s) to
which they are applied to be unique. A Foreign Key constraint forces the values in the
column(s) to which it applies to be a primary key in another table in the DataSet.
The DataTable's primary key is a property:
DataColumn[] bookcategoriesPK = new DataColumn[2];

bookcategoriesPK[0] = cn;
bookcategoriesPK[1] = loc;
bookcategories.PrimaryKey = bookcategoriesPK;
DataColumn[] authorsPK = new DataColumn[1];

authorsPK[0] = auid;
authors.PrimaryKey = authorsPK;
DataColumn[] booksPK = new DataColumn[1];

booksprimarykey[0] = ISBN;
books.PrimaryKey = booksPK;
Constraints
The other constraints on the Table are represented by the abstract base class Constraint
and its derived classes: UniqueConstraint and ForeignKeyConstraint. The base class
enables the constraints to be placed in the table's constraint collection. Primary Keys also
appear in the table's constraint collection as a unique constraint with a system-generated
name. The UniqueConstraint.IsPrimaryKey property can be used to detect primary keys.
We constrain the Category column in the Categories table to be unique. Since the last
argument to the Add method is false, this is not a primary key of the table. We do not
define a primary key for this table, only a unique constraint. In fact, we do not even have
to define any constraint on the table. Although that would violate the rules of relational
integrity, you are not forced to use the DataSet in a relational manner. If you wish you
can add a name to the constraint.
categories.Constraints.Add("Unique CategoryName
Constraint", categoryname, false);
Foreign Keys can specify what action should be taken when the primary key on which it
is based is changed. Your choices are the standard database choices: None, Cascade,
SetNull. You can also use SetDefault to set the new value to the DataColumn's
DefaultValue property. These operations can be specified for both update and delete
conditions.
In this example, a foreign key constraint is set so that all author ids in the Books table
have to be found in the Authors table. In other words, when a new book row is inserted, it
must have an author. We give this constraint a name: "Authors->Books." If the author id
is changed, the update rule forces the DataSet to change all the author ids in the related
rows to the new author id. If the author id is deleted, the DataSet will set the deleted
author ids in the Book rows to null. If we had set the DeleteRule to Cascade, a cascading
delete would be applied to all those rows in the Books table. The AcceptRejectRule
applies to editing of the DataSet, which we will cover in a future section. This rule
dictates what happens when the AcceptChanges method is invoked on a DataSet,
DataRow, or DataTable. In this case all changes are cascaded. The alternative rule would
be to take no action (None).
DataColumn[] bookauthorFK = new DataColumn[1];

bookauthorFK[0] = booksauid;
ForeignKeyConstraint fk = new
ForeignKeyConstraint("Authors->Books",
authorsPK, bookauthorFK);
fk.AcceptRejectRule = AcceptRejectRule.Cascade;
fk.DeleteRule = Rule.SetNull;
fk.UpdateRule = Rule.Cascade;
books.Constraints.Add(fk);
Data Relations
Besides constraints you can add a relation to the DataSet's DataRelation collection. A
relation connects two tables so that you can navigate between the parent and the child or
the child and the parent. When you add the relation you can optionally create and add the
equivalent unique and foreign key constraints to the parent and child tables' constraint
collections. [18]
[18]
Use the optional boolean createConstraints argument when you add a relation to
indicate whether the associated constraint should be added. If this argument is not
specified, the default is to add the associated constraint.
The Categories table is made the parent of the BookCategories table through the
Categories and CategoryName columns. In a relation both columns have to be of the
same type (string). You can use this relation to navigate by finding all the rows in the
child table that have the same value as in the parent table, or finding the row in the parent
table that is the parent of a row in the child table. Similarly the Library of Congress
number associated with a book has to be found in the Library of Congress field in the
BookCategory's Library of Congress field.
ds.Relations.Add("Category->BookCategories Relation",
categoryname, cn);
ds.Relations.Add("Book Category LOC->Book LOC Relation",
loc, bloc);
Examining the Schema Information about a DataTable
You can examine the information about a DataTable. Here is how to examine the
constraint and key information. A previous example has already shown you how to find
the DataColumns for a DataTable. Note the use of the IsPrimaryKey property on the
UniqueConstraint to detect a primary key.
foreach(DataTable t in ds.Tables)
{
Console.WriteLine(t.TableName);
Console.WriteLine("\tPrimary Key:");
for (int i = 0; i < t.PrimaryKey.Length; i++)
{
DataColumn c = t.PrimaryKey[i];
Console.WriteLine("\t\t{0}", c.ColumnName);
}
Console.WriteLine("\tConstraints:");
foreach(Constraint c in t.Constraints)
{
string constraintName;
if (c is ForeignKeyConstraint)
constraintName = "Foreign Key:" + c.ConstraintName;
else if (c is UniqueConstraint)
{
UniqueConstraint u = (UniqueConstraint)c;
if (u.IsPrimaryKey)
constraintName = "Primary Key";
else
constraintName = u.ConstraintName;
}
else
constraintName = "Unknown Name";
Console.WriteLine("\t\t{0, -40}", constraintName);

}
}
This produces the following output. Note how the relations defined as a DataRelation
appear in the table's constraint collection as a ForeignKeyConstraint instance.
PrimaryKeys appear in the constraint collection as a UniqueConstraint instance.
Constraints defined as unique constraints or foreign keys appear as you would expect in
the collection.
Categories
Primary Key:
Constraints:
Unique CategoryName Constraint
BookCategories
Primary Key:
CategoryName
LibraryofCongressNumber
Constraints:
Primary Key
Foreign Key:Category->BookCategories Relation
Constraint2
Authors
Primary Key:
AuthorId
Constraints:
Primary Key
Books
Primary Key:
ISBN
Constraints:
Primary Key
Foreign Key:Authors->Books
Foreign Key:Book Category LOC->Book LOC Relation
Note the BookCategories constraint with the system-generated name. If you examine the
code carefully, you will see we never added this constraint. Where did it come from? If
you were to look at the columns in that constraint, you would find the Library of
Congress field. The system realized that, since the CategoryName is a foreign key in
another table, the Library of Congress field should be unique.
You can also examine the relations collection on the DataSet. You can examine the parent
table and the columns in the parent table involved in the relationship. You can also
examine the child table in the relationship and its columns.
foreach(DataRelation dr in ds.Relations)
{
DataTable parentTable = dr.ParentTable;
DataTable childTable = dr.ChildTable;
Console.WriteLine(" Relation: {0} ", dr.RelationName);
Console.WriteLine(" ParentTable: {0, -10}",
parentTable);
Console.Write(" Columns: ");
for(int j = 0; j < dr.ParentColumns.Length; j++)
Console.Write(" {0, -10}",
dr.ParentColumns[j].ColumnName);
Console.WriteLine(" ChildTable: {0, -10}",
childTable);
Console.Write(" Columns: ");
for(int j = 0; j < dr.ChildColumns.Length; j++)
Console.Write(" {0, -10}",
dr.ChildColumns[j].ColumnName);
}
Here is the resulting output:
Output Relations between tables in the DataSet...

Relation: Category->BookCategories Relation
ParentTable: Categories
Columns: Category
ChildTable: BookCategories
Columns: CategoryName
Relation: Book Category LOC->Book LOC Relation
ParentTable: BookCategories
Columns: LibraryofCongressNumber
ChildTable: Books
Columns: LibraryofCongressNumber
Database Events
Several ADO.NET classes generate events.
The SqlConnection class generates the StateChange and InfoMessage events. The
SqlDataAdapter generates the RowUpdated and RowUpdating events. The DataTable
class generates the ColumnChanging, ColumnChanged, RowChanged, RowChanging,
RowDeleted, and RowDeleting events.
For example, the RowChanged event occurs after an action has been performed on a row.
Continuing with our DataEditing example, it defines a handler for the RowChanged event
in the Books table. Every time a row changes in the Books table, the event handler will
run.
books.RowChanged+=new
DataRowChangeEventHandler(Row_Changed);
private static void Row_Changed(object sender,

System.Data.DataRowChangeEventArgs e)
{
DataTable table = (DataTable)sender;
DataColumn[] primaryKey = table.PrimaryKey;
string keyName = primaryKey[0].ColumnName;
Console.WriteLine("Table " + table.TableName + " " +

e.Action.ToString() + "Row with Primary Key " +
e.Row[keyName]);
return;
}
So when the code adds some rows, including some to the Books table:
DataRow row
row = categories.NewRow();
row["Category"] = "UnitedStates:PoliticalHistory";
categories.Rows.Add(row);
...
row = authors.NewRow();
row["AuthorId"] = 1;
row["AuthorLastName"] = "Burns";
row["AuthorFirstName"] = "James M.";
authors.Rows.Add(row);
...
row = books.NewRow();
row["ISBN"] = "0-201-62000-0";
row["Title"] = "The Deadlock of Democracy";
row["LibraryofCongressNumber"] = "E183.1";
books.Rows.Add(row);
row = books.NewRow();
row["ISBN"] = "0-201-62000-3";
row["Title"] = "Freedom and Order";
row["LibraryofCongressNumber"] = "E183.1";
books.Rows.Add(row);
We get one output line for each book added, printed by the event handler:
Table Books AddRow with Primary Key 0-201-62000-0

Table Books AddRow with Primary Key 0-201-62000-3
If we were to change the ISBN numbers of the two books that were added to the same
value, a ConstraintException would be thrown. If we changed the
DataSet.EnforceConstraints property to false, however, no exception would be thrown.
Navigating Relationships
Using the schema information, we can navigate from parent table to child table and print
out the results. This cannot be done with relationships defined as ForeignKeyConstraint,
only as a DataRelation in the relations collection of the DataSet.
We previously printed out the schema information associated with the relationships. Now
we use this information to print out the parent and child rows in the relationships. By
using relationships appropriately, you can walk through the data without using relational
queries. This can be quite useful for finding all the books in a certain category, or all
order items in an order.
Note the use of the DataRow methods GetChildRows and GetParentRows to do the
navigation. For a given relation, first we navigate from parent to children, then from the
children to their parent. We also show how you can use different constructs to access the
items in the various collections.
foreach (DataRelation dr in ds.Relations)

{
Console.WriteLine(dr.RelationName);
DataTable parentTable = dr.ParentTable;
foreach(DataRow parentRow in parentTable.Rows)
{
Console.Write(" Parent Row: ");
foreach(DataColumn pc in parentTable.Columns)
Console.Write(" {0} ", parentRow[pc]);
DataRow[] childRows = parentRow.GetChildRows(dr);

for(int k = 0; k < childRows.Length; k++)
{
Console.Write(" Child Row: ");
foreach(DataColumn cc in childTable.Columns)
Console.Write(" {0} ", childRows[k][cc]);
}
}
foreach(DataRow childRow in childTable.Rows)

{
Console.Write(" Child Row: ");
for(int m = 0; m < childTable.Columns.Count; m++)
Console.Write(" {0} ", childRow[childTable.
Columns[m].ColumnName].ToString().Trim());
foreach(DataRow pRow in childRow.GetParentRows(dr))

{
Console.Write(" Parent Row: ");
for(int p = 0; p < parentTable.Columns.Count; p++)
Console.Write(" {0} ", pRow[parentTable.
Columns[p].ColumnName].ToString().Trim());
}
}
}
Next let us look at the output that this code produces. Note how we loop through each
relation. For each relation we first loop through the parent table and output each row of
the parent table with its corresponding child rows. We then loop through the child table
and output each row of the child table with its corresponding parent rows.
...
Category->BookCategories Relation
Parent Row:UnitedStates:PoliticalHistory
Child Row:UnitedStates:PoliticalHistory E183
Parent Row:UnitedStates:PoliticalHistory:Opinion
Child Row:UnitedStates:PoliticalHistory:Opinion E183.1
Parent Row:UnitedStates:PoliticalHistory:Predictions
Child Row:UnitedStates:PoliticalHistory:Predictions
E183.3
Child Row:UnitedStates:PoliticalHistory E183

Parent Row:UnitedStates:PoliticalHistory
Child Row:UnitedStates:PoliticalHistory:Predictions
E183.3
Book Category LOC->Book LOC Relation

Parent Row:UnitedStates:PoliticalHistory E183
Parent Row:UnitedStates:PoliticalHistory:Opinion E183.1
Child Row:0-201-62000-0 1
The Deadlock of Democracy E183.1
Child Row:0-201-62000-3 2
Freedom and Order E183.1
Parent Row:UnitedStates:PoliticalHistory:Opinion E183.2
E183.3
Child Row:0-201-62000-0 1
The Deadlock of Democracy E183.1
E183.1
Child Row:0-201-62000-3 2 Freedom and Order E183.1
E183.1
DataRow Editing
BeginEdit, EndEdit, CancelEdit
If you want to make multiple edits to a DataSet, and postpone the checking of constraints
and events, you can enter a dataset editing mode. You enter this mode by invoking the
BeginEdit method on the row. You leave it by invoking the EndEdit or CancelEdit row
methods.
In the DataEditing example, we violate the foreign-key constraint by adding a row with a
nonexistent author id. The foreign-key constraint exception will not be raised until the
EndEdit method is called.
Since we have called BeginEdit in the following code fragment, there is no exception
caught.
DataRow rowToEdit = books.Rows[0];

rowToEdit.BeginEdit();
try
{
rowToEdit["AuthorId"] = 21;
...
}
catch(Exception e)
{
Console.WriteLine("\n" + e.Message + " while editing a
row.");
}
However, when we invoke the EndEdit method on the row, the exception is raised.
try
{
rowToEdit.EndEdit();
}
catch(Exception e)
{
Console.WriteLine("\n" + e.Message + " on EndEdit");
}
The following message is printed out because the illegal value was still present when the
editing session was finished.
ForeignKeyConstraint Authors->Books requires the child key

values (21) to exist in the parent table. on EndEdit
DataRow Versions
Before the row changes have been accepted, both the original and the changed row data
are available. The item property [19] of the row can take a DataRowVersion to specify
which value you want. The version field can be Original, Default, Current, or Proposed.
[19]
The item property of the DataRow is the indexer for the class.
Console.WriteLine("BeginEdit called for Book AuthorId.");

rowToEdit.BeginEdit();
rowToEdit["AuthorId"] = 2;
Console.WriteLine("Current Value {0}",
rowToEdit["AuthorId", DataRowVersion.Current]);
Console.WriteLine("Proposed Value {0}",
rowToEdit["AuthorId", DataRowVersion.Proposed]);
Console.WriteLine("Default Value {0}",
rowToEdit["AuthorId", DataRowVersion.Default]);
rowToEdit.EndEdit();
Console.WriteLine("Current Value {0}",

rowToEdit["AuthorId", DataRowVersion.Current]);
Console.WriteLine("Default Value {0}",
rowToEdit["AuthorId", DataRowVersion.Default]);
Console.WriteLine("EndEdit called.");
...
This code caused the following output to be printed out:
BeginEdit called for Book AuthorId.

Current Value 1
Proposed Value 2
Default Value 2
...
EndEdit called.
Current Value 2
Default Value 2
During editing the Current and Proposed item values are available. After CancelEdit, the
Proposed value is no longer available. After EndEdit, the Proposed value becomes the
Current value, and the Proposed value is no longer available.
DataRow RowState Property
In addition to the Current and Proposed values of a field, the DataRow itself has a
property that indicates the state of the particular row. The values can be Added, Deleted,
Detached, Modified, or Unchanged.
A row is in the Detached state when it has been created, but has not been added to any
DataRow collection, or it has been removed from a collection.
The Default DataRowVersion of a field returns the appropriate row version depending on
the RowState property.
Accepting and Rejecting Changes
Calling EndEdit on a DataRow does not cause the changes to be made to the row. Calling
the AcceptChanges or RejectChanges method on the DataSet, DataTable, or DataRow
ends editing on all the contained rows of the appropriate scope. If EndEdit or CancelEdit
has not been called, these methods do it implicitly for all rows within its scope.
After the AcceptChanges method, the Current value becomes the Original value. If
EndEdit has not been called the Proposed value becomes the new Current and Original
values. If the RowState was Added, Modified, or Deleted it becomes Unchanged and the
changes are accepted.
After the RejectChanges method, the Proposed value is deleted. If the RowState was
Deleted or Modified, the values revert to their previous values, and the RowState
becomes Unchanged. If the RowState was Added, the row is removed from the Rows
collection.
Since the RowState after AcceptChanges is Unchanged, calling the DataAdapter's Update
method at this point will not cause any changes to made on the data source. Therefore,
you should call the Update method on the DataAdapter to update changes to the data
source before calling AcceptChanges on any row, table, or DataSet.
Here is the code from the case study's HotelBroker object's CancelReservation method.
Note how AcceptChanges on the DataSet is called if the SqlDataAdapter.Update method
succeeds. If an exception is thrown, or the update fails, RejectChanges is called.
public void CancelReservation(int id)

{
DataTable t = null;
try
{
t = dataset.Tables["Reservations"];
DataRow[] rc = t.Select("ReservationId = " + id + " ");
for (int i = 0; i < rc.Length; i++)

rc[i].Delete();
int NumberRows = adapter.Update(dataset,

"Reservations");
if (NumberRows > 0)
t.AcceptChanges();
else
t.RejectChanges();
}
catch(Exception e)
{
t.RejectChanges();
throw e;
}
return;
}
If you do not reject the changes on failure, the rows will still be in the DataSet. The next
time an update is requested, the update will be rejected again, because the rows are still
waiting to be updated. Since the DataSet is independent of a database, the fact that an
update occurs on the database has nothing to do with accepting or rejecting the changed
rows in the DataSet.
DataRow Errors
If there have been any data editing errors on a row, the HasErrors property on the
DataSet, DataTable, or DataRow will be set to true. To get the error, use the DataRow's
GetColumnError or the GetColunmsInError methods.
At this point we have covered more than enough material for you to understand the
database version of the Customer and HotelBroker objects in the case study. As usual, the
code is in the CaseStudy directory for this chapter.
Since there will never be any reason for the Customer object to hold any state, the
Customer object methods use SqlDataReader to access the database and return the
results. Any state that a program might need (i.e., a list of customers) could easily be
maintained in the client program and not in a middle-tier object.
The HotelBroker and HotelBookings objects are a little more complicated. As mentioned
earlier, for pedagogical reasons alone these objects would have been implemented using a
DataSet to show you how that technology would work in an application.
Nonetheless, we will see that with Web applications there might be a reason to keep some
state in the middle tier. In that scenario, the DataSet can serve as an intelligent cache.
XML Data Access
As we will discuss in the Web Services chapter, XML has many advantages for
describing data that must move between heterogeneous systems and data sources. Since
you can validate your XML against an XML schema description, you can pass it in many
situations where passing a DataSet makes no sense. [20] Since XML is text, it can pass
through firewall ports that are normally open, unlike the DCOM or RMI protocols that
require special ports to be open.
[20]
When you remote a DataSet it is remoted as XML, nonetheless, if you have to interact
with an unmanaged program you can convert the data in the DataSet to XML and send it.
As discussed in the Web Services chapter, the XML protocol used by remoting and Web
Services is not identical.
The thrust of these next sections is not to discuss XML in any great detail. We just want
to demonstrate how you can move back and forth between looking at data in XML and
looking at data with a DataSet.
XML Schema and Data
XML does not dictate how data is organized or what the meaning of XML documents are.
It only describes the rules on how the documents are put together. [21] An XML schema
describes the metadata of how the data is organized inside an XML document. XML
schemas are written in XML.
[21]
Technically speaking, XML documents in the sense that we speak of are defined by
the XML Infoset and consist of documents, elements, and attributes.
For example, XML can be used to describe data in a relational database, but an XML
schema can be used to describe relationships such as primary and foreign keys. Having
the XML schema and the data in one document or text stream is vastly simpler than
having to download each table into a dataset and then programmatically set up the
relations between the tables.
XmlDataDocument
Documents can include database output within them. For example, a sales report has an
explanation as well as the sales data that was pulled from a data source. The
XmlDataDocument class can be used to represent data in the form of an XML document.
The XmlDataDocument class inherits from XmlDocument which represents an XML

document for the .NET XML Framework classes. What makes the XmlDataDocument
particularly interesting is that you can construct an XmlDataDocument from a DataSet by
passing the DataSet instance to the XmlDataDocument constructor. The
XmlDataDocument has a read-only DataSet property so that you can work with the XML
document as relational data if that makes sense.
DataSet and XML
The DataSet has methods, WriteXml and WriteXmlSchema, that can write out the data
and schema associated with the dataset. The XML schema that the DataSet writes out is
deduced from the current set of tables, columns, constraints, and relations. Unless you
explicitly add the constraints to the DataSet, such as primary- or foreign-key
relationships, they will not be part of the schema.
The DataSet also has methods to read XML: ReadXml and ReadXmlSchema. ReadXml
can read both the data and the schema into the dataset. If a schema is not present, it will
try to infer one from the data. If it cannot infer a schema, it will throw an exception.
ReadXmlSchema will read in a schema document.
If there is no schema in an XML document, the DataSet extracts elements that would be
defined as tables according to a set of rules. The remaining elements, along with the
attributes, are then assigned as columns to the tables.
You can use the ColumnMapping property of the DataColumn class to control whether
you want columns written as XML elements or attributes. Elements that are not scalar
values become tables; attributes and scalar values are columns. The exact procedure is
described in the .NET documentation.
AirlineBrokers Database
The AirlineBrokers database will be used to study XML data access. This database can be
created using the SqlServer Enterprise Manager and the airlinebroker.sql script found in
the AcmeDatabaseScripts subdirectory of the case study. The AirlineBrokers database
represents another service that the Acme reservation system uses. Acme customers can
make airline reservations to the places they wish to go.
The database has several tables:
• Airlines: information about the various airlines in the database

• PlaneType: the various planes that the airlines use
• Flights: information about the various airlines' flights
• Customers: information about customers
• Reservations: information about the customers' reservations
Although in real life the Airline Broker and the Hotel Broker would not have the same
Customers table, for simplicity we use the same table structure, and we use the same
component to access it.
DataSet and XML
To illustrate the relationship between the relation model of the DataSet and the XML
model we will first fetch some information from the database. The DataSetXml example
uses the same commands and techniques we have studied in this chapter to extract the
data.
First the connection, DataSet, and the SqlDataAdapters for the various tables are created.
SqlConnection conn = new SqlConnection(connectString);

DataSet d = new DataSet("AirlineBroker");
SqlDataAdapter airlinesAdapter = new SqlDataAdapter();
SqlDataAdapter flightsAdapter = new SqlDataAdapter();
SqlDataAdapter planetypeAdapter = new SqlDataAdapter();
SqlDataAdapter customersAdapter = new SqlDataAdapter();
SqlDataAdapter reservationsAdapter = new SqlDataAdapter();
Then the various select commands to fetch the data are created and the dataset is filled
with the data from those tables:
airlinesAdapter.SelectCommand = new SqlCommand(

"select * from Airlines", conn);
airlinesAdapter.Fill(d, "Airlines");
flightsAdapter.SelectCommand = new SqlCommand(

"select * from Flights", conn);
flightsAdapter.Fill(d, "Flights");
planetypeAdapter.SelectCommand = new SqlCommand(

"select * from PlaneType", conn);
planetypeAdapter.Fill(d, "PlaneType");
customersAdapter.SelectCommand = new SqlCommand(

"select * from Customers", conn);
customersAdapter.Fill(d, "Customers");
reservationsAdapter.SelectCommand = new SqlCommand(

"select * from Reservations", conn);
reservationsAdapter.Fill(d, "Reservations");
We now have the data for the Airlines, Flights, PlaneType, Customers, and Reservations
tables in the data set.
Next we have the DataSet written out as an XML schema, the schema it infers from the
data. Then the DataSet writes out the data as XML.
d.WriteXmlSchema("Airlines.xsd");
d.WriteXml("Airlines.xml");
Here are some of the data that were written to the file Airlines.xml. The main element is
Airline Broker, which was the name of the DataSet. Elements at the next lower level
correspond to the various tables that were added to the database: Airlines, Flights,
PlaneType, and Customers. There were no reservations in the database. There is one set
for each row in the table. The elements under each of these tables correspond to the fields
for that particular row.
<AirlineBroker>
<Airlines>
<Name>America West</Name>
<Abbreviation>AW</Abbreviation>
<WebSite>www.americawest.com</WebSite>
<ReservationNumber>555-555-1212</ReservationNumber>
</Airlines>
<Airlines>
<Name>Delta</Name>
<Abbreviation>DL</Abbreviation>
<WebSite>www.delta.com</WebSite>
<ReservationNumber>800-456-7890</ReservationNumber>
</Airlines>
...
<Flights>
<Airline>DL</Airline>
<FlightNumber>987</FlightNumber>
<StartCity>Atlanta</StartCity>
<EndCity>New Orleans</EndCity>
<Departure>2001-10-05T20:15:00.0000000-04:00
</Departure>
<Arrival>2001-10-05T22:30:00.0000000-04:00</Arrival>
<PlaneType>737</PlaneType>
<FirstCost>1300</FirstCost>
<BusinessCost>0</BusinessCost>
<EconomyCost>450</EconomyCost>
</Flights>
...
<Flights>
<PlaneType>
<PlaneType>737</PlaneType>
<FirstClass>10</FirstClass>
<BusinessClass>0</BusinessClass>
<EconomyClass>200</EconomyClass>
</PlaneType>
...
<Customers>
<LastName>Adams</LastName>
<FirstName>John</FirstName>
<EmailAddress>[email protected]</EmailAddress>
<CustomerId>1</CustomerId>
</Customers>
</AirlineBroker>
From the structure of the data, the DataSet deduces a schema that was written to
Airlines.xsd. We discuss here an excerpt from that file. There are no relationships or
primary keys defined between any of the tables such as Airlines and Flights as in the
database, because none were defined in the DataSet. If you look at the actual generated
file, you will see that schema information was inferred for Reservations even though
there were no data in the table.
The schema preamble in the first line, reproduced here, defines the name of the schema
as AirlineBroker, and we are using two namespaces in this schema document. One,
abbreviated xsd, contains the XML Schema standard definitions. The other, abbreviated
msdata, contains Microsoft definitions.
...
<xsd:schema id="AirlineBroker" targetNamespace="" xmlns=""
xmlns:xsd=http://www.w3.org/2001/XMLSchema
xmlns:msdata="urn:schemas-microsoft-com:xml-msdata">
The next line defines an element called AirlineBroker which has an attribute that
indicates this schema came from a DataSet. That is a Microsoft defined attribute, not one
defined by the W3C Schema namespace. This element AirlineBroker is a complex type.
which means it is a structure composed of other types. This structure can have an
unlimited number of any (or even none) of the types defined in the rest of the schema.
<xsd:element name="AirlineBroker" msdata:IsDataSet="true">

<xsd:complexType>
<xsd:choice maxOccurs="unbounded">
The Airlines element is defined next. It, too, is a structure, or complex type, whose
elements, if present, appear in the structure in the order in which they were defined.
Those elements, which correspond to the columns in the database table, are all defined to
be strings that are optional. No primary keys were defined, and these strings are certainly
not optional in the database, but that was what the DataSet deduced from the set of tables,
constraints and relationships currently defined in the DataSet.
<xsd:element name="Airlines">
<xsd:complexType>
<xsd:sequence>
<xsd:element name="Name" type="xsd:string"
minOccurs="0" />
<xsd:element name="Abbreviation"
type="xsd:string" minOccurs="0" />
<xsd:element name="WebSite" type="xsd:string"
minOccurs="0" />
<xsd:element name="ReservationNumber"
type="xsd:string" minOccurs="0" />
</xsd:sequence>
</xsd:complexType>
</xsd:element>
The table, Flights, is defined similarly to Airlines. In addition to there being no primary
key here, there is no foreign key defined for Airline or PlaneType.
<xsd:element name="Flights">
<xsd:complexType>
<xsd:sequence>
<xsd:element name="Airline" type="xsd:string"
minOccurs="0" />
<xsd:element name="FlightNumber" type="xsd:int"
minOccurs="0" />
<xsd:element name="StartCity" type="xsd:string"
minOccurs="0" />
<xsd:element name="EndCity" type="xsd:string"
minOccurs="0" />
<xsd:element name="Departure" type="xsd:dateTime"
minOccurs="0" />
<xsd:element name="Arrival" type="xsd:dateTime"
minOccurs="0" />
<xsd:element name="PlaneType" type="xsd:string"
minOccurs="0" />
<xsd:element name="FirstCost" type="xsd:decimal"
minOccurs="0" />
<xsd:element name="BusinessCost"
type="xsd:decimal" minOccurs="0" />
<xsd:element name="EconomyCost"
type="xsd:decimal" minOccurs="0" />
</xsd:sequence>
</xsd:complexType>
</xsd:element>
...
</xsd:choice>
</xsd:complexType>
</xsd:element>
</xsd:schema>
We will come back to this schema definition, but for the moment let us continue to work
with this example.
Creating an XML Doc from a Dataset
We create a new XML document from the DataSet. Using an XPath query to get the top
of the document, we set up an XmlNodeReader to read through it. We can then print out
the contents of the document to the console. The XmlNodeReader class knows how to
navigate through the document.
XmlDataDocument xmlDataDoc = new XmlDataDocument(d);
XmlNodeReader xmlNodeReader = null;

try
{
XmlNode node = xmlDataDoc.SelectSingleNode("/");
xmlNodeReader = new XmlNodeReader (node);
FormatXml (xmlNodeReader);
}
catch (Exception e)
{
Console.WriteLine ("Exception: {0}", e.ToString());
}
finally
{
if (xmlNodeReader != null)
xmlNodeReader.Close();
}
...
private static void FormatXml (XmlReader reader)
{
while (reader.Read())
{
switch (reader.NodeType)
{
...
case XmlNodeType.Element:
Format (reader, "Element");
while(reader.MoveToNextAttribute())
Format (reader, "Attribute");
break;
case XmlNodeType.Text:
Format (reader, "Text");
break;
...
static string lastNodeType = "";
private static void Format(XmlReader reader, string

nodeType)
{
if (nodeType == "Element")
{
if (lastNodeType == "Element")
{
}
for (int i=0; i < reader.Depth; i++)
{
Console.Write(" ");
}
Console.Write(reader.Name);
}
else if (nodeType == "Text")
Console.WriteLine("={0}", reader.Value);
else
{
Console.Write(nodeType + "<" + reader.Name + ">" +
reader.Value);
}
lastNodeType = nodeType;
}
The results resemble the XML that the DataSet wrote to a file.
AirlineBroker
Airlines
Name=America West
Abbreviation=AW
WebSite=www.americawest.com
ReservationNumber=555-555-1212
Airlines
Name=Delta
Abbreviation=DL
WebSite=www.delta.com
Airlines
Name=Northwest
Abbreviation=NW
WebSite=www.northwest.com
Airlines
Name=Piedmont
Abbreviation=P
WebSite=www.piedmont.com
Airlines
Name=Southwest
Abbreviation=S
WebSite=www.southwest.com
ReservationNumber=1-800-111-222
Airlines
Name=United
Abbreviation=UAL
WebSite=www.ual.com
Flights
Airline=DL
FlightNumber=987
StartCity=Atlanta
EndCity=New Orleans
Departure=2001-10-05T20:15:00.0000000-04:00
Arrival=2001-10-05T22:30:00.0000000-04:00
FirstCost=1300
PlaneType=737
BusinessCost=0
EconomyCost=450
Flights
Airline=UAL
FlightNumber=54
EndCity=Los Angeles
StartCity=Boston
Departure=2001-10-01T10:00:00.0000000-04:00
Arrival=2001-10-01T13:00:00.0000000-04:00
PlaneType=767
FirstCost=1500
BusinessCost=1000
EconomyCost=300
PlaneType
PlaneType=737
FirstClass=10
BusinessClass=0
EconomyClass=200
PlaneType
PlaneType=767
FirstClass=10
BusinessClass=30
EconomyClass=300
Customers
LastName=Adams
FirstName=John
[email protected]
CustomerId=1
Schema with Relationships
If we add relationships to the schema we just created, we can use the schema to create a
typed data class to work with our database.
We could do that programmatically by adding constraints and relationships to the dataset,

as discussed earlier in the chapter, and then writing out the schema. The DataSchemaXml
example does just that. You could also create a schema document by hand, or edit the one
we generated in the previous example.
The XSD Tool directory has a schema which has been revised to add the relationships
between the tables in the AirlineBroker database. The first part of the file,
AirlineBroker.xsd, looks like the previous version except that the minOccurs=0 attribute
has been removed from all the fields because we do not allow nulls in any of them.
...
<xsd:element name="Airlines">
<xsd:complexType>
<xsd:sequence>
<xsd:element name="Name" type="xsd:string" />
<xsd:element name="Abbreviation"
type="xsd:string" />
<xsd:element name="ReservationNumber"
type="xsd:string" />
<xsd:element name="WebSite" type="xsd:string" />
</xsd:sequence>
</xsd:complexType>
</xsd:element>
...
The last section defines the relationships. Here is the definition for the Airlines table
primary key. Note the use of attributes in the msdata namespace. These attributes are
defined by Microsoft using the W3C Schema standard to express additional semantic
information about the DataSet. These extensions themselves are not a W3C standard. The
Schema standard can express constraints with the unique, key, or keyref constructs.
Nonetheless, they do not specify which unique key is the primary key.
XPath, which is used to specify relationships to other tables and fields is a W3C standard
for locating elements within an XML file. It is used when an XML constraint has to
specify to which other element it refers to.
The primary key definition states that the Airlines_PrimaryKey is a primary key defined
for the Airlines element, consisting of the sub element, Name. Note how the
msdata:PrimaryKey attribute is used in conjunction with the standard unique construct.
<xsd:unique name="Airlines_PrimaryKey"
msdata:PrimaryKey="true">
<xsd:selector xpath=".//Airlines" />
<xsd:field xpath="Name" />
</xsd:unique>
The next section constrains the Abbreviation column in an Airlines row to be unique.
<xsd:unique name="Unique_Airline_Abbreviation">
<xsd:selector xpath=".//Airlines" />
<xsd:field xpath="Abbreviation" />
</xsd:unique>
...
Reservations_x0020_CustomerId is defined to be a foreign key. The CustomerId field in

the Reservations table must be found in the CustomerId field of some row in the
Customer table.
<xsd:keyref name="Reservations_x0020_CustomerId"
refer="Customers_PrimaryKey"
<xsd:selector xpath=".//Reservations" />

<xsd:field xpath="CustomerId" />
</xsd:keyref>
The foreign key Flights_x0020_Abbrev has some rules defined for it.
<xsd:keyref name="Flights_x0020_Abbrev"
refer="Unique_Airline_Abbreviation"
msdata:AcceptRejectRule="Cascade"
msdata:DeleteRule="SetNull">
<xsd:selector xpath=".//Flights" />
<xsd:field xpath="Airline" />
</xsd:keyref>
...
Typed DataSet
An XML schema can be used to generate a dataset that is "typed." Instead of using the
index property of a collection to access an element of the dataset, you can use the name
of a column. Here is a fragment from the TypedDataSet example:
AirlineBroker.AirlinesRow UAL = a.FindByName("United");

Console.WriteLine("{0}({1}) ReservationNumber:{2}
WebSite:{3}", UAL.Name.Trim(), UAL.Abbreviation.Trim(),
UAL.ReservationNumber.Trim(), UAL.WebSite.Trim());
You can assign a meaningful name to rows as well as use strong typing to make sure you
are working with the data element you want to. If you try to set the field
UAL.ReservationNumber to an integer, the compiler will detect the mistake.
A typed DataSet inherits from the DataSet class, so that everything that is available in a
DataSet is available in a typed DataSet. If the schema of the database changes, however,
the typed dataset class must be regenerated.
Generating Typed DataSets
The XML Schema Definition Tool (Xsd.exe) is used to transform an XML schema (XSD)
to a typed data set. The syntax for doing this is:
Xsd.exe /d /l:C# filename.xsd
The /d switch indicates that a DataSet should be generated. The /l switch indicates that a
C# class should be generated.
The XSD Tool directory has a batch file that can be used to take the revised
AirlineBroker XSD and generate a typed dataset AirlineBroker.cs.
Fetching Data with a Typed DataSet
The TypedDataSet example shows how to use a typed dataset to access the Airline
Brokers database. You define your SqlConnection as usual and create an SqlDataAdapter
instance for each table you want to use. You create whatever SqlCommands you need to
work with the data. A typed DataSet is independent of a database, just like the untyped
DataSet, so it needs SqlDataAdapter to handle the database operations.
SqlConnection conn = new SqlConnection(connectString);
SqlDataAdapter airlinesAdapter = new SqlDataAdapter();

SqlDataAdapter flightsAdapter = new SqlDataAdapter();
SqlDataAdapter planetypeAdapter = new SqlDataAdapter();
SqlDataAdapter customersAdapter = new SqlDataAdapter();
SqlDataAdapter reservationsAdapter = new SqlDataAdapter();
AirlineBroker airlineBrokerDataset = new AirlineBroker();
Next the select commands are defined to fetch the data, just as for use with a regular
DataSet. For illustrative purposes, constraint checking is enabled even though it is on by
default.
airlinesAdapter.SelectCommand = new SqlCommand(

"select * from Airlines", conn);
airlinesAdapter.InsertCommand = new SqlCommand(
"insert Airlines(Name, Abbreviation, WebSite,
ReservationNumber) values(@Name, @Abbrev, @Web,
@Reserve)", conn);
airlinesAdapter.InsertCommand.CommandType =
CommandType.Text;
SqlParameter param = new SqlParameter("@Name",

SqlDbType.NChar, 40);
airlinesAdapter.InsertCommand.Parameters.Add(param);
airlinesAdapter.InsertCommand.Parameters["@Name"].
SourceColumn = "Name";
...
airlineBrokerDataset.EnforceConstraints = true;
...
Now you can fetch the data. The order is which you do this is important. If Flights data
are fetched before PlaneType data, a constraint violation exception will occur, because
the PlaneType field in the Flights table does not exist.
airlinesAdapter.Fill(airlineBrokerDataset, "Airlines");
planetypeAdapter.Fill(airlineBrokerDataset, "PlaneType");
flightsAdapter.Fill(airlineBrokerDataset, "Flights");
customersAdapter.Fill(airlineBrokerDataset, "Customers");
reservationsAdapter.Fill(airlineBrokerDataset,
"Reservations");
Displaying Data with a Typed Dataset
The strong typing makes it straightforward to display the data:
AirlineBroker.AirlinesDataTable a =
airlineBrokerDataset.Airlines;
Console.WriteLine(a.TableName);
Console.WriteLine(" {0, -18} {1, -20} {2, -20}
{3, -15}", "Name", "Abbreviation", "Web Site",
"Reservation Numbers");
for (int i = 0; i < a.Count; i++)
Console.WriteLine(" {0, -18} {1, -20} {2, -20}
{3, -15}", a[i].Name.Trim(), a[i].Abbreviation.Trim(),
a[i].WebSite.Trim(), a[i].ReservationNumber.Trim());
...
It is easy to locate the data:
AirlineBroker.AirlinesRow UAL = a.FindByName("United");

Console.WriteLine("{0}({1}) ReservationNumber:{2}
WebSite:{3}", UAL.Name.Trim(), UAL.Abbreviation.Trim(),
UAL.ReservationNumber.Trim(), UAL.WebSite.Trim());
...
Modify Data with a Typed Dataset
You modify and update the database with a typed dataset just like a regular dataset. Make
sure the correct table is specified in the Update method.
airlineBrokerDataset.Airlines.AddAirlinesRow("Southwest",
"S", "1-800-111-222", "www.southwest.com");
NumberRows = airlinesAdapter.Update(airlineBrokerDataset,
"Airlines");
if (NumberRows == 1)
Console.WriteLine("Southwest added.");
else
Console.WriteLine("Southwest not added");
Summary
ADO.NET provides classes that enable you to design and build a distributed data
architecture. You can access databases in a connected or disconnected mode depending
on your concurrency requirements. The DataSet enables you to work with data in a
relational manner without being connected to any data source. XML can be used to model
relational data inside an XML document that contains nonrelational information. A typed
DataSet gives you the ability to work in a much easier, type-safe fashion with a DataSet,
provided you have an XML Schema that defines your data.
Chapter 10. ASP.NET and Web Forms Top
An important part of .NET is its use in creating Web applications through a technology
known as ASP.NET. Far more than an incremental enhancement to Active Server Pages
(ASP), the new technology is a unified Web development platform that greatly simplifies
the implementation of sophisticated Web applications. In this chapter we introduce the
fundamentals of ASP.NET and cover Web Forms, which make it easy to develop
interactive Web sites. In Chapter 11 we cover Web Services, which enable the
development of collaborative Web applications that span heterogeneous systems.
What Is ASP.NET?
We begin our exploration of ASP.NET by looking at a very simple Web application.

Along the way we will establish a test bed for ASP.NET programming, and we will
review some of the fundamentals of Web processing. Our little example will reveal some
of the challenges in developing Web applications, and we can then appreciate the features
and benefits of ASP.NET, which we will elaborate in the rest of the chapter.
Web Application Fundamentals
A Web application consists of document and code pages in various formats. The simplest
kind of document is a static HTML page, which contains information that will be
formatted and displayed by a Web browser. An HTML page may also contain hyperlinks
to other HTML pages. A hyperlink (or just "link") contains an address, or a Uniform
Resource Locator (URL), specifying where the target document is located. The resulting
combination of content and links is sometimes called "hypertext" and provides easy
navigation to a vast amount of information on the World Wide Web.
Setting up the Web Examples
As usual, all the example programs for this chapter are in the chapter folder. To run the
examples, you will need to have Internet Information Services (IIS) installed on your
system. IIS is installed by default with Windows 2000 Server. You will have to explicitly
install it with Windows 2000 Workstation. Once installed, you can access the
documentation on IIS through Internet Explorer via the URL http://localhost, which will
redirect you to the starting IIS documentation page, as illustrated in Figure 10-1.
Figure 10-1. Internet Information Services documentation.
The management tool for IIS is a Microsoft Management Console (MMC) "snap-in," the
Internet Services Manager, which you can find under Administrative Tools in the Control
Panel. Figure 10-2 shows the main window of the Internet Services Manager. You can
Start and Stop the Web server and perform other tasks by right-clicking on Default Web
Site. Choosing Properties from the context menu will let you perform a number of
configurations on the Web server.
Figure 10-2. Internet Services Manager.
The default home directory for publishing Web files is \Inetpub\ wwwroot on the drive
where Windows is installed. You can change this home directory using Internet Services
Manager. You can access Web pages stored at any location on your hard drive by creating
a "virtual directory." The easiest way to create one is from Windows Explorer. Right-click
over the desired directory, choose Sharing..., select the Web Sharing tab, click on the Add
button, and enter the desired alias, which will be the name of the virtual directory. Figure
10-3 illustrates creating an alias NetCs, or virtual directory, for the folder
\OI\NetCs\Chap10. You should perform this operation now on your own system in order
that you may follow along as the chapter's examples are discussed.
Figure 10-3. Creating a virtual directory.

Once a virtual directory has been created, you can access files in it by including the
virtual directory in the path of the URL. In particular, you can access the file default.htm
using the URL http://localhost/NetCs/. The file default.htm contains a home page
for all the ASP.NET example programs for this chapter. See Figure 10-4.
Figure 10-4. Home page for ASP.NET example programs.

An Echo Program
The first example program for this chapter is Hello.aspx, shown as a link on the home
page. The example is complete in one file and contains embedded server code. Here is the
source code, which consists of HTML along with some C# script code. There are also
some special tags for "server controls," recognized by ASP.NET.


<%@ Page Language="C#" %>
<HTML>
<HEAD>
<SCRIPT RUNAT="SERVER">
protected void cmdEcho_Click(object Source, EventArgs e)
{
lblGreeting.Text="Hello, " + txtName.Text;
}
</SCRIPT>
</HEAD>
<BODY>
<FORM RUNAT="SERVER">Your name:
<asp:textbox id=txtName Runat="server"></asp:textbox>
<p><asp:button id=cmdEcho onclick=cmdEcho_Click Text="Echo"
runat="server" tooltip="Click to echo your name">
</asp:button></p>
<asp:label id=lblGreeting runat="server"></asp:label>
<P></P>
</FORM>
</BODY>
</HTML>
You can run the program using the URL http://localhost/ NetCs/Hello.aspx or by
clicking on the link Hello.aspx in the home page of the examples programs. The page
shows a text box where you can type in your name, and there is an "Echo" button.
Clicking the button will echo your name back, with a "Hello" greeting. The simple form
is again displayed, so you could try out other names. If you slide the browser's mouse
cursor over the button, you will see the tool tip "Click to echo your name" displayed in a
yellow box. Figure 10-5 illustrates a run of this example.
Figure 10-5. Running the Hello.aspx echo program.
This little program would not be completely trivial to implement with other Web
application tools, including ASP. The key user-interface feature of such an application is
its thoroughly forms-based nature. The user is presented with a form and interacts with
the form. The server does some processing, and the user continues to see the same form.
This UI model is second nature in desktop applications but is not so common in Web
applications. Typically the Web server will send back a different page.
This kind of application could certainly be implemented using a technology like ASP, but
the code would be a little ugly. The server would need to synthesize a new page that
looked like the old page, creating the HTML tags for the original page, plus extra
information sent back (such as the greeting shown at the bottom in our echo example). A
mechanism is needed to remember the current data that is displayed in the controls in the
form.
Another feature of this Web application is that it does some client-side processing too—
the "tooltip" displayed in the yellow box is performed by the browser. Such rich client-
side processing can be performed by some browsers, such as Internet Explorer, but not
others.
As can be seen by the example code, with ASP.NET it is very easy to implement this kind
of Web application. We will study the code in detail later. For now, just observe how easy
it is!
ASP.NET Features
ASP.NET provides a programming model and infrastructure that facilitates developing

new classes of Web applications. Part of this infrastructure is the .NET runtime and
framework. Server-side code is written in .NET compiled languages. Two main
programming models are supported by ASP.NET.
• Web Forms helps you build form-based Web pages. A WYSIWYG development
environment enables you to drag controls onto Web pages. Special "server-side"
controls present the programmer with an event model similar to what is provided
by controls in ordinary Windows programming. This chapter discusses Web
Forms in detail.
• Web Services make it possible for a Web site to expose functionality via an API
that can be called remotely by other applications. Data is exchanged using
standard Web protocols and formats such as HTTP and XML, which will cross
firewalls. We will discuss Web Services in the next chapter.
Both Web Forms and Web Services can take advantage of the facilities provided by .NET,
such as the compiled code and .NET runtime. In addition, ASP.NET itself provides a
number of infrastructure services, including state management, security, configuration,
caching, and tracing.
Compiled Code
Web Forms (and Web Services) can be written in any .NET language that runs on top of
the CLR, including C#, VB.NET, and C++ with Managed Extensions. This code is
compiled, and thus offers better performance than ASP pages with code written in an
interpreted scripting language such as VBScript. All of the benefits, such as a managed
execution environment, are available to this code, and of course the entire .NET
Framework Class Library is available. Legacy unmanaged code can be called through the
.NET interoperability services, which are discussed in Chapter 14.
Server Controls
ASP.NET provides a significant innovation known as "server controls." These controls

have special tags such as <asp:textbox>. Server-side code interacts with these controls,
and the ASP.NET runtime generates straight HTML that is sent to the Web browser. The
result is a programming model that is easy to use and yet produces standard HTML that
can run in any browser.
Browser Independence
Although the World Wide Web is built on standards, the unfortunate fact of life is that
browsers are not compatible and have special features. A Web page designer then has the
unattractive options of either writing to a lowest common denominator of browser, or else
writing special code for different browsers. Server controls help remove some of this
pain. ASP.NET takes care of browser compatibility issues when it generates code for a
server control. If the requesting browser is upscale, the generated HTML can take
advantage of these features, otherwise the generated code will be vanilla HTML.
ASP.NET takes care of detecting the type of browser.
Separation of Code and Content
Typical ASP pages have a mixture of scripting code interspersed with HTML elements. In
ASP.NET there is a clean separation between code and presentation content. The server
code can be isolated within a single <SCRIPT RUNAT="SERVER"> ... /SCRIPT> block
or, even better, placed within a "code behind" page. We will discuss "code behind" pages
later in this chapter. If you would like to see an example right away, you can examine the
second example program HelloCodebehind.aspx, with code in the file HelloCodebehind.
aspx.cs. (These files are in the top-level chapter directory.)
State Management
HTTP is a stateless protocol. Thus, if a user enters information in various controls on a

form, and sends this filled-out form to the server, the information will be lost if the form
is displayed again, unless the Web application provides special code to preserve this state.
ASP.NET makes this kind of state preservation totally transparent. There are also
convenient facilities for managing other types of session and application state.
Web Forms Architecture
A Web Form consists of two parts:
• The visual content or presentation, typically specified by HTML elements

• Code that contains the logic for interacting with the visual elements.
A Web Form is physically expressed by a file with the extension .aspx. Any HTML page
could be renamed to have this extension and could be accessed using the new extension
with identical results to the original. Thus Web Forms are upwardly compatible with
HTML pages.
The way code can be separated from the form is what makes a Web Form special. This
code can be either in a separate file (having an extension corresponding to a .NET
language, such as .cs for C#) or in the .aspx file, within a <SCRIPT
RUNAT="SERVER"> ... /SCRIPT> block. When your page is run in the Web server, the
user interface code runs and dynamically generates the output for the page.
We can understand the architecture of a Web Form most clearly by looking at the code-
behind version of our "echo" example. The visual content is specified by the .aspx file
HelloCodebehind.aspx.


<%@ Page Language="C#" Src="HelloCodebehind.aspx.cs"
Inherits= MyWebPage %>
<HTML>
<HEAD>
</HEAD>
<BODY>
<FORM RUNAT="SERVER">YOUR NAME:
</asp:button></p>
<P></P>
</FORM>
</BODY>
</HTML>
The user interface code is in the file HelloCodebehind.aspx.cs,
// HelloCodebehind.aspx.cs
using System;
using System.Web;
using System.Web.UI;
using System.Web.UI.WebControls;
public class MyWebPage : System.Web.UI.Page

{
protected TextBox txtName;
protected Button cmdEcho;
protected Label lblGreeting;

{
}
}
Page Class
The key namespace for Web Forms and Web Services is System.Web. Support for Web
Forms is in the namespace System.Web.UI. Support for server controls such as text boxes
and buttons is in the namespace System.Web.UI.WebControls. The class that dynamically
generates the output for an .aspx page is the Page class, in the System.Web.UI
namespace, and classes derived from Page, as illustrated in the code behind page in this
last example.
Inheriting from Page Class
The elements in the .aspx file, the code in the code-behind file (or script block), and the
base Page class work together to generate the page output. This cooperation is achieved
by ASP.NET's dynamically creating a class for the .aspx file, which is derived from the
"code-behind" class, which in turn is derived from Page. This relationship is created by
the "Inherits" attribute in the .aspx file. Figure 10-6 illustrates the inheritance hierarchy.
Here MyWebPage is a class we implement, derived from Page.
Figure 10-6. Hierarchy of page classes.
The most derived page class, shown as "My .aspx Page" in Figure 10-6, is dynamically
created by the ASP.NET runtime. This class extends the page class, shown as
"MyWebPage" in the figure, to incorporate the controls and HTML text on the Web
Form. This class is compiled into an executable, which is run when the page is requested
from a browser. The executable code creates the HTML that is sent to the browser.
Web Forms Page Life Cycle
We can get a good high-level understanding of the Web Forms architecture by following
the life cycle of our simple Echo application. We will use the code-behind version (the
second example), HelloCodebehind.aspx.
1. User requests the HelloCodebehind.aspx Web page in the browser.

2. Web server compiles the page class from the .aspx file and its associated code
behind page. The Web server executes the code, creating HTML, which is sent to
the browser. (In Internet Explorer you can see the HTML code from the menu
View | Source.) Note that the server controls are replaced by straight HTML. The
following code is what arrives at the browser, not the original code on the server.
3. 
4.
5. <HTML>
6. <HEAD>
7. </HEAD>
8. <BODY>
9. <form name="ctrl0" method="post"
10.action="HelloCodebehind.aspx" id="ctrl0">
11.<input type="hidden" name="__VIEWSTATE"
12.value="dDwxMzc4MDMwNTk1Ozs+" />
13.YOUR NAME: <input name="txtName" type="text"
14.id="txtName" />
15.<p><input type="submit" name="cmdEcho" value="Echo"
16.id="cmdEcho" title="Click to echo your name" /></p>
17. <span id="lblGreeting"></span>
18.<P></P>
19.</form>
20.</BODY>
</HTML>
21. The browser renders the HTML, displaying the simple form shown in Figure 10-
7. To distinguish this example from the first one, we show "YOUR NAME" in all
capitals. Since this is the first time the form is displayed, the text box is empty,
and no greeting message is displayed.
Figure 10-7. The form for the "Echo" application is displayed for the
first time.
22. The user types in a name (e.g., "Mary Smith") and clicks the "Echo" button. The
browser recognizes that a Submit button has been clicked. The method for the
form is "post" [1] and the action is "HelloCodebehind.aspx." We thus have what is
called a "post back" to the original .aspx file.
[1]
The HTTP POST method sends form results separately as part of the data body,
rather than by concatenating it onto the URL, as is done in the GET method.
23. The server now performs processing for this page. An event was raised when the
user clicked the "Echo" button, and an event handler in the MyWebPage class is
invoked.
24.protected void cmdEcho_Click(object Source, EventArgs e)
25.{
26. lblGreeting.Text="Hello, " + txtName.Text;
}
27. The Text property of the TextBox server control txtName is used to read the name
submitted by the user. A greeting string is composed and assigned to the Label
control lblGreeting, again using property notation.
28. The server again generates straight HTML for the server controls and sends the
whole response to the browser. Here is the HTML.
29....
30.<form name="ctrl0" method="post"
31.action="HelloCodebehind.aspx" id="ctrl0">
32.<input type="hidden" name="__VIEWSTATE"
33.value="dDwxMzc4MDMwNTk1O3Q8O2w8aTwyPjs+O2w8dDw7bDxpPDU+Oz47b
34.Dx0PHA8cDxsPFRleHQ7PjtsPEhlbGxvLCBNYXJ5IFNtaXRoOz4+Oz47Oz47P
35.j47Pj47Pg==" />
36.YOUR NAME: <input name="txtName" type="text"
37.value="Mary Smith" id="txtName" />
38.<p><input type="submit" name="cmdEcho" value="Echo"
39.id="cmdEcho" title="Click to echo your name" /></p>
40. <span id="lblGreeting">Hello, Mary Smith</span>
...
41. The browser renders the page, as shown in Figure 10-8. Now a greeting message
is displayed.
Figure 10-8. After a round trip a greeting message is displayed.
View State
An important characteristic of Web Forms is that all information on forms is

"remembered" by the Web server. Since HTTP is a stateless protocol, this preservation of
state does not happen automatically but must be programmed. A nice feature of ASP.NET
is that this state information, referred to as "view state," is preserved automatically by the
Framework, using a "hidden" control.
...
<input type="hidden" name="__VIEWSTATE"
value="dDwxMzc4MDMwNTk1O3Q8O2w8aTwyPjs+O2w8dDw7bDxpPDU+Oz47b
Dx0PHA8cDxsPFRleHQ7PjtsPEhlbGxvLCBNYXJ5IFNtaXRoOz4+Oz47Oz47P
j47Pj47Pg==" />
...
Later in the chapter we will examine other facilities provided by ASP.NET for managing
session state and application state.
Web Forms Event Model
From the standpoint of the programmer, the event model for Web Forms is very similar to
the event model for Windows Forms. Indeed, this similarity is what makes programming
with Web Forms so easy. What is actually happening in the case of Web Forms, though, is
rather different. The big difference is that events get raised on the client and processed on
the server. [2]
[2]
Some controls, such as the Calendar control, raise some events on the server. Also, the
Page itself raises events on the server.
Our simple form with one text box and one button is not rich enough to illustrate event
processing very thoroughly. Let's imagine a more elaborate form with several text boxes,
list boxes, check boxes, buttons, and the like. Because round trips to the server are
expensive, events do not automatically cause a postback to the server. Server controls
have what is known as an intrinsic event set of events that automatically cause a postback
to the server. The most common such intrinsic event is a button click. Other events, such
as selecting an item in a list box, do not cause an immediate postback to the server.
Instead, these events are cached, until a button click causes a post to the server. Then, on
the server the various change events are processed, in no particular order, and the button-
click event that caused the post is processed.
Page Processing
Processing a page is a cooperative endeavor between the Web server, the ASP.NET
runtime, and your own code. The Page class provides a number of events, which you can
handle to hook into page processing. The Page class also has properties and methods that
you can use. We cover some of the major ones here. For a complete description, consult
the .NET Framework documentation. The example programs in this chapter will illustrate
features of the Page class.
Page Events
A number of events are raised on the server as part of the normal processing of a page.
These events are actually defined in the Control base class and so are available to server
controls also. The most important ones are listed below.
• Init is the first step in the page's life cycle and occurs when the page is initialized.
There is no view-state information for any of the controls at this point.
• Load occurs when the controls are loaded into the page. View-state information
for the controls is now available.
• PreRender occurs just before the controls are rendered to the output stream.
Normally this event is not handled by a page but is important for implementing
your own server controls.
• Unload occurs when the controls are unloaded from the page. At this point it is
too late to write your own data to the output stream.
Page Properties
The Page class has a number of important properties. Some of the most useful are listed
below.
• EnableViewState indicates whether the page maintains view state for itself and its
controls. You can get or set this property. The default is true, view state is
maintained.
• ErrorPage specifies the error page to which the browser should be redirected in
case an unhandled exception occurs.
• IsPostBack indicates whether the page is being loaded in response to a postback
from the client or is being loaded for the first time.
• IsValid indicates whether page validation succeeded. [3]
[3]
We discuss validation later in the chapter, in the section on Server Controls.
• Request gets the HTTP Request object, which allows you to access data from
incoming HTTP requests.
• Response gets the HTTP Response object, which allows you to send response data
to a browser.
• Session gets the current Session object, which is provided by ASP.NET for storing
session state.
• Trace gets a TraceContext object for the page, which you can use to write out
trace information.
Sample Program
We can illustrate some of these features of page processing with a simple extension to our
Echo program. The page HelloPage.aspx (located in the top-level chapter directory)
provides handlers for a number of page events, and we write simple text to the output
stream, using the Response property. For each event we show the current text in the
txtName and lblGreeting server controls. In the handler for Load we also show the
current value of IsPostBack, which should be false the first time the page is accessed, and
subsequently true.


<%@ Page Language="C#" Debug="true" %>
<HTML>
<HEAD>
{
}
protected void Page_Init(Object sender, EventArgs E)
{
Response.Write("Page_Init<br>");
Response.Write("txtName = " + txtName.Text + "<br>");
Response.Write("lblGreeting = " + lblGreeting.Text +
"<br>");
}
protected void Page_Load(Object sender, EventArgs E)
{
Response.Write("Page_Load<br>");
Response.Write("IsPostBack = " + IsPostBack + "<br>");
"<br>");
}
protected void Page_PreRender(Object sender, EventArgs E)
{
Response.Write("Page_PreRender<br>");
"<br>");
}
protected void Page_Unload(Object sender, EventArgs E)
{
//Response not available in this context
//Response.Write("Page_Unload<br>");
}
</SCRIPT>
</HEAD>
<BODY>
</asp:button></p>
<P></P>
</FORM>
</BODY>
</HTML>
When we display the page the first time the output reflects the fact that both the text box
and the label are empty, since we have entered no information. IsPostBack is false.
Now enter a name and click the "Echo" button. We obtain the following output from our
handlers for the page events:
Page_Init
txtName =
lblGreeting =
Page_Load
IsPostBack = True
txtName = Robert
lblGreeting =
Page_PreRender
txtName = Robert
lblGreeting = Hello, Robert
In Page_Init there is no information for either control, since view state is not available at
page initialization. In Page_Load the text box has data, but the label does not, since the
click-event handler has not yet been invoked. IsPostBack is now true. In Page_PreRender
both controls now have data.
Click "Echo" a second time. Again, the controls have no data in Page_Init. This time,
however, in Page_Load the view state provides data for both controls. Figure 10-9 shows
the browser output after "Echo" has been clicked a second time.
Figure 10-9. Browser output after "Echo" has been clicked a second time.
Page Directive
An .aspx file may contain a page directive defining various attributes that can control
how ASP.NET processes the page. A page directive contains one or more attribute/value
pairs of the form
attribute="value"
within the page directive syntax
<@ Page ... @>

Our example program HelloCodebehind.aspx illustrates an .aspx page that does not have
any code within it. The "code-behind" file HelloCodebehind.aspx.cs that has the code is
specified using the Src attribute.


<%@ Page Language="C#" Src="HelloCodebehind.aspx.cs"
Inherits=MyWebPage %>
...
Src
The Src attribute identifies the code-behind file.
Language
The Language attribute specifies the language used for the page. The code in this
language may be in either a code-behind file or a SCRIPT block within the same file.
Values can be any .NET-supported language, including C# and VB.NET.
Inherits
The Inherits directive specifies the page class from which the .aspx page class will
inherit.
Debug
The Debug attribute indicates whether the page should be compiled with debug
information. If true, debug information is enabled, and the browser can provide detailed
information about compile errors. The default is false.
ErrorPage
The ErrorPage attribute specifies a target URL to which the browser will be redirected in
the event that an unhandled exception occurs on the page.
Trace
The Trace attribute indicates whether tracing is enabled. A value of true turns tracing on.
The default is false.
Tracing
ASP.NET provides extensive tracing capabilities. Merely setting the Trace attribute for a
page to true will cause trace output generated by ASP.NET to be sent to the browser. In
addition, you can output your own trace information using the Write method of the
TraceContext object, which is obtained from the Trace property of the Page.
The page HelloTrace.aspx illustrates using tracing in place of writing to the Response
object.


<%@ Page Language="C#" Debug="true" Trace = "true" %>
<HTML>
<HEAD>
{
}
protected void Page_Init(Object sender, EventArgs E)
{
Trace.Write("Page_Init<br>");
Trace.Write("txtName = " + txtName.Text + "<br>");
Trace.Write("lblGreeting = " + lblGreeting.Text +
"<br>");
}
...
Figure 10-10 shows the browser output after the initial request for the page. Notice that
the trace output is shown after the form, along with trace information that is generated by
ASP.NET itself.
Figure 10-10. Browser output showing trace information.

Request/Response Programming
The server control architecture is built on top of a more fundamental processing

architecture, which may be called "request/response." Understanding request/response is
important to solidify our overall grasp of ASP.NET. Also, in certain programming
situations request/response is the natural approach.
HttpRequest Class
The System.Web namespace contains a useful class HttpRequest that can be used to read
the various HTTP values sent by a client during a Web request. These HTTP values
would be used by a classical CGI program in acting upon a Web request, and they are the
foundation upon which higherlevel processing is built. Table 10-1 shows some of the
public instance properties of HttpRequest. If you are familiar with HTTP, the meaning of
these various properties should be largely self-explanatory. Refer to the .NET Framework
documentation of the HttpRequest class for full details about these and other properties.
Table 10-1. Public Instance Properties of HttpRequest

Property Meaning
AcceptTypes String array of client-supported MIME accept types
Browser Information about client's browser capabilities
ContentLength Length in bytes of content sent by the client
Cookies Collection of cookies sent by the client
Form Collection of form variables
Headers Collection of HTTP headers
HttpMethod HTTP transfer method used by client (e.g., GET or POST)
Params Combined collection of QueryString, Form, ServerVariables, and
Cookies items
Path Virtual request of the current path
QueryString Collection of HTTP query string variables
ServerVariables Collection of Web server variables
The Request property of the Page class returns a HttpRequest object. You may then
extract whatever information you need, using the properties of HttpRequest. For example,
the following code determines the length in bytes of content sent by the client and writes
that information to the Response object.
int length = Request.ContentLength;
Response.Write("ContentLength = " + length + "<br>");
Collections
A number of useful collections are exposed as properties of HttpRequest. The collections

are of type NamedValueCollection (in System. Collections.Specialized namespace). You
can access a value from a string key. For example, the following code extracts values for
the QUERY_STRING and HTTP_USER_AGENT server variables using the
ServerVariables collection.
string strQuery =
Request.ServerVariables["QUERY_STRING"];
string strAgent =
Request.ServerVariables["HTTP_USER_AGENT"];
Server variables such as these are at the heart of classical Common Gateway Interface
(CGI) Web server programming. The Web server passes information to a CGI script or
program by using environment variables. ASP.NET makes this low-level information
available to you, in case you need it.
A common task is to extract information from controls on forms. In HTML, controls are
identified by a name attribute, which can be used by the server to determine the
corresponding value. The way in which form data is passed to the server depends on
whether the form uses the HTTP GET method or the POST method.
With GET, the form data is encoded as part of the query string. The QueryString
collection can then be used to retrieve the values. With POST, the form data is passed as
content after the HTTP header. The Forms collection can then be used to extract the
control values. You could use the value of the REQUEST_METHOD server variable
(GET or POST) to determine which collection to use (the QueryString collection in the
case of GET and the Forms collection in case of POST).
With ASP.NET you don't have to worry about which HTTP method was used in the
request. ASP.NET provides a Params collection, which is a combination (union in the
mathematical sense) of the ServerVariables, QueryString, Forms, and Cookies
collections.
Example Program
We illustrate all these ideas with a simple page Squares.aspx that displays a column of
squares. How many squares to display is determined by a number submitted on a form.
The page GetSquares.aspx submits the request using GET, and PostSquares.aspx submits
the request using POST. These two pages have the same user interface, illustrated in
Figure 10-11.
Figure 10-11. Form for requesting a column of squares.
Here is the HTML for GetSquares.aspx. Notice that we are using straight HTML. Except
for the Page directive, which turns tracing on, no features of ASP.NET are used.


<%@ Page Trace = "true" %>
<html>
<head>
</head>
<body>
<P>This program will print a column of squares</P>
<form method="get" action = Squares.aspx>
How many:
<INPUT type=text size=2 value=5 name=txtCount>
<P></P>
<INPUT type=submit value=Squares name=cmdSquares>
</form>
</body>
</html>
The form tag has attributes specifying the method (GET or POST) and the action (target
page). The controls have a name attribute, which will be used by server code to retrieve
the value.
Run GetSquares.aspx and click "Squares." You will see some HTTP information
displayed, followed by the column of squares. Tracing is turned on, so details about the
request are displayed by ASP.NET. Figure 10-12 illustrates the output from this GET
request.
Figure 10-12. Output from a GET request.
You can see that form data is encoded in the query string, and the content length is 0. If
you scroll down on the trace output, you will see much information. For example, the
QueryString collection is shown.
Now run PostSquares.aspx and click "Squares." Again you will then see some HTTP
information displayed, followed by the column of squares. Tracing is turned on, so details
about the request are displayed by ASP.NET. Figure 10-13 illustrates the output from this
POST request.
Figure 10-13. Output from a POST request.

You can see that now the query string is empty, and the content length is 29. The form
data is passed as part of the content, following the HTTP header information. If you scroll
down on the trace output, you will see that now there is a Form collection, which is used
by ASP.NET to provide access to the form data in the case of a POST method.
By comparing the output of these two examples, you can clearly see the difference
between GET and POST, and you can also see the data structures used by ASP.NET to
make it easy for you to extract data from HTTP requests.
HttpResponse Class
The HttpResponse class encapsulates HTTP response information that is built as part of
an ASP.NET operation. The Framework uses this class when it is creating a response that
includes writing server controls back to the client. Your own server code may also use the
Write method of the Response object to write data to the output stream that will be sent to
the client. We have already seen many illustrations of Response.Write.
Redirect
The HttpResponse class has a useful method, Redirect, that enables server code to
redirect an HTTP request to a different URL. A simple redirection without passing any
data is trivial—you need only call the Redirect method and pass the URL. An example of
such usage would be a reorganization of a Web site, where a certain page is no longer
valid and the content has been moved to a new location. You can keep the old page live
by simply redirecting traffic to the new location.
It should be noted that redirection always involves an HTTP GET request, like following
a simple link to a URL. (POST arises as an option when submitting form data, where the
action can be specified as GET or POST.)
A more interesting case involves passing data to the new page. One way to pass data is to
encode it in the query string. You must preserve standard HTTP conventions for the
encoding of the query string. The class HttpUtility provides a method UrlEncode, which
will properly encode an individual item of a query string. You must yourself provide code
to separate the URL from the query string with a "?" and to separate items of the query
string with "&."
The folder Hotel provides an example of a simple Web application that illustrates this
method of passing data in redirection. The file default.aspx provides a form for collecting
information to be used in making a hotel reservation. The reservation itself is made on the
page Reservation1.aspx. You may access the starting default.aspx page through the URL
http://localhost/NetCs/Hotel/
As usual, we provide a link to this page in our home page of example programs. Figure
10-14 illustrates the starting page of our simple hotel reservation example.
Figure 10-14. Starting page for making a hotel reservation.

Here is the script code that is executed when the "Make Reservation" button is clicked.
private void cmdMakeReservation_Click(

object sender, System.EventArgs e)
{
string query = "City=" +
HttpUtility.UrlEncode(txtCity.Text);
query += "&Hotel=" +
HttpUtility.UrlEncode(txtHotel.Text);
query += "&Date=" +
HttpUtility.UrlEncode(txtDate.Text);
query += "&NumberDays=" +
HttpUtility.UrlEncode(txtNumberDays.Text);
Response.Redirect("Reservation1.aspx?" + query);
}
We build a query string, which gets appended to the Reservation1.aspx URL, separated
by a "?". Note the ampersand that is used as a separator of items in the query string. We
use the HttpUtility.UrlEncode method to encode the individual items. Special encoding is
required for the slashes in the date and for the space in the name "San Jose." Clicking the
button brings up the reservation page. You can see the query string in the address window
of the browser. Figure 10-15 illustrates the output shown by the browser. (Our program
does not actually make the reservation; it simply prints out the parameters passed to it.)
Figure 10-15. Browser output from making a hotel reservation.
You can turn on tracing, and the trace output should serve to reinforce the ideas we have
been discussing about request/response Web programming. In particular, you should
examine the QueryString collection, as illustrated in Figure 10-16.
Figure 10-16. The query string is used for passing parameters in redirection.
Web Applications Using Visual Studio.NET
We have examined the fundamentals of ASP.NET and have created some simple Web
pages. To carry the story further it will be very helpful to start using Visual Studio.NET.
Everything we do could also be accomplished using only the .NET Framework SDK, but
our work will be much easier using the facilities of Visual Studio. A special kind of
project, an "ASP.NET Web Application," creates the boilerplate code. The Forms
Designer makes it very easy to create Web forms by dragging controls from a palette. We
can add event handlers for controls in a manner very similar to the way event handlers are
added in Windows Forms. In fact, the whole Web application development process takes
on many of the rapid application development (RAD) characteristics typical of Visual
Basic.
In this section we will introduce the Web application development features of Visual
Studio by creating the first step of our Acme Travel Web site. We will elaborate on
specific features of ASP.NET in later sections.
Form Designers for Windows and Web

Applications
The basic look and feel of the Form Designers for Windows and Web
applications is the same. You drag controls from a toolbox. You set properties in
a Property window. You navigate between a code view and a designer view with
toolbar buttons. In the following discussion we assume you have a basic
familiarity with this visual paradigm. You may find it helpful to refer back to
Chapter 6.
Hotel Information Web Page (Step 0)
We begin by creating a simple Web page that will display information about hotels.
Dropdown listboxes are provided to show cities and hotels. Selecting a city from the first
dropdown will cause the hotels in that city to be shown in the second dropdown. We
obtain the hotel information from the Hotel.dll component, and we use data binding to
populate the listboxes. As a source for the Hotel.dll and Customer.dll components used
later, we provide a copy of the GUI application from Chapter 6, AcmeGui. The Hotel.dll
component we need in the following demonstration is in the folder
AcmeGui\Hotel\bin\Debug.
If you would like to follow along hands-on with Visual Studio, do your work in the
Demos folder for this chapter. The completed project is in CaseStudy\Step0.
Configuring Web Server Connection
Before getting started you may wish to check, and possibly change, your Visual Studio
Web Server Connection setting. The two options are File share and FrontPage. If you are
doing all your development on a local computer, you might find File share to be faster
and more convenient. To access this setting, select the Visual Studio menu Tools |
Options.… Choose Web Settings underneath Projects. You can then set the Preferred
Access Method by using a radio button, as illustrated in Figure 10-17.
Figure 10-17. Configuring Web server connection preferred access method.
Creating an ASP.NET Web Application

1. In Visual Studio select the menu File | New | Project.…
2. In the New Project dialog box choose "Visual C# Projects" as the Project Type and
"ASP.NET Web Application" as the Template.
3. Enter "AcmeWeb" as the name of your project. For the location enter an HTTP path
to a folder on your server machine. The default will be the IIS home directory
\Inetpub\wwwroot. If you have made \OI\NetCs\Chap10 into a virtual directory with
alias "NetCs", you can enter for the path http://localhost/NetCs/Demos, as
illustrated in Figure 10-18.
Figure 10-18. Creating a Visual Studio ASP.NET Web Application project.
4. Click OK. The project files will then be created in \OI\NetCs\ Chap10\Demos. The
VS.NET solution AcmeWeb.sln will then be created under MyDocuments\Visual
Studio Projects\AcmeWeb.
Using the Form Designer

1. Bring up the Toolbox from the View menu, if not already showing. Make sure the
Web Forms tab is selected.
2. Drag two Label controls and two DropDownList controls onto the form.
3. Change the Text property of the Labels to "City" and "Hotel." Resize the
DropDownList controls to look as shown in Figure 10-19.
Figure 10-19. Using the Form Designer to add controls to the form.
4. Change the (ID) of the DropDownList controls to listCities and listHotels.
Initializing the HotelBroker

1. Copy Hotel.dll from AcmeGui\Hotel\bin\Debug to Demos\ AcmeWeb\bin.
2. In your AcmeWeb, project add a reference to Hotel.dll.
3. As shown in the following code fragment, in Global.asax, add the following line near
the top of the file. (Use the View Code button to show the code.)
4. Add a public static variable hotelBroker of type HotelBroker.
5. Add code to Application_Start to instantiate HotelBroker.

// Global.asax
using System;
using System.ComponentModel;
using System.Web;
using System.Web.SessionState;
namespace AcmeWeb
{
/// <summary>
/// Summary description for Global.
/// </summary>
public class Global : System.Web.HttpApplication
{
public static HotelBroker hotelBroker;
protected void Application_Start(Object sender,
EventArgs e)
{
}
...
6. In WebForm1.aspx.cs add a using OI.NetCs.Acme; statement, and declare a static

variable hotelBroker of type HotelBroker.
...
namespace AcmeWeb
{
/// <summary>
/// Summary description for WebForm1.
/// </summary>
public class WebForm1 : System.Web.UI.Page
{
...
private static HotelBroker hotelBroker;
...
Data Binding
Next we will populate the first DropDownList with the city data, which can be obtained
by the GetCities method of HotelBroker. We make use of the data binding capability of
the DropDownList control. You might think data binding is only used with a database.
However, in .NET data binding is much more general, and can be applied to other data
sources besides databases. Binding a control to a database is very useful for two-tier,
client/server applications. However, we are implementing a three-tier application, in
which the presentation logic, whether implemented using Windows Forms or Web Forms,
talks to a business logic component and not directly to the database. So we will bind the
control to an ArrayList.
The .NET Framework provides a number of data binding options, which can facilitate
binding to data obtained through a middle-tier component. A very simple option is
binding to an ArrayList. This option works perfectly in our example, because we need to
populate the DropDownList of cities with strings, and the GetCities method returns an
array list of strings.
The bottom line is that all we need to do to populate the listCities DropDownList is to
add the following code to the Page_Load method of the WebForm1 class.
private void Page_Load(object sender, System.EventArgs e)

{
if (!IsPostBack)
{
hotelBroker = Global.hotelBroker;
ArrayList cities = hotelBroker.GetCities();
listCities.DataSource = cities;
DataBind();
}
}
The call to DataBind( ) binds all the server controls on the form to their data source,
which results in the controls being populated with data from the data source. The
DataBind method can also be invoked on the server controls individually. DataBind is a
method of the Control class, and is inherited by the Page class and by specific server
control classes.
You can now build and run the project. Running a Web application under Visual Studio
will bring up Internet Explorer to access the application over HTTP. Figure 10-20 shows
the running application. When you drop down the list of cities, you will indeed see the
cities returned by the HotelBroker component.
Figure 10-20. Running the Web page to show information about cities.
Initializing the Hotels
We can populate the second DropDownList with hotel data using a similar procedure. It
is a little bit more involved, because GetHotels returns an array list of HotelListItem
structures rather than strings. We want to populate the listHotels DropDownList with the
names of the hotels. The helper method BindHotels loops through the array list of hotels
and creates an array list of hotel names, which is bound to listHotels. Here is the
complete code, which adds the logic for initializing the hotels for the first city (which has
index 0).

{
if (!IsPostBack)
{
hotelBroker = Global.hotelBroker;
ArrayList hotels =
hotelBroker.GetHotels((string)cities[0]);
BindHotels(hotels);
DataBind();
}
}
private void BindHotels(ArrayList hotels)
{
ArrayList hotelNames = new ArrayList(hotels.Count);
foreach(HotelListItem hotel in hotels)
{
hotelNames.Add(hotel.HotelName.Trim());
}
listHotels.DataSource = hotelNames;
Selecting a City
Finally, we implement the feature that selecting a city causes the hotels for the selected
city to be displayed. We can add an event handler for selecting a city by double-clicking
on the listCities DropDownList control. The is a shortcut for adding a handler for the
primary event for the control. In the Properties window you can click on the button to
see all the events for the control. You can then double-click on the event. The second
method allows you to add a handler for any event of the control. Here is the code for the
SelectedIndexChanged event.
private void listCities_SelectedIndexChanged(object

sender,
System.EventArgs e)
{
string city = listCities.SelectedItem.Text;
ArrayList hotels = hotelBroker.GetHotels(city);
BindHotels(hotels);
DataBind();
}
Build and run the project. Unfortunately, the event does not seem to be recognized by the
server. What do you suppose the problem is?
AutoPostBack
For an event to be recognized by the server, you must have a postback to the server. Such
a postback happens automatically for a button click, but not for other events. Once this
problem is recognized, the remedy is simple. In the Properties window for the cities
DropDownList control, change the AutoPostBack property to true. (You can get back to a
display of properties from a display of events by clicking the button.) Figure 10-21
illustrates setting the AutoPostBack property.
Figure 10-21. Setting the AutoPostBack property of a DropDownList control.
Debugging
One advantage of using Visual Studio for developing your ASP.NET applications is the
ease of debugging. You can set breakpoints, single-step, examine the values of variables,
and so forth, in your code-behind files just as you would with any other Visual Studio
program. All you have to do is build your project in Debug mode (the default) and start
the program from within Visual Studio using Debug | Start (or F5 at the keyboard or the
toolbar button ).
As an example, set a breakpoint on the first line of the SelectedIndexChanged event

handler for listCities. Assuming you have set the AutoPostBack property to true, as we
have discussed, you should hit the breakpoint, as illustrated in Figure 10-22.
Figure 10-22. Debugging a code-behind file in Visual Studio.
Deploying a Web Application Created Using Visual Studio
Developing a Web application using Visual Studio is quite straightforward. You can do all
your work within Visual Studio, including testing your application. When you start a Web
application within Visual Studio, Internet Explorer will be brought up automatically. And
it is easy to debug, as we have just seen.
Deploying a Web application created using Visual Studio is also easy, but you need to be
aware of a few things. [4]
[4]
This part of the Visual Studio development environment has been the most
problematical in working with beta software. A technique we have found useful in the
beta is to edit the .csproj.webinfo file to provide an HTTP path to a new location where
the project has been moved. Then double-clicking on the .csproj file will create a new
Visual Studio solution, which you can work with. Be sure to consult the readme.txt file
for this chapter in the code distribution.
1. The Project | Copy Project... menu can be used to deploy a Web project from
Visual Studio.
2. Visual Studio precompiles Web pages, storing the executable in the bin folder.
3. The Src attribute in the Page directive is not used. Instead, the Inherits attribute is
used to specify the Page class.
4. The directory containing the Web pages must be marked as a Web application.
This marking is performed automatically by Visual Studio when you deploy the
application. If you copy the files to another directory, possibly on another system,
you must perform the marking as an application yourself, which you can do using
Internet Services Manager.
Using Project | Copy Project...

To illustrate using Visual Studio to deploy a Web project, let's deploy the Acme Hotel
Information page we have created. We will deploy it to a new directory AcmeWeb0 in the
Deploy directory for Chapter 10.
1. Using Windows Explorer, create a new directory AcmeWeb0 underneath Deploy.
2. Bring up the Copy Project dialog from the menu Project | Copy Project.…
3. Enter the following information (see Figure 10-23).
• http://localhost/NetCs/Deploy/AcmeWeb0 for Destination project folder

• File share for Web access method
• \OI\NetCs\Chap10\Deploy\AcmeWeb0 for Path
• "Only files needed to run this application" for Copy
Figure 10-23. Copying Web project files using Visual Studio.
4. You can test the deployment by using Internet Explorer. Enter the following URL:
http://localhost/netcs/deploy/AcmeWeb0/WebForm1.aspx. You should then
see the hotel information Web page displayed, and you should be able to select a city
from the City dropdown and see the corresponding hotels displayed in the Hotel
dropdown.
Precompiled Web Page
Examining the files in the folder Deploy\AcmeWeb0, you will see no code-behind file
WebForm1.aspx.cs. Instead, in the bin folder you will see the DLL AcmeWeb.dll.
Inherits Attribute in Page Directive
Examining the file WebForm1.aspx, we see there is no Src attribute. Instead, the Inherits
attribute specifies the Page class WebForm1, which is implemented in the assembly
AcmeWeb.dll.
<%@ Page language="c#" Codebehind="WebForm1.aspx.cs"

AutoEventWireup="false" Inherits="AcmeWeb.WebForm1" %>
Configuring a Virtual Directory as an Application
The identical files you copied to Deploy\AcmeWeb0 are also provided in the directory
AcmeRun\Step0. Try the URL http://localhost/
netcs/AcmeRun/Step0/WebForm1.aspx in Internet Explorer. You will obtain a
configuration error, as illustrated in Figure 10-24.
Figure 10-24. Error message when virtual directory is not configured as an

application.
The key sentence in the error message is: "This error can be caused by a virtual directory
not being configured as an application in IIS." The remedy is simple. Use Internet
Services Manager to perform the following steps.
1. Find the folder Step0 underneath AcmeRun in the virtual directory NetCs.
2. Right-click and choose properties. See Figure 10-25. Click "Create."
Figure 10-25. Configuring a virtual directory as an application in IIS.

3. You will then see "Step0" suggested as the application name. Accept all the suggested
settings and click OK.
4. Now again try http://localhost/netcs/AcmeRun/Step0/ WebForm1.aspx in

Internet Explorer. You should be successful in bringing up the application.
Moving a Visual Studio ASP.NET Web Application Project
At the time of writing there appeared to be no really clean way to move an entire
ASP.NET Web Application project so that you could continue development under Visual
Studio. The simplest approach we have found involves copying the source and bin files
and editing the .csproj.webinfo file. A "brute force" approach is outlined in the readme.txt
file for this chapter.
Our illustration will create a copy of the AcmeWeb Web application that we have been
creating in the Demos directory. Our copy will be in a directory called AcmeWeb0 on the
same machine. [5]
[5]
The detailed steps outlined worked on Beta 2. Please consult the readme.txt file for this
chapter to check for any changes in behavior in the released product.
1. In Windows Explorer create a new folder AcmeWeb0 in the Demos directory.
2. Close Visual Studio and copy all the source files, except the .sln and .suo files, from
the AcmeWeb directory to AcmeWeb0. Copy the whole bin folder.
3. Edit the file AcmeWeb.csproj.webinfo to rename Web URLPath to:
"http://localhost/NetCs/Demos/AcmeWeb0/AcmeWeb.csproj"
4. Double-click on the file AcmeWeb.csproj. This should bring up Visual Studio and
create a new solution with a project AcmeWeb.
5. Remove the (broken) reference to Hotel and add this reference back in, navigating to
bin\Hotel.dll.
6. Build the solution. When presented with a Save As dialog, save the solution by the
suggested name AcmeWeb.sln. You should get a clean build.
7. Try to run the project. You will be asked to set a start page. Set the start page as
WebForm1.aspx.
8. Build and run. If you get a configuration error, use Internet Services Manager to
configure the virtual directory as an application in IIS, as previously discussed. You
should now be able to run the application at its new location.
You can view what we have done as establishing a snapshot of Step0. You can go back to
new development in the main directory Demo\AcmeWeb, and if you want to compare
with the original version, you have Demo\AcmeWeb0 available.
Throughout this book we have been using the "Acme Travel Agency" as a case study to
illustrate many concepts of .NET. In this section we look at a Web site for the Acme
Travel Agency. The code for the Web site is in the CaseStudy directory in three
progressive steps: Step0, Step1, and Step2. Step0 corresponds to our Visual Studio.NET
demonstration from the preceding section. (A final Step3, discussed later in the chapter, is
a database version of the case study.)
In this section we will give an overview of the case study, and in the next we will discuss
some more details about Web applications, using the case study as an illustration.
Configuring the Case Study
Links are provided to the three steps of the case study on the ASP.NET example programs
"home page" for this chapter, which you can access through the URL
http://localhost/netcs/. To be able to run the Web applications, you must use IIS to
configure the directories CaseStudy/Step0, CaseStudy/Step1, and CaseStudy/Step2 as
Web applications. Follow the instructions provided in the previous section. If you want to
experiment with any of the steps in Visual Studio, you can double-click on the .csproj file
to create a Visual Studio solution.
Acme Web Site Step 1
In Step 1 we provide a simple two-page Web site. In the first page you can make
reservations, and in the second you can manage your reservations. We have hard-coded
the customer as "Rocket Squirrel," who has a CustomerId of 1.
HotelReservations.Aspx
The start page for the application is HotelReservations.aspx. Figure 10-26 shows this
page in Internet Explorer, after a reservation has been booked at the Hotel Dixie in
Atlanta.
Figure 10-26. Hotel reservations page of Acme Web site.

The code for initializing the DropDownList controls is the same as for Step 0, as is the
code for handling the SelectedIndexChanged event for the City dropdown. The key new
code is making a reservation. This code should have no surprises for you. It makes use of
the HotelBroker class, which we already have instantiated for displaying the hotels.
The design of the Web page enables a user to quickly make a number of reservations
without leaving the page. We are relying on the postback mechanism of ASP.NET. When
done making reservations, the user can follow the link "Manage My Reservations."
ManageReservations.Aspx
The second page for the application is ManageReservations.aspx. Figure 10-27 shows
this page in Internet Explorer, after reservations have been booked for Atlanta, Boston,
and Chicago.
Figure 10-27. Manage reservations page of Acme Web site.

The user can cancel a reservation by selecting a reservation in the listbox and clicking the
"Cancel Selected Reservation" button. A link is provided to the hotel reservations page.
The code for this page is quite straightforward, making use of the capability to provide
event handlers in a server-side control. Here is the code for a helper method to show the
reservations in the listbox. This code is very similar to the Windows Forms code that we
looked at in Chapter 6.
private void ShowReservations()

{
int id = Convert.ToInt32(lblHotelCustomerId.Text);
ArrayList array =
hotelBroker.FindReservationsForCustomer(id);
if (array == null)
{
return;
}
ClearReservations();
foreach (ReservationListItem item in array)
{
string rid = item.ReservationId.ToString();
string hotel = item.HotelName;
string city = item.City;
string arrive = item.ArrivalDate.ToString("d");
string depart = item.DepartureDate.ToString("d");
string number = item.NumberDays.ToString();
string str = id + "," + rid + "," + hotel + "," +
city + " ," + arrive + "," + depart + "," +
number;
listReservations.Items.Add(str);
}
}
Acme Web Site Step 2
Step 2 is the full-blown implementation of our Web site case study. Acme customers do
not interact with the Hotel Broker directly. Instead, they go through Acme's Web site. In
order to use the Web site, a customer must register, providing a user ID, name, and email
address. Subsequently, the user can log in by just providing the user ID.
AcmeLib Component
Internally, Acme maintains a database of user IDs and corresponding Hotel Customer
IDs. [6] The interface IAcmeUser encapsulates this database maintained by Acme. The
class library project AcmeLib contains a collection-based implementation of such a
database. The file AcmeTravelDefs.cs contains the definitions of interfaces and of a
structure.
[6]
The Web site is Acme's, and Acme maintains user IDs for its own customers. Acme
connects to various brokers (such as hotel and airline), and each broker will have its own
customer ID.
// AcmeTravelDefs.cs
using System;
public interface IAcmeUser

{
bool Login(string userid);
bool Register(string userid, string firstName,
string lastName, string emailAddress);
bool Unregister(string userid);
bool ChangeEmailAddress(string userid,
string emailAddress);
bool GetUserInfo(string userid, out UserInfo info);
}
public interface IAcmeAdmin

{
ArrayList GetUsers();
}
public struct UserInfo

{
public int HotelCustomerId;
}
Login will return true if userid is found. Register will register a new user with the Hotel
Broker. Methods are also provided to unregister and change email address. These
methods will call the corresponding methods of the ICustomer interface. GetUserInfo
will return a UserInfo struct as an out parameter. This struct defines an Acme user. The
method GetUsers of the IAcmeAdmin interface returns an array list of UserInfo structs.
The class Acme wraps access to the Customers class, whose methods get invoked
indirectly through methods of IAcmeUser. The class Acme also contains a public
member hotelBroker of type HotelBroker. Thus to gain complete access to the Hotel
Broker system, a client program or Web page simply has to instantiate an instance of
Acme. Here is the start of the definition of Acme.
public class Acme : IAcmeUser, IAcmeAdmin

{
public HotelBroker hotelBroker;
private Customers customers;
private ArrayList users;
private UserInfo currentUser;
public Acme()
{
users = new ArrayList();
customers = new Customers();
InitializeUsers();
}
// Initialize users with data from Customers list
private void InitializeUsers()
{
ArrayList array = customers.GetCustomer(-1);
foreach (CustomerListItem cust in array)
{
string userid = cust.FirstName;
int custid = cust.CustomerId;
User user = new User(userid, custid);
users.Add(user);
}
}
...
The class Acme also implements the interface IAcmeAdmin.
public interface IAcmeAdmin

{
ArrayList GetUsers();
}
The method GetUsers returns an array list of UserInfo.
Login.Aspx
To get a good feel for how this Web application works, it would be a good idea for you to
register and make a few reservations. You could then try logging in as another user. [7]
You can start up the application through the ASP.NET Example programs home page,
link to Acme (Step 2), or else direct enter the URL:
[7]
We are ignoring security considerations in this chapter. Security in ASP.NET will be
discussed in Chapter 12.
http://localhost/netcs/CaseStudy/Step2/Main.aspx
The start page for the application is Main.aspx. If there is no currently logged-in user, the
new user will be redirected to Login.aspx. We will examine the logic in Main.aspx
shortly. For now, let's do the experiment of registering and logging in. Figure 10-28
shows the login page. In our implementation we offer "Rocket" as a possible user ID.
Later you can quickly log in as "Rocket Squirrel" by simply clicking "Login." But now
click "Register."
Figure 10-28. Login page of Acme Web site.

RegisterNewUser.Aspx
The "Register New User" page allows the user to pick a User ID and enter some
identifying information (first name, last name, and email address). Figure 10-29 shows
this page after "John Smith" has entered information for himself. When done entering
information, the user should click "Register," which will directly bring up the Acme
Travel Agency home page, bypassing a need for a separate login.
Figure 10-29. Register new user page of Acme Web site.

Main.Aspx
The home page of the Acme Web Site is Main.aspx. Figure 10-30 shows this home page
for the user "John Smith" who has just registered. A link is provided to "Login" as a
different user, if desired. There are links for "Make a Hotel Reservation" and "Manage
Your Reservations." These pages are the same as shown previously for Step 1.
Figure 10-30. Home page of the Acme Web site.

ASP.NET Applications
An ASP.NET application consists of all the Web pages and code files that can be invoked
from a virtual directory and its subdirectories on a Web server. Besides .aspx files and
code-behind files such as those we have already examined, an application can also have a
global.asax file and a configuration file config.web. In this section we examine the
features of ASP.NET applications. We then investigate the mechanisms for working with
application state and session state and for configuring Web applications. Our illustration
will be our Acme Case Study (Step 2).
Sessions
To appreciate the Web application support provided by ASP.NET, we need to understand

the concept of a Web session. HTTP is a stateless protocol. This means that there is no
direct way for a Web browser to know whether a sequence of requests is from the same
client or from different clients. A Web server such as IIS can provide a mechanism to
classify requests coming from a single client into a logical session. ASP.NET makes it
very easy to work with sessions.
Global.asax
An ASP.NET application can optionally contain a file Global.asax, which contains code
for responding to application-level events raised by ASP.NET. This file resides in the root
directory of the application. Visual Studio will automatically create a Global.asax file for
you when you create an ASP.NET Web Application project. If you do not have a
Global.asax file in your application, ASP.NET will assume you have not defined any
handlers for application-level events.
Global.asax is compiled into a dynamically generated .NET Framework class derived
from HttpApplication.
Here is the Global.asax file for our Case Study Step 2.
using System;
using System.Web;
namespace AcmeWeb
{
{
EventArgs e)
{
HotelState.acme = new Acme();
}
protected void Session_Start(Object sender,
EventArgs e)
{
Session["UserId"] = "";
}
protected void Application_BeginRequest(
Object sender, EventArgs e)
{
}
protected void Application_EndRequest(Object sender,
EventArgs e)
{
}
protected void Session_End(Object sender,
EventArgs e)
{
}
protected void Application_End(Object sender,
EventArgs e)
{
}
}
}
The most common application-level events are shown in this code. The typical life cycle
of a Web application would consist of these events:
• Application_Start is raised only once during an application's lifetime, on the first

instance of HttpApplication. An application starts the first time it is run by IIS for
the first user. In your event handler you can initialize a state that is shared by the
entire application.
• Session_Start is raised at the start of each session. Here you can initialize session
variables.
• Application_BeginRequest is raised at the start of an individual request. Normally
you can do your request processing in the Page class.
• Application_EndRequest is raised at the end of a request.
• Session_End is raised at the end of each session. Normally you do not need to do
cleanup of data initialized in Session_Start, because garbage collection will take
care of normal cleanup for you. However, if you have opened an expensive
resource, such as a database connection, you may wish to call the Dispose method
here.
• Application_End is raised at the very end of an application's lifetime, when the
last instance of HttpApplication is torn down.
In addition to these events, there are other events concerned with security, such as
AuthenticateRequest and AuthorizeRequest. We will discuss ASP.NET security in
Chapter 12.
In the Case Study, we instantiate a single global Acme object instance in

Application_OnStart. This single instance is stored as a static data member of HotelState.
class HotelState
{
static public Acme acme;
}
In the Session_Start event handler we initialize the session variable UserId to be a blank
string. We discuss session variables later in this section.
State in ASP.NET Applications
Preserving state across HTTP requests is a major problem in Web programming, and
ASP.NET provides several facilities that are convenient to use. There are two main types
of state to be preserved.
• Application state is global information that is shared across all users of a Web
application.
• Session state is used to store data for a particular user across multiple requests to a
Web application.
Static Data Members
Static data members of a class are shared across all instances of a class. Hence static data
members can be used to hold application state.
In our case study the class HotelState has a single static member acme of the class Acme.
class HotelState
{
static public Acme acme;
}
Thus the hotelBroker and customers objects within acme will hold shared data that is the
same for all users of the application. Each user will see the same list of hotels.
If you like, you may perform a small experiment at this stage. The directory HotelAdmin
contains a special version of the Acme Web site that makes available the hotel
administration interface IHotelAdmin to the special user with user ID of "admin." When
this privileged user logins, a special home page will be displayed that provides a link to
"Administer Hotels," as illustrated in Figure 10-31.
Figure 10-31. Home page of the Acme Web site tailored for administrators.
Run this Web application, either from the "Hotel Admin" link on the example programs
home page or else via the URL http://localhost/ netcs/HotelAdmin/Main.aspx.
Log in as "admin" and follow the link to "Administer Hotels." You will be brought to a
page showing a list of all the hotels. Select the first hotel (Dixie) on the list and click the
"Delete Selected Hotel" button and then the "Refresh" button. You will now see an
updated list of hotels, as shown in Figure 10-32.
Figure 10-32. Hotel administration page after deleting the Hotel Dixie.
If your Web server is on a network, you can now try running the same Web application
from a different client. Use the URL http://<server-
name>/netcs/HotelAdmin/Main.aspx where "<server-name>" is the name of your
server machine. [8] Again log in as "admin" and go to the "Hotel Admin" page. You should
see the same list of hotels seen by the other client, with Hotel Dixie not on the list. [9]
[8]
On a local machine you can use either the machine name or "localhost."
[9]
Remember that at this point we are not using a database. Thus our example illustrates
application state preserved in memory.
Application Object
You can store global application information in the built-in Application object, an
instance of the class HttpApplicationState. You can conveniently access this object
through the Application property of the Page class. The HttpApplicationState class
provides a key-value dictionary that you can use for storing both objects and scalar
values.
For example, as an alternative to using the class HotelState with the static member acme
that we previously used, we could instead use the Application object. We make up a
string name for the key—for example, "HotelState." In Global.asax we can then
instantiate an Acme object and store it in the Application object using the following code.

EventArgs e)
{
Application["HotelState"] = new Acme();
}
You can then retrieve the Acme object associated with "HotelState" by using the index
expression on the right-hand side and casting to Acme, as illustrated in the code,
Acme acme = (Acme) Application["HotelState"];

string name = acme.CurrentUser.FirstName;
As a little exercise in employing this technique, you may wish to modify the Step 2 case
study to use the Application object in place of a static data member. The solution to this
exercise can be found in the directory ApplicationObject. [10]
[10]
In our current example of a Web application that is precompiled by Visual Studio, it is
quite feasible to use a static variable, that can be shared across pages. But if your
application is not precompiled, each page will be compiled individually at runtime, and
sharing a static variable is no longer feasible. Hence you will have to use the Application
object to share data.
Session Object
You can store session information for individual users in the built-in Session object, an
instance of the class HttpSessionState. You can conveniently access this object through
the Session property of the Page class. The HttpSessionState class provides a key-value
dictionary that you can use for storing both objects and scalar values, in exactly the same
manner employed by HttpApplicationState.
Our case study provides an example of the use of a session variable "UserId" for storing a
string representing the user ID. The session variable is created and initialized in
Global.asax.
protected void Session_Start(Object sender, EventArgs e)

{
Session["UserId"] = "";
}
We use this session variable in the Page_Load event of our home page Main.aspx to
detect whether we have a returning user or a new user. A new user is redirected to the
login page. (Note that "returning" means coming back to the home page during the same
session.)

{
// Put user code to initialize the page here
string userid = (string)Session["UserId"];
if (userid == "")
Response.Redirect("Login.aspx");
if (!IsPostBack)
{
Acme acme = (Acme) Application["HotelState"];
string name = acme.CurrentUser.FirstName;
lblUserName.Text = "Welcome, " + name;
lblLogin.Text = "(If you are not " + name +
", please login)";
}
}
There are some interesting issues in the implementation of session variables.
• Typically cookies are used to identify which requests belong to a particular

session. What if the browser does not support cookies, or the user has disabled
cookies?
• There is overhead in maintaining session state for many users. Will session state
"expire" after a certain time period?
• A common scenario in high-performance Web sites is to use a server farm. How
can your application access its data if a second request for a page is serviced on a
different machine from that on which the first request was serviced?
Session State and Cookies
Although by default ASP.NET uses cookies to identify which requests belong to a

particular session, it is easy to configure ASP.NET to run cookieless. In this mode the
Session ID, normally stored within a cookie, is instead embedded within the URL. We
will discuss cookieless configuration in the next section.
Session State Timeout
By default session state times out after 20 minutes. This means that if a given user is idle
for that period of time, the session is torn down; a request from the client will now be
treated as a request from a new user, and a new session will be created. Again, it is easy
to configure the timeout period, as we will discuss in the section on Configuration.
Session State Store
ASP.NET cleanly solves the Web farm problem, and many other issues, through a session
state model that separates storage from the application's use of the stored information.
Thus different storage scenarios can be implemented without affecting application code.
The .NET state server does not maintain "live" objects across requests. Instead, at the end
of each Web request, all objects in the Session collection are serialized to the session state
store. When the same client returns to the page, the session objects are deserialized.
By default, the session state store is an in-memory cache. It can be configured to be

memory on a specific machine, or to be stored in an SQL Server database. In these cases
the data is not tied to a specific server, and so session data can be safely used with Web
farms.
ASP.NET Configuration
In our discussion of session state we have seen a number of cases where it is desirable to
be able to configure ASP.NET. There are two types of configurations:
• Server configuration specifies default settings that apply to all ASP.NET

applications.
• Application configuration specifies settings specific to a particular ASP.NET
application.
Configuration Files
Configuration is specified in files with an XML format, which is easy to read and to
modify.
Server Configuration File
The configuration file is machine.config. This file is located within a version-specific

folder under \WINNT\Microsoft..NET\Framework. Because there are separate files for
each version of .NET, it is perfectly possible to run different versions of ASP.NET side-
by-side. Thus if you have working Web applications running under one version of .NET,
you can continue to run them, while you develop new applications using a later version.
Application Configuration Files
Optionally, you may provide a file web.config at the root of the virtual directory for a
Web application. If the file is absent, the default configuration settings in machine.config
will be used. If the file is present, any settings in web.config will override the default
settings.
Configuration File Format

Both machine.config and web.config files have the same XML-based format. There are
sections that group related configuration items together, and individual items within the
sections. As an easy way to get a feel both for the format of web.config and also for some
of the important settings you may wish to adjust, just look at the web.config file that is
created by Visual Studio when you create a new ASP.NET Web Application project.
<?xml version="1.0" encoding="utf-8" ?>

<configuration>
<system.web>

<compilation
defaultLanguage="c#"
debug="true"
/>

<customErrors
mode="Off"
/>

<authentication mode="None" />
...
</system.web>
</configuration>
Application Tracing
Earlier in the chapter we examined page-level tracing, which can be enabled with the
Trace="true" attribute in the Page directive. Page-level tracing is useful during
development but is rather intrusive, because the page trace is sent back to the browser
along with the regular response. Application tracing, which is specified in web.config,
writes the trace information to a log file, which can be viewed via a special URL.
As a demonstration of the use of web.config, let's add application tracing to our original
Hello.aspx application. The folder HelloConfig contains Hello.aspx and web.config. We
have added a trace statement in Hello.aspx.


<%@ Page Language="C#" %>
<HTML>
<HEAD>
{
Trace.Write("cmdEcho_Click called");
}
</SCRIPT>
</HEAD>
<BODY>
<p><asp:button id=cmdEcho onclick=cmdEcho_Click
Text="Echo" runat="server" tooltip="Click to echo your
name">
</asp:button></p>
<P></P>
</FORM>
</BODY>
</HTML>
We have provided a trace section in web.config to enable tracing.

<configuration>
<system.web>
<trace
enabled="true"
/>
</system.web>
</configuration>
You can run this application from Internet Explorer by simply providing the URL
http://localhost/netcs/helloconfig/hello.aspx. [11] Enter a name and click the
"Echo" button a couple of times. The application should run normally, without any trace
information included in the normal page returned to the browser.
[11]
If you get a configuration error, try configuring the directory in IIS as an application.
See "Configuring a Virtual Directory as an Application" in the section "Deploying a Web
Application Created Using Visual Studio."
Now enter the following URL: http://localhost/netcs/helloconfig/ trace.axd
(specifying trace.axd in place of hello.aspx), and you will see top-level trace information,
with a line for each trip to the server, as shown in Figure 10-33. If you click on the "View
Details" link, you will see a detailed page trace, as we saw earlier in the chapter.
Figure 10-33. Viewing the application trace log through the browser.
Session Configuration
As another example of configuration, modify the web.config file for Step 2 of the case
study to change the timeout value to be 1 minute.

<configuration>
<system.web>
...

<sessionState
mode="InProc"
stateConnectionString="tcpip=127.0.0.1:42424"
sqlConnectionString=
"data source=127.0.0.1;user id=sa;password="
cookieless="false"
timeout="1"
/>
...
</system.web>
</configuration>
Now run the application, log in, do some work, and return to the home page. You should
be welcomed by your name without having to log in again. Now do some more work,
wait more than a minute, and return to the home page. Now the session will have timed
out, and you will be redirected to log in again.
Server Controls
An important innovation in ASP.NET is server controls. They provide an event model

that is startlingly similar to Windows GUI programming, and they encapsulate browser
dependencies. They integrate seamlessly into the Visual Studio development
environment. The end result is an extremely powerful tool for Web development.
We have been using server controls from the very beginning of the chapter, where we
presented our "Hello" program. In this section we will look at server controls more
systematically, and we will see a number of examples of interesting controls.
Web Controls
The most important kind of control in ASP.NET is the Web Forms server control or just
Web control. These are new controls provided by the .NET Framework, with special tags
such as <asp:textbox>. These controls run at the server, and they generate HTML code
that is sent back to the browser. They are easy to work with, because they behave
consistently. For example, you can determine the value returned by a control by using
simple property notation.
string name = txtName.Text;
All of our previous examples of server controls in this chapter have been Web controls. In
this section, we will look at several additional kinds of Web controls, including validation
controls, list controls, and rich controls such as the Calendar control. But first we will
look at HTML server controls.
HTML Server Controls
HTML server controls provide equivalent functionality to standard HTML controls,

except that they run on the server, not on the client. In fact, the only way to distinguish an
HTML server control from an ordinary HTML control on a Web page is the presence of
the runat="server" attribute.
Here are two controls. Both are INPUT controls. The first is a server control. The second
is of type password and is a regular HTML control.
<INPUT id=txtUserId
style="WIDTH: 135px; HEIGHT: 22px" type=text size=17
runat="server"></P>
<INPUT id=""
style="WIDTH: 138px; HEIGHT: 22px" type=password size=17
name=txtPassword>
Working with HTML server controls is much like working with the Web Forms server
controls we've used already. In server-side code you access the control through a control
variable that has the same name as the id attribute. However, we are dealing with HTML
controls, so there are some differences. You access the string value of the control not
through the Text property but through the Value property. Here is some code that uses the
value entered by the user for the txtUserId control.
lblMessage.Text = "Welcome," + txtUserId.Value;
The advantage of HTML server controls for the experienced Web programmer is that they
match ordinary HTML controls exactly, so that your knowledge of the details of HTML
control properties and behavior carries over to the ASP.NET world. However, this
similarity means they carry over all the quirks and inconsistencies of HTML. For
example, rather than having two different controls for the somewhat different behaviors
of a textbox and a password control, HTML uses in both cases the INPUT control,
distinguishing between the two by the type=password attribute. Web Forms controls, in
contrast, are a fresh design and have an internal consistency. Also, as we shall soon see,
there is a much greater variety to Web Forms controls.
Html Controls Example
Let's look at an example of HTML controls. All of our server control examples in this
section can be accessed from the page ServerControls\WebForms1.aspx. (As usual, you
should use IIS to configure the folder ServerControls as an application.) The top-level
page gives you a choice of three examples,
• HTML Controls
• Validation
• Calendar
Follow the link to HTML Controls, and you will come to a login page, as illustrated in
Figure 10-34.
Figure 10-34. A login page illustrating HTML server controls.
There is a textbox for entering a user ID and a password control for entering a password.
Both of these controls are HTML INPUT controls, as shown previously. The textbox runs
at the server, and the password is an ordinary HTML control. Clicking the Login button
(implemented as a Windows Forms Button control) results in very simple action. There is
one legal password, hardcoded at "77." The button event handler checks for this
password. If legal, it displays a welcome message, otherwise an error message.
private void Login_Click(object sender, EventArgs e)

{
if (Request.Params["txtPassword"] == "77")
lblMessage.Text = "Welcome, " + txtUserId.Value;
else
lblMessage.Text = "Illegal password";
}
Since the password control is not a server control, no server control variable is available
for accessing the value. Instead, we must rely on a more fundamental technique, such as
using the Params collection. [12]
[12]
We described the various collections earlier in the chapter in the section
"Request/Response Programming." The collections are included in Table 10-1.
HTML Controls in Visual Studio
It is easy to work with HTML controls in Visual Studio. [13] The Toolbox has a palette of
HTML controls, which you can access through the HTML tab. Figure 10-35 shows some
of the HTML controls in the Visual Studio Toolbox.
[13]
But it is also confusing, because there is only one palette for HTML controls, and you
distinguish between classical HTML controls and server HTML controls by
runat="server." The Forms Designer UI for setting this attribute is described below.
Figure 10-35. HTML controls in the Visual Studio Toolbox.
You can drag HTML controls onto a form, just as we have done with Web Forms
controls. You have the option of using FlowLayout or GridLayout. The default is
GridLayout, which enables absolute positioning of controls on a form. FlowLayout is the
simplest layout, resulting in elements positioned in a linear fashion. You can set the
layout mode through the pageLayout property of the form. In our example we used
FlowLayout for the two INPUT controls and their associated labels.
The default choice for HTML controls is not to run at the server. To make an HTML
control into a server control, right-click on it in the Form Designer. Clicking on "Run As
Server Control" toggles back and forth between running on the server and not running on
the server. You can inspect the runat property in the Properties panel, but you cannot
change it there.
Validation Controls
The rest of our discussion of server controls will focus on Web controls. A very
convenient category of control is the group of validation controls. The basic idea of a
validation control is very simple. You associate a validation control with a server control
whose input you want to validate. Various kinds of validations can be performed by
different kinds of validation controls. The validation control can display an error message
if the validation is not passed. Alternatively, you can check the IsValid property of the
validation control. If one of the standard validation controls does not do the job for you,
you can implement a custom validation control. The following validation controls are
available:
• RequiredFieldValidator
• RangeValidator
• CompareValidator
• RegularExpressionValidator
• CustomValidator
There is also a ValidationSummaryControl that can give a summary of all the validation
results in one place.
An interesting feature of validation controls is that they can run on either the client or the
server, depending on the capabilities of the browser. With an upscale browser such as
Internet Explorer, ASP.NET will emit HTML code containing JavaScript to do validation
on the client. [14] If the browser does not support client-side validation, the validation will
be done only on the server.
[14]
Validation will also be done on the server, to prevent "spoofing."
Required Field Validation
A very simple and useful kind of validation is to check that the user has entered
information in required fields. Our second server control demonstration page provides an
illustration. Back on the top-level ServerControls\WebForms1.aspx page, follow the link
to "Validation" (or click the Register button from the Login page). You will be brought to
the page RegisterNewUser.aspx, as illustrated in Figure 10-36. The screenshot shows the
result of clicking the Register button after entering a UserId, a Password, and a First
Name, but leaving Last Name blank. You will see an error message displayed next to the
Last Name textbox, because that is where the validator control is on the form.
Figure 10-36. Register New User page illustrates ASP.NET validation controls.
The textboxes for First Name and Last Name both have an associated
RequiredFieldValidator control. In Visual Studio you can simply drag the control to a
position next to the associated control. You have to set two properties of the validator
control:
• ControlToValidate must be set to the ID of the control that is to be validated.

• ErrorMessage must be specified.
Then, when you try to submit the form, the validator control will check whether
information has been entered in its associated control. If there is no data in the control,
the designated error message will be displayed.
Internet Explorer supports client-side validation using JavaScript. You can verify that
ASP.NET generates suitable JavaScript by looking at the generated source code in the
browser (View | Source).
This form also requires that the UserId field not be blank. Since the primary validation of
this field is done by a regular expression validator, as discussed shortly, we will use
another technique for the required field validation. Figure 10-37 shows the location of the
various validator controls in the Visual Studio Form Designer.
Figure 10-37. Layout of validation controls for Register New User page.
We assign the id vldUserId to the required field validator control associated with the
UserId control, and we clear the error message. We also set the EnableClientScript
property to False, to force a postback to the server for the validation. The event handler
for the Register button then checks the IsValid property of vldUserId.
private void cmdRegister_Click(object sender,

System.EventArgs e)
{
if (vldUserId.IsValid)
lblMessage.Text = "Welcome, " + txtFirstName.Text;
else
lblMessage.Text = "UserId must not be blank";
}
If the control is valid, we display the welcome message, otherwise an error message.
Note that we won't even reach this handler if other validation is false.
Regular Expression Validation

The RegularExpressionValidator control provides a very flexible mechanism for
validating string input. It checks whether the string is a legal match against a designated
regular expression. Our example illustrates performing a regular expression validation of
UserId. The requirement is that the id consist only of letters and digits, which can be
specified by the regular expression [A-Za-z0-9]+
The following properties should normally be assigned for a RegularExpressionValidator

control:
• ValidationExpression (the regular expression, not surrounded by quotes)

• ControlToValidate
• ErrorMessage
You can try this validation out on our Register New User page by entering a string for
UserId that contains a nonalphanumeric character.
Rich Controls
Another category of Web Forms controls consists of "rich controls," which can have quite
elaborate functionality. The Calendar control provides an easy-to-use mechanism for
entering dates on a Web page. Our third sample server control page provides an
illustration, as shown in Figure 10-38.
Figure 10-38. Using the Calendar control to select a date.

The user can select a date on the Calendar control. The SelectedDate property then
contains the selected date as an instance of the DateTime structure. You can work with
this date by handling the SelectionChanged event. In our example page, the event handler
displays the date as a string in a textbox.
private void Calendar1_SelectionChanged(object sender,

System.EventArgs e)
{
txtDate.Text =
Calendar1.SelectedDate.ToShortDateString();
}
Database Access in ASP.NET
A great deal of practical Web application development involves accessing data in various
kinds of databases. A great thing about the .NET Framework is that it is very easy to
encapsulate a database, allowing the rest of the program to work with data in a very
generic way, without worrying about where it came from. In this section we discuss data
binding in Web Forms controls, and we then present a database version of our Acme
Travel Agency Web site.
Data Binding in ASP.NET
ASP.NET makes it easy to display data from various data sources by permitting a Web
Forms control to be bound to data source. The data source can be specified in a variety of
ways—for example, by directly giving a connection string to a database. This form of
data binding is quite convenient in a two-tier type of application, where the presentation
layer talks directly to the database. In three-tier applications it is more convenient to bind
to some data structure that is returned by a middle-tier component, which does the actual
connection to the database. Our Acme case study illustrates this approach. The Hotel.dll
and Customer.dll components encapsulate access to a SQL Server database through the
HotelBroker and Customers classes. Methods such as GetCities return an ArrayList, and
the array list can be bound to a Web Forms control. [15]
[15]
The component could be hidden behind a Web Service, which will be illustrated in
Chapter 11. We can still use data binding in such a scenario, by binding to an array list.
We will look at two examples of data binding. The first, mentioned earlier in the chapter,
illustrates binding to an ArrayList. The second illustrates binding to a DataTable through
a DataView.
Binding to an Arraylist
It is extremely simple to bind to an array list. The case study code, beginning with Step 1,
provides an illustration. You may wish to bring up Step 1 of the case study and examine
the code in CaseStudy\Step1\ MakeReservations.aspx.cs. When the page is loaded, the
DropDownList control listCities is initialized to display all the cities in the database of
the hotel broker. The GetCities method returns the cities as strings in an array list. The
following code will then cause the cities to be displayed in the dropdown.

DataBind();
The DataBind method of the Page class causes all the Web Forms controls on the page to
be bound to their data sources, which will cause the controls to be populated with data
from the data sources. You could also call the DataBind method of a particular control.
Binding to a Datatable
As we saw in Chapter 9, ADO.NET defines a very useful class, the DataTable, which can
be used to hold data from a variety of data sources. Once created, a data table can be
passed around and used in a variety of contexts. One very useful thing you can do with a
data table is to bind it to a Web Forms control. Since a data table is self-describing, the
control can automatically display additional information, such as the names of the
columns. We illustrate with the DataGrid control.
To run this example, you need to have SQL Server or MSDE installed on your system,
and you should also have set up the Acme database, as described in Chapter 9. The
example Web page is DataGridControl/ShowHotels.aspx. As usual, you should use IIS to
configure the folder DataGridControl as an application. This page will display all the
hotels in the Acme database in a data grid, with appropriate headings, as illustrated in
Figure 10-39. When you work with Web Forms controls you can easily change styles,
such as fonts and colors, by setting properties appropriately.
Figure 10-39. Displaying hotels in the Acme database using a DataGrid control.
The relevant C# code is in the files Global.asax.cs and ShowHotels. aspx.cs. The first
thing we need to do is to create an instance of the HotelBroker class. We create a single
instance, once, when the application starts up.
// Global.asax.cs
using System;
using System.Web;
namespace DataGridControl
{
{
public static HotelBroker hotelBroker;
EventArgs e)
{
}
...
In the Page_Load method we get the hotels from the Hotel Broker, call a helper method,
CreateDataSource, to obtain an ICollection interface reference (the data binding is very
general, and any collection can be used), assign the data source, and bind. We are using
the DataTable to hold data obtained from the middle-tier component.

{
if (!IsPostBack)
{
// Need to load this data only once.
ArrayList array = Global.hotelBroker.GetHotels();
dgHotels.DataSource= CreateDataSource(array);
dgHotels.DataBind();
}
}
It is in the helper method CreateDataSource that the interesting work is done. A data table
is created and populated with hotel data obtained from the Hotel Broker.
private ICollection CreateDataSource(ArrayList array)

{
if (array == null)
{
return null;
}
DataTable dt = new DataTable();
DataRow dr;
dt.Columns.Add(new DataColumn("City", typeof(string)));
dt.Columns.Add(new DataColumn("Hotel", typeof(string)));
dt.Columns.Add(new DataColumn("Rooms", typeof(int)));
dt.Columns.Add(new DataColumn("Rate", typeof(decimal)));
foreach(HotelListItem hotel in array)
{
dr = dt.NewRow();
dr[0] = hotel.City.Trim();
dr[1] = hotel.HotelName.Trim();
dr[2] = hotel.NumberRooms;
dr[3] = hotel.Rate;
dt.Rows.Add(dr);
}
Acme Travel Agency Case Study (Database Version)
We have illustrated many concepts of ASP.NET with our Acme Travel Agency case study.
For simplicity we used a version of the case study that stored all data as collections in
memory. This way you did not have to worry about having a database set up properly on
your system, so you could focus on just ASP.NET. Also, the results are always
deterministic, since sample data is hardcoded.
Now, however, we would like to look at the "real" case study, based upon our
HotelBroker database, and the database version of the Hotel.dll and Customer.dll
components created in Chapter 9.
AcmeCustomerDatabase
The Acme Travel Agency maintains its own database of customers. Customers register
with Acme through the Web site. The following information is stored in Acme's database:
• LoginName
• Password
• HotelBrokerCustomerId
• AirlineBrokerCustomerId
Currently we use LoginName (corresponding to what we called "UserId" earlier in the

chapter) and HotelBrokerCustomerId. The AirlineBrokerCustomerId field will facilitate
Acme adding an airplane reservation system later. A Password field is also provided for
possible future use.
To set up the database, all you need to do is to run the script acmedb.sql, which is located
in the directory AcmeScript. This script assumes you have SQL Server installed on
partition c:. If your installation is in a different partition, edit the script accordingly.
AcmeLibDb Component
The directory AcmeLibDb contains a class library project for building an AcmeLib
component that encapsulates access to the AcmeCustomerDatabase. This component also
wraps access to HotelBroker and Customers, providing the Web pages with a very easy
programming model.
Acme Web Site (Step 3)
The Step 3 version of the Acme Web site is in CaseStudy\Step3. As usual, you will need
to use IIS to configure this directory as an application. You can start it from the URL
http://localhost/netcs/CaseStudy/Step3/Login.aspx
You should find the code very easy to understand, because it relies on the same interfaces
as the implementation we used earlier based on collections.
Summary
ASP.NET is a unified Web development platform that greatly simplifies the

implementation of sophisticated Web applications. In this chapter we introduced the
fundamentals of ASP.NET and Web Forms, which make it easy to develop interactive
Web sites. Server controls present the programmer with an event model similar to what is
provided by controls in ordinary Windows programming. This high-level programming
model rests on a lower-level request/response programming model that is common to
earlier approaches to Web programming and is still accessible to the ASP.NET
programmer.
The Visual Studio.NET development environment includes a Form Designer, which

makes it very easy to visually lay out Web forms, and with a click you can add event
handlers. ASP.NET makes it very easy to handle state management. Configuration is
based on XML files and is very flexible. There are a great variety of server controls,
including wrappers around HTML controls, validation controls, and rich controls such as
a Calendar. Data binding makes it easy to display data from a variety of data sources.
In the next chapter we cover Web Services, which enable the development of
collaborative Web applications that span heterogeneous systems.
Chapter 11. Web Services Top
Distributing functionality and data beyond the enterprise in which they were developed is
the next step in component technology. Developers can integrate into their applications a
much more extensive set of services than they could ever hope to develop on their own.
Our Acme Reservation System case study is a simple example. The Acme Travel Agency,
by using the reservation systems of the airlines and hotels, can provide a wider range of
services to their clients.
One vendor will not be able to supply the necessary distributed technology infrastructure.
At the very minimum, the worlds of Java, .NET, mobile computers, and legacy systems
will continue. Fortunately, TCP/IP and HTTP have established themselves as industry
standard networking protocols and can be the basis for any attempt to interconnect
heterogeneous systems. HTTP is a text-based protocol, so using the industry standard
XML to describe the interactions of these systems makes sense. Web Services use XML-
and HTTP-based protocols to provide an industry standard to allow diverse systems to
interconnect.
Web Services is the second part of the .NET distributed computing story. If all the
applications and services that need to interconnect are all based on the Common
Language Runtime, .NET remoting can be used. Its advantage is that you can remote any
.NET data structure through the remoting serialization. Environments that do not run
.NET, however, cannot handle the full range of .NET data types. Hence, Web Services
transmit only a much more limited set of data structures that can be expressed in the
XML-based protocols that Web Services use today. The versions of the SOAP protocol
used by Web Services and by .NET remoting have different programming models. The
latter offers full CLR fidelity. The former is constrained by interoperability standards.
Besides the ability of heterogeneous systems to interconnect, Web Services allow

business partners to share information or integrate with legacy systems without having to
write specialized interconnection applications. Even within a single enterprise you will be
able to integrate information from internal and external sources. If Web Services are to be
more than just distributed application development, however, the necessary financial,
reliability, security, and legal infrastructure have to be developed.
Protocols
Behind the Web Services technology are several protocols: XML, XML Namespaces,
XML Schema, SOAP, and WSDL. Some of these are formal W3C industry standards.
Some, like WSDL, are just gaining widespread use without yet being codified in a
standard.
XML
XML is a W3C industry standard [1] that provides a way to structure documents to provide
relationships between the basic elements of the document. Elements can also have
descriptive information called attributes. Elements can be composed of other elements, so
they can have complex structure. Since such documents can be represented as text, [2]
XML can provide a platform-neutral way to represent data that are transmitted over a
network. In particular, as text it can go safely through a firewall because HTTP port 80
will invariably be open. Here is an example of an XML document that describes a
CustomerList composed of several customers.
[1]
Technically, W3C final documents are called recommendations. However, we will refer
to them as standards or specifications. W3C documents that have not reached
recommendations status are referred to by their W3C names: proposed recommendations,
candidate recommendations, last call working drafts, working drafts, and notes.
[2]
But they do not have to be text. You can build programs using the abstractions defined
in the W3C proposed recommendation Information Set. Using these abstractions, such as
document, namespace, element, character, and attribute, to represent the hierarchy of an
XML document, you are independent of the particular format in which the XML is
stored. Mobile solutions will probably use a more efficient binary format for XML
encoding rather than text. The XML Schema Recommendation is written based on the
Infoset, not the angle-bracket syntax. The Information Set assumes the existence of XML
namespaces.
<CustomerList>
<Customer>
<FirstName>John</FirstName>
<LastName>Smith</LastName>
</Customer>
<Customer>
<FirstName>Mary</FirstName>
<LastName>Jones</LastName>
</Customer>
</CustomerList>
XML Namespaces
A set of elements and attributes in an XML document can be referred to as a vocabulary.

This is particularly useful if this vocabulary can model information that might be reused.
For example, we could have vocabularies for financial or chemical information.
Namespaces not only allow these vocabularies to be uniquely named in order to prevent
conflicts, but allow them to be reused.
The following example XML document uses a namespace attribute to uniquely identify
the elements <FirstName>, <LastName>, and <EmailAddress> from any other
definitions that might use the same tag names with a different meaning or context. The
example also shows that abbreviations can be used with namespaces. This is very
convenient if multiple namespaces are used in a document.
<Customer xmlns:c=
"urn:uuid 28833F1C-CBE4-4042-9B35-BF641DFB35DC">
<c:FirstName>John</c:FirstName>
<c:LastName>Smith</c:LastName>
<c:EmailAddress>[email protected]</c:EmailAddress>
</Customer>
A Uniform Resource Identifier (URI) is used to identify a particular XML namespace. A

URI can either be a Uniform Resource Locator (URL) or a Uniform Resource Name
(URN). Both represent a unique name. URLs are the familiar Web site addresses, which
are unique because they are given out by a central naming authority. A URN is just a
unique string. For example, you could use a URN defined by a GUID [3] such as
urn:uuid:28833F1C-CBE4-4042-9B35-BF641DFB35DC. [4] URIs used for namespaces
do not have to resolve to any location on the Web.
[3]
A GUID, or Globally Unique Identifier, is a 128-bit identifier that is guaranteed to be
unique. GUIDs are widely used in COM. You can generate your own GUIDs using the
tool guidgen.exe (Windows UI) or uuidgen.exe (command-line UI). These tools are in the
directory …\Microsoft Visual Studio.NET\Common7\Tools.
[4]
GUIDs are used in the examples for simplicity and to reinforce the idea that
uniqueness, but not existence, is required for a namespace identifier. In real systems
URL-based names are used whether or not the URLs actually exist.
XML Schema
XML with namespaces, however, does not assign any semantics to the data. The XML
Schema specification (XSD) defines a basic set of data types and the means to define new
data types. In other words, an XML Schema can assign meaning to the structure of a
document. The schema itself is written in XML. The CustomerList document described
previously could be defined by the following schema:
<schema xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:c="http://www.acme.com/Customer"
targetNamespace="http://www.acme.com/CustomerList">
<xsd:complexType name="Customer">
<xsd:sequence>
<xsd:element name="FirstName" type="xsd:string" />
<xsd:element name="LastName" type="xsd:string" />
<xsd:element name="EmailAddress" type="xsd:string"/>
</xsd:sequence>
</xsd:complexType>
</schema>
The targetNamespace element defines the name of the schema being defined. This
particular string uses the XSD defined element "string." Using XSD, we can restrict the
range of values, specify how often particular instances occur, as well as provide attributes
to the elements. The schema itself is written in XML. Both the document and its
associated schema can be validated and managed as XML documents. The same
document, interpreted by two different schemas, will have two different meanings.
SOAP
While XML schemas can define the types used by the data, you need a set of conventions
to describe how the data and their associated type definitions are transmitted. SOAP, the
Simple Object Access Protocol, uses XML as a wire protocol to do just this.
While SOAP can use XML schema types to describe the transmitted types, it was
designed before the XML Schema specification was finished, so there are some
divergences between the two. The reason is that XML Schema describes a hierarchy or
tree structure. SOAP wants to be able to represent objects, and objects can have far more
complicated relationships than a hierarchy. Classes, for example, can have multiple
parent classes. As we will discuss later, this has some implications for Web Services. The
W3C is currently working on reconciling SOAP with XML Schema.
SOAP 1.1 can be used with several transport protocols, not just HTTP.
The use of SOAP for Web Services on Microsoft platforms is not unique to .NET.
Microsoft has released the SOAP Toolkit that has allowed Windows-based platforms to
develop Web Services. The support for SOAP, however, is built into .NET. The SOAP
Toolkit does contain, however, the SOAP Trace Utility, which is useful for tracking raw
and formatted SOAP messages.
WSDL
Objects contain both state and behavior. Schemas define the data. WSDL, the Web
Services Description Language, defines the methods and the data associated with a Web
Service. As the simple example we shall describe shortly demonstrates, WSDL is not
necessary for writing Web Services. It is important, however, if you want to be able to
automatically generate classes that can call Web Services, or do anything that requires
automatic machine intervention with Web Services. [5] Otherwise, you would have to craft
and send the SOAP messages by hand.
[5]
This is similar to VB 6's use of type libraries to make COM programming simpler. Of
course, WSDL is a complete description of the Web Service, unlike a type library's
incomplete description of a COM object and interfaces.
As you will see in the following example, the SOAP that is used to describe the Web
Service's transport format is defined in the WSDL. WSDL is a W3C note.
Web Service Architecture
Besides handling ASP.NET, Microsoft's Internet Information Server (IIS) can handle Web
Services, since they come in as HTTP requests. These requests are encoded in the URL or
as XML. IIS then creates the required object to fulfill the Web Service request. IIS then
calls the object's method that is associated with the request. Any returned values are
converted to XML and returned to the client, using the HTTP protocol.
The Add Web Service Example
To illustrate how this works under Microsoft .NET, we will build a simple Web Service to
illustrate this architecture and how the associated protocols are used. Our Web Service
will simply add two numbers. To make things clear we will build the Web Service, Add,
in the simplest possible way.
By writing code in a file with the suffix asmx and placing it in a subdirectory of the IIS
root directory we can have a simple Web Service. [6] IIS has the concept of virtual
directories, so that the actual directory does not have to physically be under the IIS root
directory. The easiest way to do this is to enable WebSharing on the file folder. Select the
folder in the NT Explorer, right-click on the folder, and select Sharing on the context
menu. Use the Web Sharing tab to make the directory a virtual directory for IIS.
[6]
By default this directory is \inetpub\wwwroot.
The file add.asmx first defines the language used to write the Web Service, and the class
that has the definitions. That class inherits from the WebService class in the namespace
System.Web.Services. Note the use of the WebService attribute to define a namespace for
the service. This file is found in the WebService subdirectory of the SimpleWebService
directory for this chapter. You should make WebService a virtual directory with alias
SimpleWebService, as described in the previous paragraph.
A method of that class can be used as a Web Service if the attribute WebMethod is
applied to it.
<%@ WebService language="C#" class="Test" %>

using System;
using System.Web;
using System.Web.Services;
[WebService(Namespace=
"urn:uuid:10C14FCF-BF4A-477a-BFE7-41B9F2A4514E")]
public class Test: WebService
{
[WebMethod]
public long Add(long x, long y)
{
return x + y;
}
}
We will put this file in a directory called SimpleWebService. [7]

[7]
You can use a code-behind page here if you wish. WebServices created with VS.NET
do reference a code-behind page in the asmx file.
A Client Program for the Add Web Service
Internet Explorer can be used as a simple client program that uses the HTTP GET
protocol's URL encoding of a Web Service request. Using http://
localhost/SimpleWebService/Add.asmx as the address, Figure 11-1 shows the result.
Figure 11-1. Web Service request in Internet Explorer.
By clicking on the Add link you will get a form enabling you to submit a request to the
Add Service. In addition, the form describes the various HTTP protocols that can be used
for submitting the request. For our purposes, two protocols are worth mentioning: HTTP
GET and SOAP.
The HTTP GET protocol is worth exploring because the form that appears in IE uses it.
The protocol has boldfaced placeholders for data that has to be entered:
GET /SimpleWebService/add.asmx/Add?x=string&y=string
HTTP/1.1
...
The data entered into the form is added to the URL in the standard way that any HTTP
GET request is made. Data are returned as:
...
<long xmlns="urn:uuid:10C14FCF-BF4A-477a-BFE7-
41B9F2A4514E">long</long>
Figure 11-2 shows values entered into the form. By pressing the Invoke button, you can
call the Web Service.
Figure 11-2. Values entered on the Internet Explorer form.
An IE window will appear with the part of the HTTP response data generated by the Web
Service that contains the actual returned value:
...
<long xmlns=
"urn:uuid:10C14FCF-BF4A-477a-BFE7-41B9F2A4514E">9</long>
This is exactly the format that appeared in the description of the protocol with the answer
(9) substituted for the placeholder. HTTP GET, however, can handle only simple types.
The more interesting protocol is SOAP. Both the SOAP HTTP POST request and
response are described with placeholders for information that has to be provided in the
actual call. Those placeholders are in boldface type.
First, let us look at the SOAP HTTP POST request. The first part is a set of HTTP
headers. The XML for the SOAP protocol is in the data (entity-body) section of the HTTP
request, which is always separated from the headers by a blank line. The content-length
header is the length of the data, which is dependent on the size of the parameters in the
data section.
The method header identifies the file to which the request is directed. It could also name
an object that is to handle the request (endpoint). The SOAPAction header indicates the
name of the method, qualified by a namespace, to be invoked for the Web Service. [8]
[8]
For those with a COM background, you can think of the namespace for the method as
equivalent to the GUID that identifies and interface (IID).
SOAP uses XML to specify the parameters of the method. [9] The SOAP body contains
the parameters for the method call. In a real method call, the long placeholders would be
replaced by the actual parameters to be passed to the Web Service method.
[9]
The parallel to IDL is WSDL, which we will discuss shortly. SOAP is analogous to
NDR, the wire format used for DCOM calls. All these parallels to COM appear in Don
Box's March 2000 MSDN article "A Young Person's Guide to The Simple Object Access
Protocol."
POST /SimpleWebService/Add.asmx HTTP/1.1

Host: localhost
Content-Type: text/xml; charset=utf-8
Content-Length: length
SOAPAction:
"urn:uuid:10C14FCF-BF4A-477a-BFE7-41B9F2A4514E/Add"
<?xml version="1.0" encoding="utf-8"?>

<soap:Envelope
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/">
<soap:Body>
<Add xmlns=
"urn:uuid:10C14FCF-BF4A-477a-BFE7-41B9F2A4514E">
<x>long</x>
<y>long</y>
</Add>
</soap:Body>
</soap:Envelope>
Next the HTTP response is described. The long placeholder will be replaced by the actual
value returned.
HTTP/1.1 200 OK
Content-Length: length

<soap:Envelope xmlns:xsi="http://www.w3.org/2001/XMLSchema-
instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"
<soap:Body>
<AddResponse xmlns=
<AddResult>long</AddResult>
</AddResponse>
</soap:Body>
</soap:Envelope>
WSDL
SOAP does not describe the Web Service interface. While you could encode the SOAP
yourself, it would be nice to be able to generate proxy classes for the client to use.
Otherwise you would have to understand all the details of the SOAP specification and
how to parse the returned XML.
WSDL provides a description of the Web Service interface. Here is the WSDL
description for our SimpleWebService which has one method, Add. We have omitted the
WSDL for invocations of the Web Service that do not use SOAP. The <types> section
defines the types:
• Add is used in when SOAP invokes the Web Service.

• AddResponse is used when the SOAP Web Service invocation returns.
Add has two elements, each occurring exactly once. Both are defined with the XSD type
long, and they have the names x and y. The return parameter, whose name is
AddResponse, has one element which occurs once named AddResult defined with the
XSD type long. Note how these types were used in the SOAP definitions we looked at
previously.
...
<types>
...
<s:element name="Add">
<s:complexType>
<s:sequence>
<s:element minOccurs="1" maxOccurs="1" name="x"
type="s:long" />
<s:element minOccurs="1" maxOccurs="1" name="y"
type="s:long" />
</s:sequence>
</s:complexType>
</s:element>
<s:element name="AddResponse">
<s:complexType>
<s:sequence>
<s:element minOccurs="1" maxOccurs="1"
name="AddResult" type="s:long" />
</s:sequence>
</s:complexType>
</s:element>
...
</types>
The <message> section relates the types to their use as parameters.
<message name="AddSoapIn">
<part name="parameters" element="s0:Add" />
</message>
<message name="AddSoapOut">
<part name="parameters" element="s0:AddResponse" />
</message>
...
The <portType> section relates the Web Service to the individual Web methods defined
by the <operation> elements. If there had been more Web methods in the Web Service,
there would have been more operation elements associated with the portType. [10] Each
method's input and output operation is associated with the appropriate message defined
previously.
[10]
For those of you keeping score, this is analogous to a COM interface.
<portType name="TestSoap">
<operation name="Add">
<input message="s0:AddSoapIn" />
<output message="s0:AddSoapOut" />
</operation>
</portType>
...
The <binding> section defines the encodings and protocols to be used for each operation.
<binding name="TestSoap" type="s0:TestSoap">

<soap:binding
transport="http://schemas.xmlsoap.org/soap/http"
style="document" />
<operation name="Add">
<soap:operation soapAction=
style="document" />
<input>
<soap:body use="literal" />
</input>
<output>
<soap:body use="literal" />
</output>
</operation>
</binding>
...
The <service> section relates the Web Service to its port and how it is invoked.
<service name="Test">
<port name="TestSoap" binding="s0:TestSoap">
<soap:address location=
"http://localhost/SimpleWebService/Add.asmx" />
</port>
...
</service>
...
Proxy Classes
The wsdl tool can be used to read the WSDL description and generate a proxy class that
will make the SOAP calls for you. Since C# is the default language, and SOAP the
default protocol, the following command will generate a proxy class file with the name
addproxy.cs:
wsdl /out:addproxy.cs
http://localhost/SimpleWebService/Add.asmx?WSDL
The generated proxy defines a constructor and three methods. The constructor sets the
URL which this Web Service uses. One of the methods represents a synchronous,
blocking call on the Web Service. The other two methods correspond to the asynchronous
design pattern discussed in Chapter 8. If you want to call the Web Service
asynchronously you can use the BeginXXX and the EndXXX methods associated with
the proxy. [11] The proxy class has the same name as the WebService class.
[11]
Of course in this particular case XXX=Add.
The Invoke method of the SoapHttpClientProtocol class will make the HTTP request and
process the HTTP response associated with the transmitted and received SOAP packets.
This example is found in the SimpleAddClient subdirectory under the SimpleWebService
directory.
...
public class Test :
System.Web.Services.Protocols.SoapHttpClientProtocol
{
...
public Test()
{
this.Url ="http://localhost/SimpleWebService/Add.asmx";
}
...
public long Add(long x, long y)
{
object[] results = this.Invoke("Add",
new object[] {x, y});
return ((long)(results[0]));
}
...
public System.IAsyncResult BeginAdd(long x, long y,
System.AsyncCallback callback, object asyncState)
{
return this.BeginInvoke("Add",
new object[] {x, y}, callback, asyncState);
}
...
public long EndAdd(System.IAsyncResult asyncResult)
{
object[] results = this.EndInvoke(asyncResult);
return ((long)(results[0]));
}
}
You can then write a program to use the proxy classes to issue a Web Service request.
public class AddClient

{
{
Test z = new Test();
long f = z.Add(1, 2);
Console.WriteLine(f);
return;
}
}
Web Service Client with Raw SOAP and HTTP
To show you what the SoapHttpClientProtocol class does, the final client program for this
example uses sockets to send both the HTTP headers and the SOAP directly and to
receive the response from the Web Service. This example is the RawAddClient
subdirectory of the SimpleWebService.
The main routine first reads in a file that has the SOAP headers for the service to be
called. It returns the length of the content, which will have to be placed in one of the
HTTP POST headers.
long contentLength;
StringBuilder contentData = BuildContent("SoapAdd.txt",
out contentLength);
StringBuilder requestHeader = BuildHeader(contentLength);
It then connects to the server, sends the data, and receives the response, which it writes
out to the console.
IPEndPoint endPoint = new

IPEndPoint(Dns.Resolve(httpServer).AddressList[0],
httpPort);
Socket sock = new Socket(AddressFamily.InterNetwork,
SocketType.Stream, ProtocolType.Tcp);
sock.Connect(endPoint);
...
sock.Send(header, header.Length, 0);
sock.Send(content, content.Length, 0);
...
bytes = sock.Receive(receivedData, receivedData.Length,
0);
Console.WriteLine(ASCII.GetString(receivedData, 0,
bytes));
sock.Close();
...
The routine BuildHeader just builds a standard HTTP POST request with the addition of
the SOAPAction header.
StringBuilder sb = new StringBuilder(1024);

sb.Append("POST /SimpleWebService/Add.asmx HTTP/1.1\r\n");
sb.Append("Host: localhost\r\n ");
sb.Append("Content-Type: text/xml; charset=utf-8 \r\n");
string line = "Content-Length: " +
contentLength.ToString() + "\r\n"
sb.Append(line);
sb.Append("SOAPAction: \"urn:uuid:
10C14FCF-BF4A-477a-BFE7-41B9F2A4514E/Add\"\r\n ");
sb.Append("\r\n");
...
BuildContent just reads a file to a buffer and calculates the size of the buffer in bytes.
contentLength = 0;
String line;
while ((line = fileStream.ReadLine()) != null)
{
sb.Append(line);
sb.Append("\r\n");
contentLength += line.Length + 2;
}
fileStream.Close();
...
Based on our previous discussion, the SOAP file, SoapAdd.txt, looks as we would expect
it to. The input parameters "9" and "3" appear as the WSDL would dictate.

<soap:Envelope
<soap:Body>
<Add xmlns=
<x>9</x>
<y>3</y>
</Add>
</soap:Body>
</soap:Envelope>
The program first writes out the HTTP POST request. First come the standard HTTP
headers with a special SOAPAction header, then the SOAP encoding of the request.
POST /SimpleWebService/Add.asmx HTTP/1.1

Host: localhost
Content-Length: 393
SOAPAction:

<soap:Envelope
<soap:Body>
<Add xmlns=
<x>9</x>
<y>3</y>
</Add>
</soap:Body>
</soap:Envelope>
The program then writes out the response. Again, the HTTP headers come first, then the
SOAP encoding of the result, "12."
...
HTTP/1.1 200 OK
Server: Microsoft-IIS/5.0
Date: Mon, 17 Sep 2001 02:11:30 GMT
Cache-Control: private, max-age=0
Content-Length: 383
<soap:Envelope
xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:xsd="http://www.w3.org/2001/XMLSchema">
<soap:Body>
<AddResponse xmlns=
<AddResult>12</AddResult>
</AddResponse>
</soap:Body>
</soap:Envelope>
SOAP Differences
Before we finish our basic examination of SOAP and WSDL, a more detailed look at the
relationship of SOAP, WSDL, and the XML Schema specification is in order. As
mentioned earlier, the SOAP encodings used by .NET remoting differ from those used by
Web Services and the XML serializer.
To illustrate the differences between the two, we will take the same program and serialize
it to disk and use it as a Web Service. The program builds a circular list of two customer
items. The two programs are found in the SOAP Differences directory.
The first program, SOAP Formatter, creates a circular list and then serializes it to disk
using the .NET SOAP formatter. Although it is superfluous to do so, we derive the Test
class from the WebService class to demonstrate that what makes the difference is the way
SOAP is serialized, not the basic idea of Web Services.
using System.Web.Services;
using System;
using System.IO;
using System.Runtime.Serialization.Formatters.Soap;
[Serializable]
{
public string name;
public long id;
public Customer next;
}
public class Test: WebService

{
static void Main()

{
Test test = new Test();
Customer list = test.GetList();
FileStream s = new FileStream("cust.xml",
FileMode.Create);
SoapFormatter f = new SoapFormatter();
f.Serialize(s, list);
s.Close();
}
public Customer GetList()

{
Customer cust1 = new Customer();
cust1.name = "John Smith";
cust1.id = 1;
Customer cust2 = new Customer();

cust2.name = "Mary Smith";
cust2.id = 2;
cust2.next = cust1;
Web Service Class
As we have previously demonstrated, a Web Service is nothing but an HTTP request. As

such, a Web Service can access the intrinsic objects associated with its HTTP request.
These are the same intrinsic objects discussed in the section "State in ASP.NET
Applications" in the previous chapter. The WebService class has properties that access
these intrinsic objects.
You need not derive your Web Service class from the framework WebService class. You
can derive your Web Service class from a different base class if necessary. In this case
you can use the current HttpContext to access the intrinsic objects. The WebService class
inherits from MarshalByRefObject, however, so if you want your Web Service class to be
remotable, and you do inherit from a different base class, make sure that class also
inherits from MarshalByRefObject. The HttpContext enables you to get information
about an HTTP request. By using the static Current property, you can get access to the
current request.
We will now build a Web Service inside Visual Studio.NET that will illustrate the use of
these intrinsic objects inside a Web Service. As Figure 11-4 demonstrates, choose
ASP.NET Web Service from the New Project dialog box in Visual Studio.NET.
Figure 11-4. Visual Studio.NET New Project dialog with ASP.NET Web Service
project selected.
When you click the OK button, VS.NET will setup a Web Service project for you. By
default, the Web Service files are placed in a subdirectory of the IIS directory on your
hard drive. By default, projects are placed in a VSWebCache\MachineName subdirectory
under the Documents and Settings directory for the logged in user. Figure 11-5 shows the
resulting VS.NET project.
Figure 11-5. Visual Studio.NET Web Services project.
Our Web Service will have several methods that demonstrate how to use the intrinsic
objects. As you will see, this is really no different from their use in ASP.NET. Two of the
methods will illustrate the use of application and session state by calculating a cumulative
sum of numbers.
In the global.asax file we initialize our sum to zero in the appropriate event handlers.
Global.asax has the same function in Web Services as it does for ASP.NET, as discussed
in the previous chapter in the section "ASP.NET Applications." Since the Global class
inherits from System.Web.HttpApplication, it can access the Application and Session
intrinsic objects.

{
EventArgs e)
{
Application["TotalSum"] = 0.0;
}
protected void Session_Start(Object sender, EventArgs e)

{
Session["SessionSum"] = 0.0;
}
...
Renaming the Service1.asmx file to arithmetic.asmx. we define several Web methods. By

setting the EnableSession argument to the WebMethod constructor to true, we turn on
session state for the SessionSum method. Every time a new session is started, the sum is
reset to zero. On the other hand, for the CumulativeSum Web method, EnableSession is
set to its default value or false, so that the sum is reset to zero only when the Web Service
application is restarted. The Application intrinsic object is used from the HttpContext
object to show how that class is used.
It should be clear from this code that HttpApplication, WebService, and HttpContext all
reference the same intrinsic objects. If you need to save state for the application or
session of a Web Service, you can use the collections associated with
HttpApplicationState and HttpSessionState to do so.
...
[WebMethod(EnableSession = true)]
public double SessionSum(double x)
{
Session["SessionSum"] = (double)Session["SessionSum"]+x;
return (double)Session["SessionSum"];
}
[WebMethod]
public double CumulativeSum(double x)
{
double sum = (double) Application["TotalSum"];
sum = sum + x;
Application["TotalSum"] = sum;
return (double)HttpContext.Current.Application
["TotalSum"];
}
...
The GetUserAgent method show how to use the Context object to access information
about the request. We return what kind of application is accessing the Web Service. The
GetServerInfo method accesses the Server intrinsic object.
[WebMethod]
public string GetUserAgent()
{
return Context.Request.UserAgent;
}
[WebMethod]
public string GetServerInfo()
{
string msg = "Timeout for " + Server.MachineName + " = "
+ Server.ScriptTimeout + "; Located at " +
Server.MapPath("");
return msg;
}
The ArithmeticClient console program demonstrates the use of the Web Service. We can
create a proxy class from within VisualStudio.NET. On the Project Menu, select Add Web
Reference and type in the address of the Web Service in the Address edit box, followed
by a carriage return. Information about the Arithmetic Web Service will appear as in
Figure 11-6.
Figure 11-6. Visual Studio.NET display of Arithmetic Web Service information.
Click on the Add Reference button to add the Web reference. This will add a
WebReferences set of subdirectories below the current project that will contain the proxy
class and the wsdl file for the Web Service. To the client program we will have to
reference the proxy class's namespace:
using ArithmeticClient.localhost;
We then calculate a sum using the total held by the Application intrinsic object. Next we
calculate a sum for the total held by the Session intrinsic object.
Arithmetic a = new Arithmetic();

double sum;
for (int i = 0; i < 5; i++)
{
sum = a.CumulativeSum(i);
Console.WriteLine("Adding {0},
Application sum is now {1}", i, sum);
}
double sessionSum;
for (int i = 0; i < 5; i++)
{
sessionSum = a.SessionSum(i);
Session sum is now {1}", i, sessionSum);
}
This will give us the following output. The exact numbers for the application-based sum
will depend on how many times you have run the application.
Adding 0, Application sum is now 90

Adding 0, Session sum is now 0
We now create another instance of the proxy class and make the same method calls.
Arithmetic a2 = new Arithmetic();

for (int i = 0; i < 5; i++)
{
sum = a2.CumulativeSum(i);
Application sum is now {1}", i, sum);
}
for (int i = 0; i < 5; i++)

{
sum = a2.SessionSum(i);
Session sum is now {1}", i, sum);
}
We get the following output. Notice how the application sum continues to increase, while
the session bases sum starts again from zero. A new browser session is not the only way
to start a new Web Service session.

Finally we call the GetUserAgent and GetServerInfo Web methods.
Console.WriteLine(a2.GetUserAgent());
Console.WriteLine(a2.GetServerInfo());
The output will look something like this:
Mozilla/4.0 (compatible; MSIE 6.0; MS Web Services

Client Protocol 1.0.2914.16)
Timeout for MICAH = 90; Located at
f:\inetpub\wwwroot\Arithmetic
The next step in the case study is to make the Customer and Hotel components of the
Hotel Broker available as a Web Service. This Web Service is found in the
HotelBrokerWebService subdirectory of the case study for this chapter. This Web Service
will be used both by Acme's customers to make reservations as well as by administrators
for maintenance tasks associated with the Hotel Broker.
The proxy classes themselves are built into a proxies assembly. Two batch files that can
be used to create the proxy classes and build the assembly are located in the
WebServiceProxies subdirectory of the case study.
In the HotelBrokerAdministration subdirectory you will find a version of the admin

program that uses the proxies assembly instead of the Customer and Hotel assemblies. In
the AcmeWeb2 subdirectory for the case study you will find a version of AcmeLib that
references the proxies assembly instead of the Customer and Hotel assemblies. All
references to the Customer and Hotel components in the Acme reservation Web page and
HotelBrokerAdministration programs have been removed.
Since at this stage in the book you have a lot of experience with .NET, we do not spell out
the details of building the various pieces of the case study. Please consult the file
readme.txt in the CaseStudy directory if you would like some pointers.
Customer Web Service
To implement the customer Web service we created a file, CustomerWebService.asmx,

that uses the Customer component to implement the details of the Web Service:
[WebService(Namespace=
"urn:uuid:10C14FCF-BF4A-477a-BFE7-41B9F2A4514E")]
class CustomerWebService
{
private Customers customers;
public CustomerWebService()
{
customers = new Customers("HotelBroker");
}
[WebMethod]
{
int customerId;
customerId = customers.RegisterCustomer(firstName,
lastName, emailAddress);
return customerId;
}
[WebMethod]
public void UnregisterCustomer(int customerId)
{
customers.UnregisterCustomer(customerId);
}
[WebMethod]
[XmlInclude(typeof(CustomerListItem))]
public ArrayList GetCustomer(int customerId)
{
ArrayList ar;
ar = customers.GetCustomer(customerId);
return ar;
}
[WebMethod]
public void ChangeEmailAddress(int customerId,
{
customers.ChangeEmailAddress(customerId,
emailAddress);
}
}
The only new attribute is XmlInclude, which allows the XmlSerializer used to create the
SOAP protocol to serialize a custom type, in this case CustomerListItem. This attribute is
found in the System.Xml.Serialization namespace. Nonetheless, if you examine the proxy
class for this Web Service, which is found in the WebServiceProxies directory, you will
see that GetCustomer proxy (customerproxy.cs) returns only an array of objects.
public object[] GetCustomer(int customerId)
Although the attribute instructs the serializer to save the custom type, the SOAP protocol
understands only how to transmit a generic object type. So the AcmeLib code (Acme.cs)
has to treat the return type as an object and then extract the custom type from it.
object[] al = customers.GetCustomer
(hotelCustomerId);
foreach(CustomerListItem cust in al)
{
currentUser.HotelCustomerId = hotelCustomerId;
currentUser.FirstName = cust.FirstName;
currentUser.LastName = cust.LastName;
currentUser.EmailAddress = cust.EmailAddress;
}
All the other ArrayLists in the Customer and Hotel Web Services are treated as arrays of
objects where the appropriate type has to be extracted. Arrays that use types such as
strings and integers, however, need no special treatment by the XmlSerializer.
For the HotelBroker Web Service, the Hotel assembly itself was modified to be a Web
Service. The HotelWebService.asmx file has to make reference only to the HotelBroker
class in the Hotel assembly, which is located in the bin subdirectory of the Web Service.
<%@ WebService Language="C#"

class ="OI.NetCs.Acme.HotelBroker, Hotel" %>
The code is the same as the previous version of the component except for addition of the
necessary attributes to convert the code to a Web Service. Since Web Service names have
to be unique, we had to use the MessageName property of the WebMethod attribute to
give one of the overloaded GetHotels methods a unique name.
[WebMethod(MessageName="GetAllHotels")]
[XmlInclude(typeof(HotelListItem))]
public ArrayList GetHotels()
The code in Acme.cs is modified where necessary to handle the generic object[] arrays
that are returned instead of the CLR specific ArrayList type.
Design Considerations
Network latency is a major performance consideration. Hence, the number of requests

made over the network to a Web Service or a database should be minimized. In the
HotelBroker Web Service, the reservations for a customer are kept in the dataset as a
cache, so that only for database modifications does a database request have to be made.
The same is true for tracking the hotels and cities, although there is a trade-off here, since
an administrator might add a new hotel. However, that operation is not likely to occur
during the relatively short time a customer is making a reservation. Of course, these types
of data could be cached inside the Web form itself, so a call to the Web Service would be
unnecessary.
HTTP is a stateless protocol and therefore so is SOAP. Minimizing state will help your
applications and Web Services to scale better, because objects (such as database
connections) can be pooled or reused much more easily, and less memory is required so
that more resources are available to handle more requests. This means treating your Web
Service objects as endpoints of communication, not as full-fledged objects. Our case
study has not really done this, because we wanted to illustrate the use of certain
technologies, and the proper way to partition functionality really depends on the details of
your actual application and network latencies.
You can also use the CacheDuration property on a Web method or the Cache property of
the HttpContext class to cache information to avoid network overhead.
Summary
Web Services provide a means to extend component functionality across the network
between platforms and languages from different vendors. Unlike .NET remoting,
however, the types that can be used are much more limited.
Nonetheless, if you start your design from the point of view of the XML Schema
specification and then build your WSDL and Web Service classes, you will have a much
greater chance of being able to interoperate.
Chapter 12. Security Top
While Security considerations are fundamental to application design and should not be
left for last, pedagogically it is easier to talk about security once the .NET application
model, ASP.NET, and Web Services have already been introduced. This chapter
introduces to you the fundamentals of .NET security. [1]
[1]
Pedagogical reasons also dictate the form of the sample code. It is easier to
demonstrate security by starting with an open environment and then showing you how to
restrict operations. Real systems should start with the most restrictive security and then
open up only as needed.
Security prevents a user or code from doing things it should not be allowed to do.
Traditionally, security has focused on restricting user actions. .NET allows restrictions to
be placed on executing code. For example, you can prevent certain sections of code from
accessing certain files. This is particularly useful when you have public Web sites or
services where it is impractical to create user accounts, and lock down files or other
resources, for an unknown number of users. It is critical when you are executing code
that was created by third parties.
It is important to realize that .NET security sits on top of the underlying operating
system's security system. For the purposes of this chapter, the underlying operating
system is assumed to be Windows 2000. While we will discuss some security issues
associated with the underlying infrastructure, including Microsoft's Internet Information
Server (IIS), we will go into some detail only with those parts of the security story that
are relevant to .NET. [2]
[2]
For more information about secure Web-based applications, read Designing Secure
Web-Based Applications for Microsoft Windows 2000 by Michael Howard.
To give an example of the interaction of .NET security and the operating system, code
always runs under some identity, or in other words, as some user id. Irrespective of the
file creation .NET security permissions, if the file Access Control List (ACL) denies you
the right to create a file, you will be unable to create a file.
What makes the security story so difficult to tell is that it often seems that you have to
understand everything before you can do anything. For this reason, we will tell the
security story several times, each time with a little more detail. At the end you will be
able to understand the whole story.
The security story starts with an attempt to answer to two questions. The first is the
authentication question: Who are you? The second is the authorization question: Do you
have the right to do what you want? Under .NET this story takes two branches, because
the "you" can be either a user identity or an identity associated with an assembly.
We start with a brief telling of the security story by showing how both these types of
security exist in .NET. Although it is not needed immediately, a brief excursion into
Internet security follows, so that we can use that information when we need it. Then we
start the detailed narrative with role-based security in .NET.
User-Based Security
From the perspective of traditional user-based security, the authentication question is:
Who is the identity attempting to do the action? An identity is typically a user or account
name. Credentials are what you present to prove who you are; they are evidence
presented for verification. A credential might be your password, a smart card, or a
biometric device. The user's credentials must be verified with some security authority. An
example of this is verifying a user's password against their login name based on a
database of user names and encrypted passwords. Systems that allow unverified access
are said to allow anonymous access. In security lingo the identity that can be
authenticated is referred to as the principal.
The authorization question is: Can the identity perform the action they want? The
principal is then compared to some list of rights to determine whether access is allowed.
For example, when you access a file, your user name is compared with an Access Control
List for the action you want to do in order to determine whether you can access the file.
Of course, access is not always all or nothing. You might have read, but not modify rights
to a file.
In a multitier architecture, the identity under which the server executes is often very
powerful, and you want to restrict the ability of the client that makes a request to some
subset of privileges the server has. In other cases, such as anonymous access, the server
may not know who the client really is. The server then impersonates the client. Code
executes under the identity of the client, instead of the server. In the case of anonymous
access, the server runs under the identity of some preset user account.
Windows security under .NET, and ASP.NET security, are based on the concepts of user-
based security.
Code Access Security
One of the challenges of the software world of third-party components and downloadable
code is that you open your system to damage from executing code from unknown
sources. You might want to restrict Word macros from accessing anything other than the
document that contains them. You want to stop potentially malicious Web scripts. You
even want to shield your system from bugs of software from known vendors. To handle
these situations, .NET security includes Code Access Security (CAS).
Code Access Security can be applied to verifiable code only. During JIT compilation, the
verification process examines the MSIL to verify its type safety. As discussed previously,
type-safe code can only access memory locations it is supposed to. Pointer operations are
not allowed, so that methods can be entered or left only from well-defined entry points
and exit points. You cannot calculate an address and enter code at an arbitrary point.
Disallowing pointer operations means that random memory access cannot happen; code
can behave only in a restricted manner. [3]
[3]
Of course, bugs are still possible, but bugs cannot overwrite the stack, overrun a buffer,
or do anything that could be exploited to cause the program to do anything that it does
not have the security rights to do. If you give your code unlimited rights, then you do
have potential problems. This is especially true of the unmanaged code permission that
we will discuss later on.
Security Policy
Code Access Security is based on the idea that you can assign levels of trust to assemblies
and restrict the operation of the code within those assemblies to a certain set of
operations. Code-based security is also referred to as evidence-based security. The name
evidence stems from the fact that a set of information (or evidence) is used by the CLR to
make decisions about what this code is allowed to do. A piece of evidence might be the
location from which the code was downloaded, or its digital signature. Security policy is
the configurable set of rules that the CLR uses to make those decisions. Security policy is
set by the machine administrators. Security policy can be set at the enterprise, machine,
user, or application domain level.
Permissions
Security policy is defined in terms of permissions. Permissions are objects that are used
to describe the rights and privileges of assemblies to access other objects or undertake
certain actions. Assemblies request to be granted certain permissions. Security Policy
dictates what permissions will be granted to an assembly.
Examples of the classes that model permissions include:
• SecurityPermission that controls access to the security system. This includes the
right to call unmanaged code, control threads, control principals, app domain,
evidence and the like.
• FileIOPermission that controls access to the file system.
• ReflectionPermission that controls access to nonpublic metadata and the dynamic
generation of modules, types, and members.
All the permission classes inherit from the CodeAccessPermission base class, so they all
behave in the same way.
Attributes can be applied to the assembly to represent a request for certain permissions.
The CLR will use metadata to determine what permissions are being requested. Based on
the code's identity and trust level, the CLR will use security policy to determine whether
it can grant those permissions.
Code can programmatically demand (request) that its callers have certain permissions
before it will execute certain code paths. If the demand fails, the CLR will throw a
System.Security.SecurityException. Whenever you demand a permission, you have to be
prepared to catch that exception and handle the case where the permission was not
granted. Most programmers will not have to demand permissions, because the .NET
framework libraries will do that for you on your behalf. You still have to be prepared,
though, to handle the exceptions.
Code can also request that permissions it has been granted be restricted or denied. This is
important for code that uses third-party components or relies on third-party Web scripts.
Since such code may have a lower level of trust than your own code, you might want to
restrict the available rights while that code is running. When it is finished running, you
can restore the level of permissions back.
Determining the identity of the code is equivalent to the authentication question of

traditional security. The authorization question is based on the security permissions that
are given or taken away from an assembly.
Many of the classes that support permissions are found in the

System.Security.Permissions namespace. Some are found in the System. Net and
System.Data namespaces
Internet Security
You can use the Internet Protocol Security (IPSec) to restrict access to your computer to
certain IP addresses. Of course, you need to know the IP addresses of your clients. The
advantage is that you do not have to change your client application, ASP.NET code, or
Web Service code to use it. This is impractical for public Web sites or services where you
do not know who your clients are.
Internet Information Server
While the focus of this chapter is .NET security, some knowledge of IIS Security is
important. Since both Web Services and ASP.NET use IIS, your IIS settings do affect
.NET security.
In the previous chapters on ASP.NET and Web Services, we have used the default settings
of Anonymous access. Anonymous access does not require a user name or password to
access an account. You run under some default user account. Anonymous access is useful
for public Web sites and services that do their own authentication by asking for a user
name or password or by some other means. In such a scenario you could use ASP.NET
forms-based authentication. You can build forms to get the user name and password and
then validate them against a configuration file or database.
Internet Information Services supports the major HTTP authentication schemes. These
schemes require you to configure IIS appropriately. These schemes are listed in Table 12-
1. In each of these scenarios IIS authenticates the user if the credentials match an existing
user account. Secure Sockets Layer (SSL) is used whenever you need to encrypt the
HTTP communication channel. SSL can degrade performance. We do not discuss SSL in
this chapter.
Table 12-1. llS Authentication Schemes
Scheme Type of Authentication
Basic User and password information is effectively sent as plain text. This is
standard HTTP authentication and is not secure.
Basic over SSL Basic authentication, but the communication channel is encoded, so
that the user name and password are protected.
Digest Uses secure hashing to transmit user name and password. This is not a
completely secure method because the hash codes stored on the server
are reversible.[4] It was introduced in HTTP 1.1 to replace Basic
authentication.
Windows Traditional Windows security using NTLM or Kerberos protocols. IIS
Integrated authenticates if credentials match a user account. Cannot be used across
Security proxies and firewalls. NTLM is the legacy Windows security protocol.
Certificates Client obtains a certificate that is mapped to a user account.
over SSL
[4]
See the discussion of hash codes in Chapter 7. A message digest is another name for the
result of applying a hash code to a message.
You will also have to adjust access to the necessary files (graphics, data store files, etc.)
and other resources (i.e., databases) to those user accounts (authorization). For public
Web sites and Web services this approach is not useful because users will not have user
accounts.
Microsoft has introduced the Passport authentication scheme. While ASP.NET does have
support for Passport (System.Web.Security. PassportIdentity class) on the server side, as
of this writing developer tools to handle the client side for Passport authentication do not
yet exist. Passport avoids the problem of requiring specific accounts on specific
machines. We will not discuss Passport in this chapter.
The security specification for SOAP is being worked on by the W3C. You could create
your own custom authentication using SOAP messages. Since XML is transmitted as text,
you want to run using Secure Sockets Layer to encrypt the messages (especially if you
use tags such as <user> and <password>. In general, secure data has to be encrypted
when using SOAP.
Role-Based Security in .NET [5]

[5]
The discussion in this section is relevant for intranets or other scenarios where users
will have Windows user accounts on the servers or domains. See the later section "Forms-
Based Authentication" for a discussion of security appropriate to the scenario of a public
Web site.
Most people have at least an intuitive understanding of users and passwords. MTS and
COM+ have provided an easy-to-understand security system based on roles. The best
place to start a more detailed look at .NET security is with identities and roles. We will
look at this from the point of view first of a Windows application and then of ASP.NET.
Principals and Identities
Each thread has associated with it a CLR principal. The principal contains an identity
representing the user id that is running that thread. The static property
Thread.CurrentPrincipal will return the current principal associated with the thread.
Principal objects implement the IPrincipal interface. IPrincipal has one method and one
property. The Identity property returns the current identity object, and the method
IsInRole is used to determine whether a given user is in a specific role. The
RoleBasedSecurity example illustrates the use of principals, identities, and roles.
Currently there are two principal classes in the .NET framework: WindowsPrincipal and
GenericPrincipal. The GenericPrincipal class is useful if you need to implement your own
custom principal. The WindowsPrincipal represents a Windows user and its associated
roles.
Since the RoleBasedSecurity example is a Windows (console) application, we will have a

WindowsPrincipal associated with the CurrentPrincipal property.
...[6]
IPrincipal ip = Thread.CurrentPrincipal;
WindowsPrincipal wp = ip as WindowsPrincipal;
if (wp == null)
Console.WriteLine("Thread.CurrentPrincipal is NOT a
WindowsPrincipal");
else
Console.WriteLine("Thread.CurrentPrincipal is a
WindowsPrincipal");
...
[6]
The program starts out with a demand for a SecurityPermission and then proceeds to
set the AppDomain principal policy. While the reasons for this will be discussed later, the
quick answer is to make sure that the example functions properly on your machine. If you
get an exception, you will have to set the policy on your local machine to allow you to
run the example. On a vanilla system with a standard install, this should not happen.
What to do if it does happen is discussed later in the chapter.
An identity object implements the IIdentity interface. The IIdentity interface has three
properties:
• Name is the string associated with the identity. This is given to the CLR by either
the underlying operating system or the authentication provider. ASP.NET is an
example of an authentication provider.
• IsAuthenticated is a Boolean value indicating whether the user was authenticated
or not.
• AuthenticationType is a string that indicates which authentication was used by the
underlying operating system or authentication provider. Examples of
authentication types are: Basic, NTLM, Kerberos, Forms, or Passport.
Substitute Name of Your Machine in the Examples

In several of the examples the machine name MICAH is used. You should
substitute the appropriate machine or domain name when you run the samples on
your computer.
There are several types of identity objects. Since this is a Windows program, we will
have a WindowsIdentity object associated with the WindowsPrincipal. The example next
prints out the property information associated with the identity object.
IIdentity ii = ip.Identity;
Console.WriteLine("Thread.CurrentPrincipal Name: {0}
Type: {1} IsAuthenticated: {2}", ii.Name,
ii.AuthenticationType, ii.IsAuthenticated);
On my machine this is printed out:
Thread.CurrentPrincipal Name: MICAH\mds Type: NTLM

IsAuthenticated: True
The operating system on the machine MICAH using the NTLM protocol has
authenticated the user running this program to be "mds." The sample then validates that
this is indeed a WindowsIdentity object. The WindowsIdentity object has additional
properties and methods besides those of the IIdentity interface. One of them is the Win32
account token id associated with the currently running user.
WindowsIdentity wi = wp.Identity as WindowsIdentity;

if (wi != null)
Console.WriteLine("WindowsPrincipal.Identity Name: {0}
Type: {1} Authenticated: {2} Token: {3}", wi.Name,
wi.AuthenticationType, wi.IsAuthenticated, wi.Token);
You can use the name of the user to decide (authorize) whether the user has the rights to
undertake certain actions by refusing to execute certain code paths.
.NET Windows Roles
Instead of checking each individual user name, you can assign users to roles. You can
then check to see if a user belongs to a certain role. The standard administrators group is
an example of how a role works. You do not have to individually assign a user identity all
the privileges that an administrator has and then check to see if individual users have
certain privileges. Instead, you just assign the user to the administrators group. Code then
checks to see if a user is in the administrators group before attempting actions such as
creating a new user. .NET roles are separate from COM+ roles.
You define roles by defining groups in NT4 or Windows2000. Each group represents one
role. Go to the Control Panel and select Administrative Tools. From the Administrative
Tools list select Computer Management. In the Computer Management MMC snap-in
expand the Local Users and Groups node. As Figure 12-1 shows, if you select Groups
you will see all the Groups defined on your machine.
Figure 12-1. Groups defined on a machine.
Some groups, such as Administrators and Guests, are "built in" because they are
predefined for you. CustomerAdmin is a user-defined group that represents
administrators who have the right to modify Acme customer information.
To add a new group to the local machine, right-mouse-click on the Groups node and
select "New Group." A dialog box you can fill in pops up. Figure 12-2 shows this dialog
box filled for a new group entitled "HotelAdmin" which is designed to have all users on
the machine who can add or modify information about hotels in the HotelBroker system.
Clicking the Create button will add the group to the system. You can use the Add and
Remove buttons to add or remove users from the group.
Figure 12-2. Dialog to create a HotelAdmin group.
To modify an existing group, select that group, right-mouse-click, and select Properties.
Clicking the Add button will bring up a dialog of all users on the system. You can then
select users and add them to the group. Figure 12-3 shows a user about to be added to the
HotelAdmin group. The Remove button is used to remove users from the group.
Figure 12-3. User JaneAdmin about to be added to the HotelAdmin group. User
mds has already been added.
In addition to creating a HotelAdmin group, you should also create a CustomerAdmin
group with JaneAdmin as a member using the same procedure we just described. Note
that the JaneAdmin user need not, and in fact should not, be a member of the
Administrators group. Users should run with the minimum privilege required. Within
code you qualify the name using the domain or machine name. The CustomerAdmin role
is referred to as "MICAH\\CustomerAdmin." For groups that are preinstalled, such as the
Administrators group, you use the "BUILTIN" prefix—for example,
"BUILTIN\\Administrators." To avoid translation and internationalization problems, the
System.Security.Principal.WindowsBuiltInRole enumeration can be used to refer to built-
in roles. Instead of using the "BUILTIN\\Administrators" string you can refer to the
Administrators group as WindowsBuiltInRole.Administrator.
The RoleBasedSecurity example now checks to see if the current user is in a role. You
can either pass the role as a string or use the WindowsBuiltInRole enumeration.
Remember to modify the programs to use the name of your machine when you run the
book samples on your computer.
string adminRole = "MICAH\\CustomerAdmin";

bool inRole = wp.IsInRole(adminRole);
Console.WriteLine("In CustomerAdmin role?: {0}", inRole);
inRole = wp.IsInRole(WindowsBuiltInRole.Administrator);
Console.WriteLine("Is in Administrators group: {0}",
inRole);
inRole = wp.IsInRole(WindowsBuiltInRole.Guest);
Console.WriteLine("Is in Guests group: {0}", inRole);
inRole = wp.IsInRole(WindowsBuiltInRole.User);
Console.WriteLine("Is in Users group: {0}", inRole);
Other Identity Classes
Now let us look in more detail at the other Identity classes. Currently there are four in the
.NET Framework:
• FormsIdentity is used by the FormsAuthenticationModule class. We will discuss

this class when we discuss ASP.NET forms authentication.
• GenericIdentity can represent any user. This class is used with the
GenericPrincipal for generic or custom identities and principals.
• PassportIdentity is used with Passport authentication. Since we do not discuss
Passport, we will not discuss this class.
• WindowsIdentity represents a Windows user. A Windows Principal instance will
have a WindowsIdentity instance as its Identity property. For authenticated users,
the type of authentication used (NTLM, Kerberos, etc.) is available.
Note that the properties of the IIdentity interface are read-only and therefore cannot be
modified.
Even if your users are unauthenticated, you can get the WindowsIdentity for any thread
using the static method WindowsIdentity.GetCurrent to get the WindowsIdentity instance
of the current user. [7] You can then use the WindowsPrincipal constructor to build a
WindowsPrincipal instance from this WindowsIdentity.
[7]
We discuss what this represents in the next section.
The HotelBrokerAdminstration program has been modified so that you cannot run it if
you are not in the HotelBrokerAdmin role. See the file MainAdminForm.cs in the
directory HotelBrokerAdministration Roles.
static void Main()

{
...
IPrincipal ip;
ip = Thread.CurrentPrincipal;
string hotelAdminRole = "MICAH\\HotelAdmin";
bool inRole = ip.IsInRole(hotelAdminRole);
if (inRole == false)
{
MessageBox.Show("You cannot run this program since you
are not a Hotel Administrator.",
"Acme Customer Management System",
MessageBoxButtons.OK,
MessageBoxIcon.Exclamation);
return;
}
Application.Run(new MainAdminForm());
}
ASP.NET Roles
Now that we have a fundamental understanding about principals, identities and roles, we
can apply it to our AcmeReservationSystem Web site. The Web site has been modified so
that you can choose to link to a HotelAdministration page where you can add, modify, or
delete the hotels that are part of the HotelBroker system. This example is found in the
Step0 subdirectory of the ASP.NET Roles directory. To run this example, make sure that
the Step0 directory is a virtual directory with the name AcmeWebSecurityStep0. Figures
12-4 and 12-5 show the new Web pages.
Figure 12-4. The new Acme Home Page with the link to the administration
page.
Figure 12-5. The administration page for the AcmeReservation system.

Since at this stage in the book you have a lot of experience with .NET, we do not spell out
the details of building the various ASP.NET examples. Please consult the file readme.txt
in the current chapter directory if you would like some pointers.
At this point there is no security associated with these pages. Anyone who can log into
the Web site can access the administration page and modify the hotel information. We
have also modified the login page to print out the current principal and identity
information associated with the application as well as the information associated with the
current WindowsIndentity.
string text;
IPrincipal ip;
ip = Thread.CurrentPrincipal;
string principalText = "CurrentPrincipal is of type " +
ip.GetType().ToString();
IIdentity ii = ip.Identity;
principalText = principalText + "\n " +
"Is user authenticated?: " +
ii.IsAuthenticated.ToString();
text = principalText;
WindowsIdentity wi = WindowsIdentity.GetCurrent();
string identityText = "Current Windows Identity: " + "\n
" + "Name: " + wi.Name + "\n IsAuthenticated?:" +
wi.IsAuthenticated + "\n AuthenticationType:" +
wi.AuthenticationType;
text = text + "\n" + identityText;
IdentityInfo.Text = text;
As Figure 12-6 illustrates, looking at the information on the login page we find that we
have an unauthenticated generic principal for the thread, yet the current WindowsIdentity
indicates that we are running as the authenticated SYSTEM account. What does this
mean? In the previous examples we used the IsInRole method associated with the
CurrentPrincipal. But that user is now not authenticated, so that method will always
return false!
Figure 12-6. Principal and identity information about the Step0

AcmeWebSecurity Web site.
Operating System Identity and CLR Identity
As we mentioned at the start of the chapter, .NET security sits on top of the underlying
operating-system security. The identity associated with the thread by the CLR and the
identity associated with the thread by the underlying operating system are not the same.
The identity of the thread from the operating-system perspective is reflected by the
setting of the WindowsIdentity object returned by the static Windows.Identity.GetCurrent
method. The CLR identity is reflected by the value of the Thread.CurrentPrincipal object.
[8]
To go back to the example mentioned at the start of the chapter, if you access a file
from within .NET, both the managed and unmanaged identities must have rights to the
file within their respective environments.
[8]
The reason why these were identical in the RoleBasedSecurity example is that we set
the application domain principal policy in the example to be
PrincipalPolicy.WindowsPrincipal. With the default ASP.NET settings in the config.web,
the PrincipalPolicy.UnauthenticatedPrincipal policy is used. For that policy,
Thread.CurrentPrincipal returns an unauthenticated GenericPrincipal object. We will
discuss principal policy later.
What values the current WindowsIdentity and Thread.CurrentPrincipal have are set in
two places: IIS Settings and the ASP.NET configuration files.
Unauthenticated Users
Every machine that runs .NET has a machine.config file that has the default configuration
for the computer. This file is found in the
\WINNT\Microsoft.NET\Framework\v1.0.2914\CONFIG directory, where v1.0.2914
would be replaced by the version of Microsoft.NET that is running on your machine. A
Web or Web Service application may have a config.web file that has the configuration
settings for that application. The settings for config.web affect all applications in the
directory where it lives and all its subdirectories. Config.web files in the subdirectories
override the settings in the higher-level directories.
If you look in the settings in config.web for the Step0 project, you will see the following
settings:
<identity impersonate="false" />

The first value sets the unmanaged identity returned by the current WindowsIdentity.
Since it is set to false, the default operating system identity that ASP.NET runs as will be
the SYSTEM account. Since this has broad privileges on the local machine, the Web
application can run unimpeded, but this is undesirable from a security perspective, as we
will discuss later. The second sets the managed, CLR-based identity returned by
Thread.CurrentPrincipal. Setting it to "None" means use the default or GenericPrincipal.
The login page displays the current security configurations, as was shown in Figure 12-6.
Here is the relevant output.
CurrentPrincipal is of type
System.Security.Principal.GenericPrincipal
Is user authenticated?: False
Name:
Current Windows Identity:
Name: NT AUTHORITY\SYSTEM
IsAuthenticated?:True
AuthenticationType:NTLM
If you do not have a config.web file, the authentication mode set in machine.config is
"Windows." Now if we set the authentication mode in our local config.web to
"Windows," we see the following output:
System.Security.Principal.WindowsPrincipal
Name:
Name: NT AUTHORITY\SYSTEM
Thread.CurrentPrincipal now returns a WindowsPrincipal, but it is still unauthenticated

and has no name associated with it. Other values for the authentication mode are Forms
and Passport.
Let us set the authentication mode back to "None." But now let us set the identity
impersonate to "true":
<identity impersonate="true" />

Here are the results:
System.Security.Principal.GenericPrincipal
Name:
Name: MICAH\IUSR_MICAH
Where does the identity MICAH\IUSR_MICAH [9] come from? This user is the identity
that is set in the properties for this Web application for anonymous access. Select this
Web application in the Internet Services Manager, right-mouse-click, and select
Properties. Navigate to the Directory Security tab. Click on the Edit button associated
with Anonymous access and authentication control. Note that the Anonymous access
checkbox is checked. Click the Edit button associated with Account used for anonymous
access and you will see this user account listed. Figure 12-7 shows the related dialog
boxes. You could change this setting to some other account, but this is the default value
set when IIS is installed.
[9]
As usual, MICAH is the name of my machine. Yours will be different.
Figure 12-7. Internet Services Manager settings for anonymous access.
Reset the authentication mode back to "Windows" and run again. [10] We still do not see
an authenticated principal for the managed Thread.CurrentPrincipal identity.
[10]
To duplicate the results in the next section make sure you reset the authentication
mode back to "Windows" now.
Authenticated Users
Now let us use the Internet Services Manager to set our Web application to use Windows
Integrated Security instead of anonymous access, as shown in Figure 12-8. Right-click
over "AcmeWebSecurityStep0" in the left pane and choose Properties from the context
menu. We uncheck the anonymous access box and check the Integrated Windows
authentication box.
Figure 12-8. Internet Services Manager settings for authenticated access.

Running our application gets the following results:
Is user authenticated?: True
Name: MICAH\Administrator
Name: MICAH\Administrator
We now have an authenticated Thread.CurrentPrincipal whose identity is the same as the

current WindowsIdentity. They are associated with whatever user account is currently
logged in. Both the managed and unmanaged principals are the same. Now uncheck the
Integrated Windows authentication box and check the Basic authentication box in the
Internet Services Manager dialog, as shown in Figure 12-9.
Figure 12-9. Internet Services Manager settings for Basic authentication.

If you run the application now, you will have a user and password dialog appear when
your run the Web application, as shown in Figure 12-10.
Figure 12-10. Dialog for entering a Windows user name and password.
You now have to enter the user name and password associated with an account on the
system. Again, when the login page appears, both the Thread.CurrentPrincipal and
current WindowsIdentity identities are the same, but they are associated with whichever
user account you entered into the dialog box, as shown:
Is user authenticated?: True
Name: MICAH\JaneAdmin
Name: MICAH\JaneAdmin
How did the identity associated with the CurrentPrincipal get set to be the same as the
WindowsIdentity? ASP.NET sets the CurrentPrincipal to match the HttpContext.User
property. In a Windows application you have no choice but to use the
Thread.CurrentPrincipal. Within ASP.NET it is safer to use the HttpContext.User
property. Within ASP.NET you can access the HttpContext.User property through the
User object. Step 1 of ASP.NET Roles adds the following code to the Page_Load method
of main.aspx.cs:
if (User.IsInRole("MICAH\\HotelAdmin"))
HotelAdminLink.Visible = true;
else
HotelAdminLink.Visible = false;
The Internet Services Manager security should be set to at Windows Integrated security.
The following settings are still in web.config:
<identity impersonate="true" />

<authentication mode="Windows" />.
Therefore, any user logged into Windows who is a member of the HotelAdmin group,
will see the Administration link, otherwise the link will not appear. Of course, what name
you enter into the login page has nothing to do with what you see. It is the identity
associated with the thread that matters.
If you want to test your Web application as a different user, you do not have to log out
and log in as that user. Navigate to Internet Explorer on the Start Menu, and right-mouse-
click while holding down the shift key. You will see a menu item "Run As..." (see Figure
12-11). Select it, and in the dialog box that comes in, log in as the user you want to use.
That particular instance of Internet Explorer will be running under that user identity.
Figure 12-11. RunAs Menu item to run as a different user.

Problems with Impersonation
It would seem that we need only make sure that the user id the thread impersonates is a
member of the HotelAdmin group and does not have any more privileges than are needed
(i.e., is not System or an administrator, with no ACL rights to any unnecessary files on
the server) and then everything will be just fine.
Unfortunately, life is not so simple. Impersonation was designed to be used by a server to

alter its rights by running a thread as another user. When the server is done impersonating
a user, however, it can revert to its original set of rights by calling the RevertToSelf
Win32 API. If you call out to a third party or any unmanaged code DLL running in your
process, and it made a call to RevertToSelf, it would be running as SYSTEM. As
SYSTEM this DLL could cause havoc on your system if it were malicious or just buggy.
Step 2 of ASP.NET Roles uses the following code to do this: [11]

[11]
You use the PInvoke interop facility to access the Win32 function. D11Import and
PInvoke are discussed in Chapter 14.
using System.Runtime.InteropServices;
...
[DllImport("Advapi32.dll")]
public static extern bool RevertToSelf();
...
string text;
text = "Windows Identity: " +
WindowsIdentity.GetCurrent().Name + "\n";
text = text + "CLR Identity: " + User.Identity.Name +
"\n";
text = text + "Calling RevertToSelf()...\n";
bool bRet = RevertToSelf();
text = text + "Windows Identity: " +
WindowsIdentity.GetCurrent().Name + ``\n´´;
text = text + "CLR Identity: " + User.Identity.Name +
"\n";
txtInfo.Text = text;
On the Acme Home Page, calling RevertToSelf changes the identity of the thread from
the point of view of unmanaged code. The identity from the CLR perspective is
unchanged. The HotelAdmin link will be visible or not, depending on the original
impersonated identity. Figure 12-12 shows the results.
Figure 12-12. Acme Home page showing changes in Thread Identities.
To avoid running as the SYSTEM account, you can set the identity of the process that
your Web application runs under. [12], [13] If you look in machine.config under the
<processModel> tag, you will find the enable, userName, and password attributes.
[12]
SYSTEM is the identity of the Process Token for your application. Unless
impersonating, all threads in the process would use that token. Calling RevertToSelf
removes the impersonation from the thread and reverts back to whatever identity the
Process Token had.
[13]
On IIS 5, the identity is inherited from inetinfo.exe. If you configure inetinfo and
iisadmin to run with a different identity, that will be the identity of the aspnet_wp
process. On IIS 6 with Windows.NET server, ASP.NET does not run its own process
model and inherits identity from the IIS worker process. This worker process is
configurable and defaults to Network.Service. This is a much better default then
SYSTEM.
<processModel enable="true"
...
userName="SYSTEM" password="AutoGenerate"
...
/>
By default, your application process runs under the SYSTEM account. You can modify
this value in the machine.config file only. We could change the value to be a specific user
name: [14]
[14]
You will have to stop and start the WWW service on your machine to make the
changes to machine.config effective.
...
userName="JaneAdmin" password="xyz"
...
/>
Figure 12-13 shows the results. [15] As you can see, the password for this user is written in
plain text inside machine.config. By default, machine.config is readable by everyone, so
if you use this approach, rights to that file should be restricted.
[15]
If you have problems running with a user id you supply here, that id will probably
need ACL rights to various system directories on your machine, such as the ASP.NET
temporary file directory.
Figure 12-13. Results of using RevertToSelf when a specific user is set in

<processModel>.
You can also use a setting called MACHINE which uses an account called ASPNET,
which is a member of the guest group only and therefore has limited access rights. The
ASP.NET setup does add access rights for that user to system directories to enable it to
run, such as the \WINNT\Microsoft.NET\Framework\vx.xxx directory. Figure 12-14
shows the results of this setting.
Figure 12-14. Results of using RevertToSelf when generic ASP.NET account is

used in <processModel>.
...
userName="machine" password="AutoGenerate"
...
/>
To summarize, if impersonation is turned off, as Figure 12-15 shows, then you would run
as whatever identity is specified in the process model. If you use anonymous access, then
Figure 12-16 shows the results you would expect, that the CLR thread identity is
unauthenticated.
Figure 12-15. Impersonation turned off, MACHINE specified in

<processModel>.
Figure 12-16. Impersonation turned off, anonymous access, MACHINE

specified in <processModel>.
This discussion also makes clear that the ACLs on machine.config and web.config should
be set so that only administrators can modify the file. Who can read the files should be
restricted appropriately. You also have to guard against someone browsing and
downloading from those files.
Specifying Users and Groups for Access in Web.Config
ASP.NET allows you to specify groups and users who are allowed to access the Web site.
Inside the <authorization> section of web.config you can use the <allow> and <deny>
elements with user accounts or groups. To specify groups you use the roles attribute; to
specify users you use the users attribute. The asterisk (*) symbol used with one of those
elements means all. A question mark (?) used with a user attribute means "anonymous
access."
<allow roles="MICAH\HotelAdmin" users="MICAH\Peter">

<deny users="MICAH\John">
A reference to a specific user overrides their membership in a group or a wildcard. Deny

references take precedence over allow references. These settings do not help you assign
users to particular roles or prevent access to different areas of the Web site. Only access
to the entire Web site is controlled.
Forms-Based Authentication
The previous discussion is relevant for intranets or other scenarios where users will have
Windows user accounts on the servers or domains. Furthermore, Windows Integrated
Security does not work across firewalls or proxies. For public Web sites we need another
approach.
The alternative approach is to bring up a login form to authenticate the user. We will look
at two of the several approaches to login forms that are possible within .NET. Step 0 of
the FormsBasedAuthentication example uses the .NET FormsAuthentication class and
the config.web file. Step 1 of the example uses a database login to illustrate using an
external database.
Forms Authentication and Authorization
.NET Forms-based authentication uses web.config, login form, and a cookie to

authenticate the user. [16] Typically in this scenario you will set up the Web site for
anonymous access so that no users will be screened out by IIS. Here is the web.config file
section for the Step 0 example:
[16]
You do not have to use a cookie, but it is used for automatic authentication.
<authentication mode="Forms">
<forms name = "HotelBrokerCookie" path="/"
loginUrl="Login.aspx"
protection="All" timeout="10">
<credentials passwordFormat=´´Clear´´>
<user name="Natasha" password="Natasha" />
<user name="Adams" password="Adams" />
<user name="peter" password="peter" />
</credentials>
</forms>
</authentication>
<authorization>
<allow users="Natasha,peter" />
<deny users="*" />
</authorization>
The authentication mode is set to Forms. This means that the User.Identity object will be
a FormsIdentity instance if the user is authenticated. The forms element has several
attributes that define how the authentication is set up. The name attribute is the name of
the cookie. The path attribute indicates where on the site the cookie is valid; "/" indicates
the entire site. The loginUrl indicates where the login form resides. The protection
attribute indicates how the cookie should be encrypted. "All" indicates that the cookie
should be validated and encrypted. Other options are None, Encryption, and Validation.
Timeout indicates the number of minutes before the cookie becomes invalid (expires).
The credential elements indicate how the password should be stored in the configuration
file. For simplicity we have used clear text. You could also specify SHA1 or MD5 to
encrypt the passwords. [17] If passwords are stored in web.config, it should be secured
against download (which is the default). Passwords for the configuration file can be
encrypted with the static FormsAuthentication method
[18]
HashPasswordForStoringInConfigFile.
[17]
These encryption formats are discussed in Chapter 7.
[18]
Storing passwords in a configuration file is convenient for development and testing
work. If you do your own validation, as we do with the database example, you do not
need to use the web.config file.
The user elements indicate the user names and passwords. The authorization section, as
discussed earlier, determines which authenticated users are authorized to access the Web
site.
Since this example uses redirection and cookie validation, a user should attempt to access
the main page, http://localhost/
FormsBasedAuthenticationStep0/default.aspx, instead of the login.aspx file. If a
valid cookie does not exist on the system, the user will be sent to the login page. If a valid
cookie exists, it will be used to validate the user. If the user is validated, they will go
straight to the default.aspx page. If users went straight to the login page, they would have
to log in every time, even with a valid cookie.
Here is the code for handling the Login button event:
private void Login_Click(object sender, EventArgs e)

{
if (FormsAuthentication.Authenticate(txtUserId.Text,
txtPassword.Text))
{
FormsAuthentication.RedirectFromLoginPage(
txtUserId.Text, true);
}
else
{
lblErrorMessage.Text = "Could not authenticate user.";
}
For simplicity, the Password text box does not hide the password. A password text box
that hides the password was discussed in Chapter 10.
The FormsAuthentication class's Authenticate method validates the user name and
password from the web.config file. If a valid cookie was on the system, the user is not
redirected to the login page. RedirectFromLoginPage creates a cookie, and redirects the
user to the default.aspx page. If the second argument is true, a persistent cookie is placed
on the user's system. Persistent cookies are a security risk, because the cookie can be
stolen as it is transmitted (hijacked). You should use SSL to protect the cookie. You can
remove the session or persistent cookie with the SignOut method. The check of the
authorization section of web.config to see if the user has the rights to access the page is
done on each request.
If you run the Step 0 example only Natasha, peter, and Adams will be authenticated.
However, only Natasha and Peter will be authorized to use the site. Of course, only
Natasha will be found in the database of Acme customers. That test has been moved to
default.aspx to distinguish it from the forms authentication done in login.aspx.
Default.aspx can refer to the name of the user through the User object. The type of
identity object is FormsIdentity.
bool ok = HotelState.acme.Login(User.Identity.Name);
If you succeed and log in as Natasha once, subsequent tries will succeed without the login
page because we have created a persistent cookie. To avoid persistent cookies, set the
second argument to RedirectFrom LoginPage to false.
The application, however, runs under the identity of the system process or thread, not the
identity of the user name that is logged in. Hence, if you want to use role-based security
in ASP.NET with Forms authentication, you will have to create your own roles by using a
GenericPrincipal.
Database Login Validation
It is fairly straightforward to add password validation against a database as in Step 1 of

the FormsBasedAuthentication example. When a user is registered, the password is
hashed and is stored in the Acme database. [19]
[19]
This assumes that you do not have to mail the password to a user that has forgotten it.
Then you have to use two-way encryption.
string password =
FormsAuthentication.HashPasswordForStoringInConfigFile
(txtPassword.Text, "MD5");
bool ok = HotelState.acme.Register(txtUserId.Text,
password, txtFirstName.Text, txtLastName.Text,
txtEmailAddress.Text);
Before logging in, the password is again hashed and compared with the version stored in
the database.
string password =
FormsAuthentication.HashPasswordForStoringInConfigFile
(txtPassword.Text, "MD5");
bool ok = HotelState.acme.Login(txtUserId.Text, password);
With this approach you would have to maintain your own data store to track who is or is
not a hotel administrator in order to decide who can see the hotel administration page.
Code Access Permissions
Code needs permissions in order to access a resource such as a file, or perform some
operation. Security Policy (discussed later in the chapter) will give certain permissions to
each assembly. Code access permissions can be requested by code. The CLR will decide
which permissions to grant based on the security policy for that assembly. We will not
discuss how to write a custom permission.
Here are some examples of Code access permissions:
• DNSPermission controls access to Domain Name servers on the network.

• EnvironmentPermission controls read or write access to environment variables.
• FileIOPermission controls access to files and directories.
• FileDialogPermission allows files selected in an Open dialog box to be read. This
is useful if FileIOPermission has not been granted.
• ReflectionPermission controls the ability to access nonpublic metadata and emit
metadata.
• RegistryPermission controls the ability to access and modify the registry.
• SecurityPermission controls the use of the security subsystem.
• SocketPermission controls the ability to make or accept connections on a
transport address.
• UIPermission controls the user of various user-interface features including the
clipboard.
• WebPermission controls making or accepting connections on a Web address.
The use of these permissions is referred to as Code Access Security because this
permission is based not on the identity of the user running the code, but on whether the
code itself has the right to take some action.
Simple Permission Code Request
The SimplePermissionCodeRequest example first requests permission to access a file. If

the CLR does not grant that request, the CLR will throw a SecurityException inside the
file constructor. However, this code first tests to see if it has that permission. If it does
not, it just returns instead of trying to access the file. [20]
[20]
We have not yet discussed how you set security policy so you do not yet know how to
grant or revoke this permission. By default, however, code running on the same machine
that it resides on has this permission. This is another example of how difficult it is to talk
about security without knowing the whole picture.
This step is generally superfluous because the CLR will do the demand inside the
constructor, but often you want to check permissions before you execute some code to
ascertain whether you have the rights you need.
The FileIOPermission class models the CLR file permissions. A full path must be
supplied to its constructor, and we use the Path class we discussed in Chapter 8 to get the
full path. We are asking for read, write, and append file access. Other possible access
rights are NoAccess or PathDiscovery. The latter is required to access information about
the file path itself. You might want to allow access to the file, but you may want to hide
information in the path such as directory structure or user names.
The demand request checks to see if we have the required permission. The Demand
method checks all the callers on the stack to see if they have this permission. In other
words, we want to make sure not only that the assembly this code is running in has this
right, but that all the assemblies this code is running on behalf of have this permission. If
an exception was generated, we do not have the right we demanded, so we exit the
program.
string filename = ".\\read.txt";

string fileWithFullPath = Path.GetFullPath(filename);
try
{
FileIOPermission fileIOPerm = new
FileIOPermission(FileIOPermissionAccess.AllAccess,
fileWithFullPath);
fileIOPerm.Demand();
}
catch(Exception e)
{
return 1;
}
try
{
FileInfo file = new FileInfo(filename);
StreamReader sr = file.OpenText();
string text;
{
}
sr.Close();
}
catch(Exception e)
{
}
Even if the code has the CLR read permission, the user must have read permission from
the file system. If the user does not, an UnauthorizedAccessException will be thrown
when the OpenText method is called.
You have to be careful in passing objects that have passed a security check in their
constructor to code in other assemblies. Since the check was made in the constructor, no
other check is made by the CLR to ascertain access rights. The assembly you pass the
object to may not have the same rights as your assembly. If you were to pass this FileInfo
object to another assembly that did not have the CLR read permission, it would not be
prevented from accessing the file by the CLR, because no additional security check
would be made. This is a design compromise for performance reasons to avoid making
security checks for every operation. This is true for other code access permissions as
well.
How a Permission Request Works
To determine whether code is authorized to access a resource or perform an operation, the

CLR checks all the callers on the stack frame, making sure that each assembly that has a
method on the stack can be granted the requested permission. If any caller in the stack
does not have the permission that was demanded, a SecurityException is thrown.
Less trusted code cannot use trusted code to perform an unauthorized action ("luring
attack"). The procedures on the stack could come from different assemblies that have
different sets of permissions. For example, an assembly that you build might have all
rights, but it might be called by a downloaded component that you would want to have
restricted rights (so it doesn't open your email address book).
As discussed in the next sections, you can modify the results of the stack walk by using
Deny or Assert methods on the CodeAccessPermission base class.
Strategy for Requesting Permissions
Code should request permissions that it needs before it uses them, so that it is easier to
recover if the permission request is denied. For example, if you need to access several
key files, it is much easier to check to see if you have the permissions when the code
starts up rather than when you are halfway through a delicate operation and then have to
recover. Users could be told up front that certain functions will not be available to them.
Or, as we will discuss later, you could use assembly permission requests, and then fail to
load if the required permissions are not present. The problem is that you may not know
what permissions request will succeed because you do not know what assemblies will
have callers on the stack when the request is made.
You should not request permissions that you do not need. This will minimize the chances
that your code will do damaging things from bugs or malicious third-party code and
components. In fact you can restrict the permissions you have to the minimum necessary
to prevent such damage. For example, if you do not want a program to read and write the
files on your disk, you can deny it the right to do so.
Denying Permissions
One can apply the Deny method to the permission. Even though security policy would
permit access to the file, any attempt to access the file will fail. The
SimplePermissionCodeDenial example demonstrates this. Instead of demanding the
permission, we invoke the Deny method on the FileIOPermission object.
...
try
{
fileIOPerm.Deny();
Console.WriteLine("File Access Permission Removed");
}
catch(SecurityException se)
{
Console.WriteLine(se.Message);
}
We then try to read the file using the ReadFile method. Why we do this inside another
method will be explained shortly. Since the permission was denied, the FileInfo
constructor will throw a SecurityException.
...
try
{
string text;

{
Console.WriteLine(""+ text);
}
sr.Close();
}
{
Console.WriteLine("Could not read file: " +
se.Message);
}
We then call the static RevertDeny method on the FileIOPermission class to remove the
permission denial, and we attempt to read the file again. This time the file can be read.
The call to Deny is good until the containing code returns to its caller or a subsequent call
to Deny. RevertDeny removes all current Deny requests.
...
FileIOPermission.RevertDeny();
...
ReadFile();
We then invoke the Deny method to once again remove the permission.
Asserting Permissions
The Assert method allows you to demand a permission even though you do not have
access rights to do so. You might also want to assert a permission because other calls in
the call chain do not have the right, even though your assembly does. You can only assert
permissions that your assembly has been granted. If this were otherwise, it would be
trivial to circumvent CLR security. [21]
[21]
You also need the permission to assert.
The test program code now asserts the FileIOPermission and then attempts to read the
file.
...
...
fileIOPerm.Deny();
...
fileIOPerm.Assert();
...
ReadFile();
ReadFileWithAssert(fileIOPerm);
...
ReadFile();
But the file read fails! The assertion is good only within the method that called. The
ReadFileWithAssert method can read the file because it asserts the permission within the
method and then attempts the read. Assert stops the permission stack walk from checking
permissions higher in the stack frame and allows the action to proceed, but it does not
cause a grant of the permission. Therefore, if code further down the stack frame (like
ReadFile) tries to demand the denied permission (as the FileInfo constructor does), a
SecurityException will be thrown. [22] Similarly, Deny prevents callers higher in the stack
frame from an action, but not on the current level.
[22]
This is true as well for code above you on the stack frame.
public static void ReadFileWithAssert(FileIOPermission f)
{
...
f.Assert();
...
string text;

{
Console.WriteLine(""+ text);
}
sr.Close();
...
}
Remember that the assert applies only to IO operations done in this routine for the
specific file that was passed the FileIOPermission constructor. The call to Assert is good
until the containing code returns. Hence, ReadFile fails again when it is attempted after
ReadFileWithAssert returns. RevertAssert removes all current Assert requests.
Assert opens up security holes, because some caller in the stack frame might be able to
use the routine that calls assert to violate security. Any use of Assert should be subject to
a security review.
Other Permission Methods
PermitOnly specifies the permissions that should succeed. You specify what resources
you want to access. The call to PermitOnly is good until the containing code returns, or a
subsequent call to PermitOnly. RevertPermitOnly removes all current PermitOnly
requests. RevertAll removes the effect of Deny, PermitOnly, and Assert.
SecurityPermission Class
The SecurityPermission class controls "metapermissions" that govern the CLR security
subsystem. Let us look again at the RoleBasedSecurity example from earlier in the
chapter. It used the AppDomain.SetPrincipalPolicy method to set the application
domain's principal policy:
AppDomain ap = AppDomain.CurrentDomain;
ap.SetPrincipalPolicy(PrincipalPolicy.WindowsPrincipal);
The type of principal returned by Thread.CurrentPrincipal will depend on the Application

Domain's Principal Policy. An Application Domain can have one of three authentication
policies as defined by the System.Security.PrincipalPolicy enumeration:
• WindowsPrincipal uses the current user associated with the thread.
Thread.CurrentPrincipal returns a WindowsPrincipal object.
• UnauthenticatedPrincipal uses an unauthenticated user. Thread.CurrentPrincipal
returns a GenericPrincipal object. This is the default.
• NoPrincipal returns null for Thread.CurrentPrincipal.
You set the policy with the SetPrincipalPolicy method on the AppDomain instance for the
current application domain. The static method AppDomain.CurrentDomain will return
the current instance. This method should be called before any call to
Thread.CurrentPrincipal, because the principal object is not created until the first attempt
to access that property.
In order for the RoleBasedSecurity example to set the principal policy it needs to have
the ControlPrincipal right. To ascertain if the executing code has that right, you can
demand that SecurityPermission before you change the policy. A SecurityException will
be thrown if you do not have that permission.
...
SecurityPermission sp = new SecurityPermision(
SecurityPermissionFlag.ControlPrincipal);
try
{
sp.Demand();
}
{
Console.WriteLine(se.Message);
return 1;
}
We first construct a new SecurityPermission instance, passing to the constructor the

security permission we want to see whether we have the right to use.
SecurityPermissionFlag is an enumeration of permissions used by the SecurityPermission
class. The ControlPolicy permission represents the right to change policy. Obviously, this
should be granted only to trusted code. We then demand (request) the permission.
As mentioned earlier, you can only assert permissions that your assembly actually has. So
rogue components cannot just assert permissions when running within your code. You
can either set security policy or use the SecurityPermission class to prevent components
from calling Assert. Construct an instance of the class with the
SecurityPermissionFlag.Assertion value and then Deny the permission. Other actions you
can control with the SecurityPermission class include the ability to create and manipulate
application domains, specify policy, allow or disallow execution, control whether
verification is performed, or access unmanaged code.
Unmanaged Code
Asserts are necessary for controlling access to unmanaged code, since managed code
should not call unmanaged code directly.
In order to call unmanaged code you need the unmanaged code permission. [23] Since the
CLR performs a stack walk to check whether all the callers have unmanaged code
permission, you would have to grant all code the unmanaged code permission. Hence,
assemblies other than your own trusted ones could perform operations through the Win32
API calls and subvert the framework's security system. [24]
[23]
As with all the other "security permissions" this is technically a flag on the
SecurityPermission class, but the common parlance is to call them permissions.
[24]
The underlying operating system identity that is running the program must have the
rights to perform the operating system function.
Better would be to make calls through wrapper classes that are contained in an assembly
that has the managed-code right. The code in the wrapper class would first ascertain that
the caller has the proper CLR rights by demanding the minimal set of permissions
necessary to accomplish the task (such as writing to a file). If the demand succeeds, then
the wrapper code can assert the right to managed code. [25] No other assembly in the call
chain then needs to have the managed-code right.
[25]
By demanding first, then asserting, you ensure that a luring attack is not in progress.
For example, if you ask the .NET file classes to delete a file, they first demand the delete
permission on the file. If that permission is granted, then the code asserts the managed
code permission and calls the Win32 API to perform the delete.
Attribute-Based Permissions
The SimplePermissionAttributeRequest example shows how you can use attributes to

make permission requests. This example uses an attribute to put in the metadata for the
assembly that you need to have the ControlPrincipal permission to run. This enables you
to query in advance which components conflict with security policy.
[assembly:SecurityPermission(
SecurityAction.RequestMinimum,ControlPrincipal=true)]
public class pp
{
...
The SecurityAction enumeration has several values, some that can be applied to a class or
method and some that can be applied to an assembly as in this example. For assemblies
these are RequestMinimum, RequestOptional, and RequestRefuse. RequestMinimum
indicates to the metadata those permissions the assembly requires to run.
RequestOptional indicates to the metadata permissions that the assembly would like to
have, but can run without. RequestRefuse indicates permissions that the assembly would
like to be denied. [26]
[26]
An assembly would do this to prevent code from another assembly executing on its
behalf from having this permission.
If you change the attribute in this example to RequestRefuse and run it, you will find that
the assembly will load, but you will get a SecurityException when you attempt to change
the policy.
Other values apply to classes and methods. LinkDemand is acted upon when a link is
made to some type. It requires your immediate caller to have a permission. The other
values apply at runtime. InheritanceDemand requires a derived class to have a
permission. Assert, Deny, PermitOnly, and Demand do what you would expect.
Here is an example of a FileIOPermission demand being applied to a class through an

attribute. AllAccess is being demanded of the file. A full file path is required.
[FileIOPermission(SecurityAction.Demand,
All = "c:\\foo\\read.txt")]
public class Simple
...
Principal Permission
Role-based security is controlled by the PrincipalPermission class. The

PrincipalPermission example uses this class to make sure that the user identity under
which the program is being run is an administrator. We do that by passing the identity
name and a string representing the role to the constructor. Once again, we use the
Demand method on the permission to check the validity of our permission request.
PrincipalPermission PrincipalPerm = new

PrincipalPermission(wi.Name, adminRole);
try
{
PrincipalPerm.Demand();
Console.WriteLine("Code demand for an administrator
succeeded.");
}
catch(SecurityException)
{
Console.WriteLine("Demand for Administrator failed.");
}
If the running user were an administrator the demand would succeed; otherwise it would
fail with an exception being thrown. The code then checks to see if the user with the
name JaneAdmin (not a system administrator, but part of the CustomerAdmin group) and
the designated role is running.
string customerAdminRole = "MICAH\\CustomerAdmin";

PrincipalPermission pp;
pp = new PrincipalPermission("MICAH\\JaneAdmin",
customerAdminRole);
try
{
pp.Demand();
Console.WriteLine("Demand for Customer Administrator
succeeded.");
}
{
Console.WriteLine("Demand for Customer Administrator
failed.");
}
The CodeAccessPermission base class has methods for creating permissions that are the
union or the intersection of several permissions. PrincipalPermission does not derive
from CodeAccessPermission because it is based on the identity associated with the code,
not on the rights of the code itself. Nonetheless, it shares the same idioms with the
CodeAccessPermission derived classes.
Next the example code sees if either of these two administrators is the identity of the
running code.
string id1 = "MICAH\\Administrator";

string id2 = "MICAH\\mds";
PrincipalPermission pp1 = new PrincipalPermission(id1,

adminRole);
PrincipalPermission pp2 = new PrincipalPermission(id2,
adminRole);
IPermission ipermission = pp2.Union(pp1);

try
{
ipermission.Demand();
Console.WriteLine("Demand for either administrator
succeeded.");
}
{
Console.WriteLine("Demand for either administrator
failed.");
}
The code then sees whether any administrator is the identity of the running code. [27]
[27]
A null user and a role as arguments to mean anyone in that role is not an intuitive use
of null.
PrincipalPermission pp3 = new PrincipalPermission(null,

adminRole);
try
{
pp3.Demand();
Console.WriteLine("Demand for any administrator
succeeded.");
}
{
Console.WriteLine("Demand for any administrator
failed.");
}
If the users are unauthenticated, even if they do belong to the appropriate roles, the
Demand will fail.
PermissionSet
You can deal with a set of permissions through the PermissionSet class. The
AddPermission and RemovePermission methods allow you to add instances of a
CodeAccessPermission derived class to the set. You can then Deny, PermitOnly, or
Assert sets of permissions instead of individual ones. This makes it easier to restrict what
third-party components and scripts might be able to do. The PermissionSet example
demonstrates how this is done.
We first define an interface IUserCode that our "trusted" code will use to access some
"third-party" code. While in reality this third-party code would be in a separate assembly,
to keep the example simple we put everything in the same assembly.
public interface IUserCode

{
int PotentialRogueCode();
}
public class ThirdParty : IUserCode

{
public int PotentialRogueCode()
{
try
{

string text;

{
}
sr.Close();
}
catch(Exception e)
{
}
return 0;
}
}
Our code will create a new instance of the "third party" which would cause the code to be
loaded into our assembly. We then invoke the OurCode method passing it the "third-
party" code.
...
{
ThirdParty thirdParty = new ThirdParty();
OurClass ourClass = new OurClass();
ourClass.OurCode(thirdParty);
return 0;
}
Now let us look at the OurCode method. It creates a permission set consisting of
unrestricted user interface and file access permissions. It then denies the permissions in
the permission set.
...
public void OurCode(IUserCode code)
{
UIPermission uiPerm = new
UIPermission(PermissionState.Unrestricted);
FileIOPermission fileIOPerm = new
FileIOPermission(PermissionState.Unrestricted);
PermissionSet ps = new
PermissionSet(PermissionState.None);
ps.AddPermission(uiPerm);
ps.AddPermission(fileIOPerm);
ps.Deny();
...
The "third-party" code is then called. After it returns, the permission denial is revoked
and the "third-party" code is called again.
int v = code.PotentialRogueCode();
CodeAccessPermission.RevertDeny();
...
v = code.PotentialRogueCode();
The first time, the code execution fails; the second time it succeeds. Each stack frame can
only have one permission set for denial of permissions. If you call Deny on a permission
set, it overrides any other calls to Deny on a permission set in that stack frame.
Code Identity
The characteristics by which a particular assembly can be identified are its identity
permissions. An example would be an assembly's strong name or the Web site that
generated the code. Based on the evidence provided by the loader or trusted host, identity
permissions are granted by the CLR.
Identity Permission Classes
To identity running code, there are several identity permission classes.
• PublisherIdentityPermission models the software publisher's digital signature.

• SiteIdentityPermission models the Web site where code originated.
• StrongNameIdentityPermission models the strong name of an assembly.
• ZoneIdentityPermission models the zone where the code originated.
• URLIdentityPermission models the URL and the protocol where the code
originated.
These permissions represent evidence that can be used to determine security policy.
Identity permissions are not code access permissions.
Evidence
Security policy is based on a set of rules that administrators can set. The .NET security
system can use those rules to enforce the policy. The evidence, represented by the identity
permissions, is used to determine which policy to apply.
The AppDomain class has a function ExecuteAssembly which causes an assembly to run.
One argument to the method is an Evidence instance argument. This Evidence class is a
collection of objects that represent the identity of the assembly. This class is a collection
of objects that represent evidence.
The Evidence example illustrates this. This example gets the collection of evidence
associated with a strongly named assembly and prints out the associated values.
Evidence ev = AppDomain.CurrentDomain.Evidence;
IEnumerator iEnum = ev.GetEnumerator();
bool bNext;
Console.WriteLine("Evidence Enumerator has {0} members",

ev.Count);
bNext = iEnum.MoveNext();
while (bNext == true)
{
object x = iEnum.Current;
Type t = x.GetType();
Console.WriteLine(t.ToString());
if (t == typeof(System.Security.Policy.Zone))
{
Zone zone = x as Zone;
Console.WriteLine(" " +
zone.SecurityZone.ToString());
}
else if (t == typeof(System.Security.Policy.Url))
{
Url url = x as Url;
Console.WriteLine(" " + url.Value.ToString());
}
else if (t == typeof(System.Security.Policy.Hash))
{
Hash hash = x as Hash;
byte[] md5Hash = hash.MD5;
byte[] sha1Hash = hash.SHA1;
Console.WriteLine(" MD5 Hash of Assembly:");
Console.Write(" ");
for(int i = 0; i < md5Hash.Length; i++)
Console.Write(md5Hash[i]);
Console.WriteLine(" SHA1 Hash of Assembly:");
Console.Write(" ");
for(int i = 0; i < sha1Hash.Length; i++)
Console.Write(sha1Hash[i]);
}
else if (t == typeof(System.Security.Policy.StrongName))
{
StrongName sn = x as StrongName;
Console.WriteLine(" StrongName of Assembly is: {0}
version: {1}", sn.Name, sn.Version);
Console.WriteLine(" Assembly public key:");
Console.Write(" ");
Console.WriteLine(sn.PublicKey.ToString());
}
bNext = iEnum.MoveNext();
}
The example's output would look something like this:
Evidence Enumerator has 3 members

System.Security.Policy.Zone
MyComputer
System.Security.Policy.Url
file:///F:/Book/Chap12/Evidence/bin/Debug/Evidence.exe
System.Security.Policy.StrongName
StrongName of Assembly is: Evidence version: 1.0.0.0
Assembly public key:
0024000004800000940...
...D4E1C67A3509E6C9B385EA897BA
System.Security.Policy.Hash
MD5 Hash of Assembly:
14332230461041081341241322151846823019516744
SHA1 Hash of Assembly:
821331711844749119991571111431431822382322311431771
39171
The evidence associated with the Zone for this assembly is MyComputer. The Url
evidence is the location on disk of the assembly. The Hash evidence can give us the MD5
and SHA1 hashes of the assembly. The StrongName evidence tells us information about
the unique assembly name.
Some of this evidence is convertible to the associated identity permissions. For example,
the Zone class has a CreateIdentityPermission method which returns an IPermission
interface that represents the ZoneIdentityPermission instance associated with this piece of
evidence. The Url and StrongName classes have similar methods.
Another way of looking at the identity permissions is that they answer a series of
questions:
• Who published (signed) it?

• What is the name of the assembly?
• What Web site or URL did it come from?
• What zone did the code originate from?
The creator of the application domain (host) can also provide evidence by passing in an
Evidence collection when the ExecuteAssembly method is called. Of course, that code
must have the ControlEvidence permission. The CLR is also trusted to add evidence,
since after all, it enforces the security policy. Evidence is extensible; you can define
evidence types and use it in security policy.
Security Policy
Now that we understand evidence, and how the evidence about an assembly is gathered,
we can discuss security policy. Based on the evidence for an assembly, the assembly is
assigned to a code group. Associated with each code group is a set of permissions that
represent what code associated with that code group can do.
Security Policy Levels
Security policy is set at several levels. The permissions allowed are defined by the
intersection of the policy levels. These levels are enterprise, machine, application domain,
and user. If there is a conflict between permissions assigned from a particular level, the
more restrictive version overrides. So enterprise policy can override all the machines in
the enterprise, and machine policy can override all policies for a application domain or a
particular user.
Code Groups
The enterprise, machine, and user policy levels are a hierarchy of code groups.
Associated with each code group is a set of permissions. Code that meets a specified set
of conditions belongs to a particular code group.
The root node is referred to as "All_Code." Below this level is a set of child nodes, and
each of these children can have children. Each node represents a code group. If code
belongs to a code group, it might be a member of one of its children. If it does not belong
to a given code group it cannot belong to any of its children.
By evaluating the evidence you assign code a group. By assignment to a group you get an
associated set of permissions. This set of conditions corresponds to a named permission
set. Since code can belong to more than one group, the set of permissions which can be
granted to code is the union of all the permission sets from the all groups it belongs to.
Therefore code policy is determined in two steps. For each level, the permissions for an
assembly are determined by the union of all the permission sets to which it belongs. Each
level then effectively has one permission set. Then each of these permission sets is
intersected so that the most restrictive of each permission setting is the final value. For
example, if the machine level gives all access to an assembly, but the user level restricts
the file IO permissions to just read, the assembly will have unlimited permissions for
everything but file IO, where it will just have the read permission.
Code groups can have two attributes. The exclusive attribute dictates that code will never
be allowed more permissions than associated with the exclusive group. Obviously, code
can belong to only one group marked exclusive. The level final attribute indicates that no
policy levels below this one are considered when calculating code group membership.
The order of levels is enterprise, machine, user, application domain.
Named Permission Sets
A named permission set consists of one or more code access permissions that have a
name. An administrator can associate a code group with this permission set by means of
this name. More than one code group can be associated with a named permission set.
Administrators can define their own named permission sets, but several are built in:
• Nothing: no permissions (cannot run).

• Execution: only permission to run, but no permissions that allow use of protected
resources.
• Internet: the default policy permission set suitable for content from unknown
origin.
• LocalIntranet: the default policy permission set for within an enterprise.
• Everything: all standard (i.e., built-in) permissions except permission to skip
verification.
• FullTrust: full access to all resources protected by permissions.
Of the built-in named permission sets only the Everything set can be modified. You can
define custom permission sets.
Altering Security Policy
Security policy is stored in several XML-based configuration files. Machine security

configuration is in the security.config file that is stored in the
\WINNT\Microsoft.NET\Framework\vx.x.xxxx\CONFIG directory. User security
configuration is in the security.config file that is stored in the \Documents and
Settings\UserName\Application Data\Microsoft\CLR Security Config\vx.x.xxxx
directory.
It is not recommended that you edit these XML files directly. The Code Access Security
Policy tool (caspol.exe) is a command-line tool that can be used to modify enterprise,
machine, and user policy levels.
The .NET Admin Tool introduced in Chapter 7 provides a more friendly interface to
changing policy. Figure 12-17 shows the code groups and permission sets defined for the
machine and the current user security policy levels as they appear in the left pane in the
.NET Admin Tool.
Figure 12-17. Permission sets and groups for machine and user policy.
Let us use this tool to examine the current policies in the machine level. First let us look
at the named permission sets. As you can see from Figure 12-18, on the machine level no
new named permission sets have been created; only the default ones are present. If you
select the Internet permission set and in the right pane select view permissions, you can
then select any permission and look at its settings. Figure 12-18 shows the settings for
User Interface permission in the Internet named permission set.
Figure 12-18. Permissions for User Interface permission in machine-level

Internet named permission set.
Figure 12-19 shows the properties for the Internet Zone code group on the machine
policy level. You can see that Zone identity permission is chosen for this group, and the
value associated with it is the Internet zone. On the permission set tab, you can view or
select the named permission set associated with the Internet zone.
Figure 12-19. Properties dialog for Internet zone, machine policy level.
To illustrate how security policy affects running code we use a slightly modified version
of the Evidence example. Besides writing out the associated evidence, the Policy example
also prints out the contents of a file.

try
{
string fileWithFullPath = Path.GetFullPath(filename);
string text;

{
}
sr.Close();
}
catch(Exception e)
{
}
Figure 12-20 shows the two new code groups and the one permission set we will define at
the user policy level to control security policy for this assembly.
Figure 12-20. Revised user policy level for Policy example.
We will define a new permission set called TestStrongName and two new code groups,
TestStrongNameGroup and My_Computer_Zone. The new permission set definition is in
Figure 12-21.
Figure 12-21. TestStrongName permission set definition.

This new permission set is created by selecting the Permission Sets node below the level
in which you want to create it (in this case User). Right-mouse-click and select New. The
initial Create Permission Set dialog comes up and can be filled in as in Figure 12-22.
Figure 12-22. Initial Create Permission Set dialog.

Clicking the Next button brings up the dialog in Figure 12-23. Use the Add and Remove
buttons to define the permissions you want to include in this permission set.
Figure 12-23. PermissionSet definition dialog.

To define the permission itself, select the permission, click the properties dialog, and
make the appropriate choices. Figure 12-24 shows the dialog that appears for the User
Interface permission. To modify an existing permission set, select it, right-mouse-click,
and select the Change Permissions item. A dialog similar to Figure 12-24 will appear.
Figure 12-24. Permission modification dialog.

Now this permission has to be associated with a code group. How do assemblies get
assigned to code groups? We have already explained that each code group maps to one
piece of evidence. Figure 12-25 is a diagram of the User Code Level with its three
groups.
Figure 12-25. Diagram of user-level policy groups.
Figure 12-26 shows that the TestStrongNameGroup is defined to be the strong name
associated with the policy.exe assembly. Figure 12-27 shows the TestStrongName
permission set associated with the TestStrongNameGroup. Note that there is no
FileIOPermission. This code group was created by selecting the parent group (in this case
All_Code) and selecting New from its context menu and filling in the information asked
for by the wizard. Dialogs similar to Figures 12-26 and 12-27 will appear.
Figure 12-26. Membership condition for TestStrongNameGroup.
Figure 12-27. Permissions associated with TestStrongNameGroup from

TestStrongName permission set.
The My_Computer_Zone group is defined to encompass all code on this computer. It is
defined in a similar fashion as the TestStrongNameGroup. The membership condition is
Zone, and the MyComputer zone is picked as the associated value. FullTrust is selected
for its associated permission set. The All_Code group encompasses all code on the
machine. It grants no rights to any code. It already existed by default, and we changed its
permission set to Nothing. It grants no rights.
To find out how an assembly matches the code groups, its evidence its compared with the
membership conditions for the group. All code that resides on the current machine (as
opposed to another machine on the network or the Internet) matches the All_Code and
My_Computer_Zone group. Only policy.exe matches the membership condition for the
TestStrongNameGroup. The tree is walked from parent to child node; if a parent node
does not match, no further navigation down the tree is done. On a given level the rights
assigned to the assembly are the union of all the groups that it matches. In this case, even
though policy.exe matches a group that does not give it the FileIOPermission, it gets that
permission from the My_Computer_Zone group which grants FullTrust to code.
A similar analysis of the enterprise and machine levels reveals that they also grant code
from this machine FullTrust. So if you run policy.exe, it will run.
Now modify the TestStrongNameGroup on its General tab to be exclusive, as indicated in
Figure 12-28. This will cause any code that belongs to this group to get its rights from
only this group. Since policy is determined by the intersection of all the three levels,
policy.exe will not have the FileIOPermission. If you try to run it, you will see that it
cannot read the file.
Figure 12-28. Making TestStrongNameGroup.
Summary
In this chapter we have attempted to explain the basics of .NET security. .NET security
comes in two flavors: user identity security and code access security. The former
determines which identity is executing code. The latter determines what rights the
executing code has. Between the two you have the basic tools to provide robust
applications.
What is missing from .NET security right now is distributed identity and distributed code
access security. Remote code cannot use policy information to make decisions, and
identity is not automatically transferred through remote calls.
Chapter 13. Tracing and Debugging in .NET Top
Complicated applications cannot be put under the debugger to find out what went wrong.
Duplicating, or even understanding, what conditions are needed to replicate the problem
is often difficult. The System.Diagnostics namespace has several classes that help you
instrument your application. [1]
[1]
The security of your Web Site or Web Service is enhanced by using tracing and
debugging output. You do not want to give out information in an error message that could
be used to compromise your system. Capturing that information in a trace or debug log
allows the program to generate a generic error message for the user. You could also
assign an identifier to the user message that is also recorded with the log message. If
necessary that id could be used to help the user diagnose any problems with their system.
Instrumenting your application for debugging and tracing will enable you to make your
applications more robust. It also illustrates the common pattern of how the framework
divides classes into separate tasks (writing the output, controlling the output, and the
output destination) so that you can customize parts and still rely on the Framework
classes for the rest. The mechanics of instrumenting your application has three aspects.
The Trace and Debug classes are used to generate the debug or trace output. These
classes have identical methods and properties that allow you to write diagnostic output.
They do not, however, specify the destination of the output.
The Listeners classes are used to direct the output to various destinations, although a
default destination does exist.
Finally, there are mechanisms for turning on or off the instrumentation. You can set the
DEBUG and TRACE compilation flags to have different tracing for debug or release
builds. You can have the output of the Trace and Debug classes depend on the conditional
evaluation of expressions. Or you can control the verbosity of the output, depending on
your need for information, using the BooleanSwitch and TraceSwitch classes.
The TraceDemo Example
The TraceDemo example illustrates the use of the diagnostic functionality. If you run the
example, you will get the following output:
Trace Listeners:
Default
This was compiled with a DEBUG directive!

This was compiled with a TRACE directive!
Debug Boolean Switch disabled at startup.
Debug Boolean Switch enabled!
Trace Switch Startup Value = Warning
TraceError!
TraceWarning!
Trace Listeners:
Console.Out Listener
Output File Listener
Refer to this output in the ensuing discussion. You will also find a file called output.txt on
your computer in the directory where this program ran.
Enabling Debug and Trace Output
To use the Debug class, the DEBUG flag must be defined or else the methods of this
class will not be compiled into the executable or library. Similarly, to use the Trace class
the TRACE flag must be defined. This way you can have different diagnostics for release
and debug builds. These constants can be set in the Visual Studio.NET Project |
Properties | Configuration Properties | Build Window's conditional compilation constants
shown in Figure 13-1.
Figure 13-1. Visual Studio window for setting conditional compilation

constants.
You can also define the constants in your source files or supply the definition to the
compiler's command line.
Using the Debug and Trace Classes
The useful methods and properties are static. The overloaded WriteLine and Write are
used to write debug or trace output. The overloaded WriteLineIf and WriteIf write output
if the condition in their first argument is true.
Debug.WriteLine("This was compiled with a DEBUG

directive!");
Trace.WriteLine("This was compiled with a TRACE
directive!");
...
Debug.WriteLineIf(DebugBooleanSwitch.Enabled, "Debug
Boolean Switch enabled at startup.");
Debug.WriteLineIf(!DebugBooleanSwitch.Enabled,
"Debug Boolean Switch disabled at startup.");
Output is indented with the Indent and Unindent methods. The indentation size is
controlled with the IndentSize property.
Trace.Indent();
...
Trace.IndentSize = 10;
You can also set the indentation size in the application configuration file.
<?xml version="1.0"?>
<configuration>
<system.diagnostics>
<trace indentsize="15" />
</system.diagnostics>
</configuration>
The Assert method can check an assertion. The AutoFlush property and the Flush method
control the flushing of the output buffer.
Using Switches to Enable Diagnostics
Switches give you finer grain control over the diagnostic output. You can use the
BooleanSwitch class to turn output on or off based on the value of its Enabled property.
The TraceSwitch class gives you five hierarchical levels of control for its Level property:
TraceError, TraceWarning, TraceInfo, TraceVerbose, and Off. These values are part of the
TraceLevelEnumeration. Setting a lower Trace level means that the higher ones are set as
well. For example, if the TraceWarning level is set, both the TraceError and
TraceWarning levels are enabled.
DebugBooleanSwitch.Enabled = true;
Debug.WriteLineIf(DebugBooleanSwitch.Enabled, "Debug
Boolean Switch enabled!");
...
Trace.WriteLineIf(TraceLevelSwitch.TraceError,
"TraceError!");
The constructors for these switches take two parameters. The first is the name of the
switch, the second is a text description of the switch. Both BooleanSwitch and
TraceSwitch classes inherit from the abstract class Switch. You can write your own
customized switch classes by inheriting from the Switch class. Note that the Enabled
property of the BooleanSwitch and the Level and named level properties of the
TraceSwitch are not part of the Switch class.
Enabling or Disabling Switches
You can use settings in your application configuration file to enable or disable a switch at
startup. This can also be done programmatically.
Configuration File Switch Settings
You can set the switch's initial setting in the application's configuration file.
<configuration>
<system.diagnostics
<switches>
<add name="DebugSwitch" value = "0" />
<add name="TraceSwitch" value = "2" />
</switches>
</system.diagnostics>
</ configuation>
If no values are found, the initial value of the Enabled property of the BooleanSwitch
with the name DebugSwitch is set to false and the TraceSwitch's Level property is set to
TraceOff.
Programmatic Switch Settings
The Enabled property of the BooleanSwitch can be set to true or false. The Level
property of the TraceSwitch can be set to one of the options of the TraceLevel
enumeration: TraceOff, TraceError, TraceWarning, TraceInfo, TraceVerbose. You can get
the level of the TraceSwitch's setting by examining the TraceError, TraceWarning,
TraceInfo, TraceVerbose properties.
Using Switches to Control Output
You can test the value of the switch before you write, debug, or trace output. You can do
this with an if statement, or as an argument to one of the Trace or Debug classes'
methods.
Trace.WriteLineIf(TraceLevelSwitch.TraceError,
"TraceError!");
Trace.WriteLineIf(TraceLevelSwitch.TraceWarning,
"TraceWarning!");
Trace.WriteLineIf(TraceLevelSwitch.TraceInfo,
"InfoMessage!");
Trace.WriteLineIf(TraceLevelSwitch.TraceVerbose,
"VerboseMessage!");
Since you can set these values outside of your program's code, you can select the
circumstances under which you get a particular level of debug or trace output. For
example, you can turn on TraceVerbose output if you really need a high level of
diagnostics, but turn it off after you have found the problem.
TraceListener
Classes derived from the abstract class TraceListener represent destinations for the
diagnostic output. The TextWriterTraceListener is designed to direct output to a
TextWriter, Stream, or FileStream. Console.Out is an example of a commonly used
output stream. The EventLogTraceListener class allows you to send output to an
EventLog. You can create your own event logs with the EventLog's static method
CreateEventSource method. The DefaultTraceListener sends output to the debugging
output window. Default Debug output can be viewed in Visual Studio.NET's Output
window or with utilities (such as DBMon, which is included with this project). You can
customize where output appears by implementing your own class derived from
TraceListener.
Listeners Collection
Both the Debug and Trace classes have a static Listeners collection. This collection of
TraceListeners represents a list of TraceListener objects that want to receive the output
from the Debug or Trace class. Listeners are added to or removed from the collection just
as with any other .NET collection.
TextWriterTraceListener ConsoleOutput = new

TextWriterTraceListener(Console.Out,
"Console.Out Listener");
Trace.Listeners.Add(ConsoleOutput);
Stream OutputFile = File.Create("output.txt");

TextWriterTraceListener OutputFileListener = new
TextWriterTraceListener(OutputFile,
"Output File Listener");
Trace.Listeners.Add(OutputFileListener);
Trace.Listeners.Remove("Default");
In this code extract, the OutputFileListener in the example will send the Trace output to a
file called output.txt. The DefaultTraceListener is added automatically to the Listener
collections. Any of the listeners, including the default listener, can be removed from the
collection by invoking the collection's Remove method. To list all listeners in the
collection:
foreach(TraceListener tr in Trace.Listeners)
{
Console.WriteLine("\t" + tr.Name);
}
Summary
Instrumenting your application for degrees of debugging and diagnostic output is a

common program task. The diagnostic classes exemplify the way .NET provides classes
to handle standard programming tasks so you can concentrate on developing the business
logic of your programming, not on building infrastructure. On the other hand, they also
exemplify how the .NET classes are partitioned so that you can customize the
infrastructure using as much or as little of the other classes as you require.
Chapter 14. Interoperability Top
Microsoft .NET is a powerful platform, and there are many advantages in writing a new
application within the .NET Framework. However, a typical application is not a world
unto itself, but is built from legacy components as well as new components, and
interoperability is very important. We discussed one kind of interoperability in Chapter 11
in connection with Web Services. Using the SOAP protocol it is possible for .NET
applications to call Web Services on other platforms, including Unix, mainframes, and
mobile devices.
In this chapter we will look at another kind of interoperability, the interfacing of managed
and unmanaged code running under Windows. The dominant programming model in
modern Windows systems is the Component Object Model, or COM. There exist a great
many legacy COM components, and so it is desirable for a .NET program, running as
managed code, to be able to call unmanaged COM components. The converse situation,
in which a COM client needs to call a .NET server, can also arise. [1] Apart from COM,
we may also have need for a .NET program to call any unmanaged code that is exposed
as a DLL, including the Win32 API. The .NET Framework supports all these
interoperability scenarios through COM Interoperability and the Platform Invocation
Services or PInvoke.
[1]
COM interop is the only mechanism provided for unmanaged code to call managed
code.
In this chapter we assume that you understand the concepts behind the legacy
technologies.
Calling COM Components from Managed Code
The first interoperability scenario we will look at is managed code calling COM
components. The .NET Framework makes it easy to create a Runtime Callable Wrapper
(RCW), which acts as a bridge between managed and unmanaged code. The RCW is
Figure 14-1. A Runtime Callable Wrapper between managed and unmanaged

code.
You could implement an RCW assembly yourself, using the PInvoke facility (described
in a later section) to call into the necessary APIs, such as CoCreateInstance and the
IUnknown methods directly. But that is not necessary, because the Tlbimp.exe tool can
read type library information, and automatically generate the appropriate RCW for you.
Visual Studio.NET makes it even easier when you add a reference to a COM object in
Solution Explorer. We will examine both of these facilities, as we look at some examples
of COM components and .NET clients.
The Tlbimp.exe Utility
The Tlbimp.exe utility (Type Library to .NET Assembly Converter) program is provided
in the \Program Files\Microsoft.NET\FrameworkSDK\Bin directory. It is used to
generate managed classes that wrap unmanaged COM classes. The resulting RCW is a
.NET component (i.e., a managed DLL assembly) that managed client code can use to
access the COM interface methods that are implemented in the COM component. The
Tlbimp tool is a command line program that reads COM type library information, and
generates a managed wrapper class along with the associated metadata, and places the
result into the RCW assembly. You can view the resulting contents in this assembly using
the Ildasm tool. The command line syntax for Tlbimp is shown below.
Tlbimp TypeLibName [options]

Where options may contain the following:
/out:FileName Assembly file name
/namespace:Namespace Assembly Namespace
/asmversion:Version Assembly version number
/reference:FileName Reference assembly
/publickey:FileName Public key file
/keyfile:FileName Key pair file
/keycontainer:FileName Key pair key container
/delaysign Delay digital signing
/unsafe Suppress security checks
/nologo Suppress displaying logo
/silent Suppress output except errors
/verbose Display extra information
/primary Make primary interop assembly
/sysarray SAFEARRAY as System.Array
/strictref Only /reference assemblies
/? or /help Display help information
When the Tlbimp tool imports a COM type library, it creates a .NET namespace with the
same name as the library defined in the type library (that is the name of the actual library,
not the name of the type library file that contains it). Tlbimp converts each COM coclass
defined in the type library into a managed .NET wrapper class in the resulting .NET
assembly that has one constructor with no parameters. Each COM interface defined in the
type library is converted into a .NET interface in the resulting .NET assembly.
Consider the typical COM IDL file library statement shown below that would be used to
create a type library using Midl.exe. The resulting type library (TLB) or DLL file would
cause Tlbimp.exe to generate an assembly containing metadata, including the namespace
BANKDUALLib, a managed wrapper class named Account2, and a managed interface
named IAccount2.
library BANKDUALLib
{
importlib("stdole32.tlb");
importlib("stdole2.tlb");
[
uuid(04519632-39C5-4A7E-AA3C-3A7D814AC91C),
helpstring("Account2 Class")
]
coclass Account2
{
[default] interface IAccount2;
};
};
Once you have used Tlbimp.exe to generate the wrapper assembly, you can view its
contents using the Ildasm tool, as shown in Figure 14-2. Note that the namespace shown
by Ildasm.exe is BANKDUALLib, the name of the interface is IAccount2, and the
wrapper class is named Account2.
Figure 14-2. Ildasm.exe showing contents of a COM wrapper assembly.
Demonstration: Wrapping a Legacy COM Server
The best way to get a feel for how this wrapping process works is to perform the
operations yourself. The .NET client program is in the directory NetClient. The directory
LegacyComServer contains the following files:
BankDual.dll COM server DLL

BankDual.tlb Type library
reg_bankdual.bat Batch file to register the server
unreg_bankdual.bat Batch file to unregister the server
BankConsole.exe Client executable file
The source code for the client and server are in the directories ClientSource and
ServerSource respectively. Both programs are written in Visual C++, and project files are
provided for Visual C++ 6.0. Unless you have Visual C++ 6.0 installed on your system in
addition to Visual Studio.NET, you will not be able to build these projects, but that will
not prevent you from running the program and creating an .NET client.
This COM server implements a simple bank account class that has Deposit and Withdraw
methods and a Balance property. The simple code [2] is shown in Account2.cpp in the
ServerSource directory.
[2]
We will not discuss the somewhat intricate infrastructure code provided by this
ATLbased COM server. Such "plumbing" is much easier with .NET. Our focus is on
calling COM components, not implementing them.
STDMETHODIMP CAccount2::get_Balance(long *pVal)

{
*pVal = m_nBalance;
return S_OK;
}
STDMETHODIMP CAccount2::Deposit(long amount)
{
m_nBalance += amount;
return S_OK;
}
STDMETHODIMP CAccount2::Withdraw(long amount)
{
m_nBalance -= amount;
return S_OK;
}
Register the COM Server
The first step is to register the COM server. You can do that by running the batch file
reg_bankdual.bat, which executes the command,
regsvr32 bankdual.dll
You can now see the registration entries using the Registry Editor (regedit.exe) or the
OLE/COM Object Viewer (oleview.exe). The latter program is provided on the Tools
menu of Visual Studio.NET. It groups related registry entries together, providing a
convenient display. You can also perform other operations, such as instantiating objects.
Figure 14-3 shows the entries for the Account2 class that is implemented by this server.
We have clicked the little "+" in the left-hand pane, which instantiates an object and
queries for the standard interfaces. You can release the object by right-clicking over the
class and choosing Release Instance from the context menu.
Figure 14-3. OLE/COM Object Viewer showing registry entries.

Run the COM Client
You can now run the COM client by double-clicking on BankConsole.exe in Windows
Explorer. The starting balance is shown, followed by a withdrawal of 25, and the balance
is shown again. Here is the source code, in the file BankConsole.cpp in ClientSource:
// BankConsole.cpp
#include <stdio.h>
#include <stdlib.h>
#include <objbase.h>
#include "bankdual.h"
#include "bankdual_i.c"
IAccount2* g_pAccount;
void ShowBalance()
{
long balance;
HRESULT hr = g_pAccount->get_Balance(&balance);
printf("balance = %d\n", balance);
}
int main(int argc, char* argv[])

{
// Initialize COM
HRESULT hr = CoInitializeEx(NULL,
COINIT_APARTMENTTHREADED);
// Instantiate Account object, obtaining interface
// pointer
hr = CoCreateInstance(CLSID_Account2, NULL,
CLSCTX_SERVER, IID_IAccount2, (void **) &g_pAccount);
// First obtain and display initial balance
ShowBalance();
// Deposit 25 and show balance
hr = g_pAccount->Deposit(25);
ShowBalance();
// Clean up
g_pAccount->Release();
CoUninitialize();
printf("Press enter to quit: ");
char buf[10];
gets(buf);
return 0;
}
For simplicity, no error checking is done. Robust code should check the HRESULT that is
returned from each of the COM calls. Here is the output from running the client program:
balance = 150
balance = 125
Press Enter to quit:
Import the Type Library (TlbImp.exe)
In order to call the COM component from managed code, we must create an RCW. We
can do that by running the TlbImp.exe utility that we have discussed. We will run this
utility from the command line, in the directory NetClient, where we want the RCW
assembly to wind up. We provide a relative path to the type library file [3] BankDual.tlb in
the directory LegacyComServer. What we have to type is shown in bold.
[3]
The file BankDual.dll also contains the type library and could have been used in place
of BankDual.tlb.
tlbimp ..\legacycomserver\bankdual.tlb
TlbImp - Type Library to .NET Assembly Converter Version
1.0.2914.16
Copyright (C) Microsoft Corp. 2001. All rights reserved.
Type library imported to BANKDUALLib.dll
The RCW assembly that is created is BANKDUALLib.dll, taking its name from the name
of the type library, as discussed earlier.
Implement the .NET Client Program

It is now easy to implement the .NET client program. The code is in the file NetClient.cs
in the directory NetClient.
// NetClient.cs
using System;
using BANKDUALLib;
class NetClient
{
{
Account2 acc;
acc = new Account2();
acc.Withdraw(25);
}
}
As with the COM client program, for simplicity we do no error checking. In the .NET
version we should use exception handling to check for errors. The RCW uses the
namespace BANKDUALLib, based on the name of the type library.
You must add a reference to BANKDUALLib.dll. In the Visual Studio Solution Explorer
you can right-click over References, choose "Add Reference," and use the ordinary .NET
tab of the Add Reference dialog.
Build and run the project inside of Visual Studio. You should see the following output:
balance = 150
balance = 125
Once you have added a reference to a RCW, you have all the features of the IDE
available for .NET assemblies, including Intellisense and the Object Browser. You can
bring up the Object Browser from View | Other Windows | Object Browser. Figure 14-4
illustrates the information shown.
Figure 14-4. Object Browser showing information about the RCW.

Import a Type Library Using Visual Studio
When you are using Visual Studio you can import a COM type library directly, without
first running TlbImp.exe. To see how to do this, use Solution Explorer to delete the
reference to BANKDUALLib.dll. In fact, delete the file itself, and delete the bin and obj
directories of NetClient. Now right-click over References, choose "Add Reference," and
this time select the COM tab from the Add Reference dialog. The listbox will show all the
COM components with a registered type library. Select "BankDual 1.0 Type Library," as
Figure 14-5. Add a reference to a COM component in Visual Studio.

Now click OK. You will see a message telling you that a "primary interop assembly" is
not registered for this type library. You will be invited to have a wrapper generated for
you, as illustrated in Figure 14-6. Click "Yes." The generated RCW is the file
Interop.BANKDUALLib_1_0.dll in the directory bin\Debug. You should be able to build
and run the .NET client program.
Figure 14-6. Visual Studio will create a primary interop assembly.
The primary interop assembly that was created by Visual Studio is normally created by
the publisher of the COM component. This can be done using the TlbImp.exe utility with
the /primary option.
Wrapping a COM Component with a Pure Vtable Interface

Dual Interfaces
Our example legacy COM component BankDual.dll had a dual interface IAccount2. This
means that the interface could be called by both an early-binding COM client using the
vtable and also by a late-binding client using IDispatch. The IDL file BankDual.idl
specifies the interface IAccount2 as dual.
[
object,
uuid(AAA19CDE-C091-47BF-8C96-C80A00989796),
dual,
helpstring("IAccount2 Interface"),
pointer_default(unique)
]
interface IAccount2 : IDispatch
{
[propget, id(1), helpstring("property Balance")] HRESULT
Balance([out, retval] long *pVal);
[id(2), helpstring("method Deposit")] HRESULT
Deposit([in] long amount);
[id(3), helpstring("method Withdraw")] HRESULT
Withdraw([in] long amount);
};
An example of late-binding is VBSCript code for client-side scripting on a Web page.

The directory BankHtml contains the file Bank.htm with an HTML form and VBScript
code to exercise our bank account server.


<HTML>
<HEAD>
<TITLE>Bank test page for Account object</TITLE>
<SCRIPT LANGUAGE="VBScript">

</SCRIPT>
<FORM NAME = "Form1" >

Amount <INPUT NAME="txtAmount" VALUE="" SIZE=8>
<P>
Balance <INPUT NAME="txtBalance" VALUE="" SIZE=8>
<P>
<INPUT NAME="btnCreate" TYPE=BUTTON VALUE="Create">
<INPUT NAME="btnDestroy" TYPE=BUTTON VALUE="Destroy">
<INPUT NAME="btnDeposit" TYPE=BUTTON VALUE="Deposit">
<INPUT NAME="btnWithdraw" TYPE=BUTTON VALUE="Withdraw">
</FORM>
</BODY>
</HTML>
The createobject function instantiates a COM object using late binding, referencing a
program ID rather than a CLSID. This is perfectly legitimate, because BankDual.dll
implements a dual interface on the Account2 object. Since this is client-side script, we
can exercise it locally in Internet Explorer, simply double-clicking on bank.htm in
Windows Explorer. This will bring up Internet Explorer and show the form. You can click
the Create button and instantiate an object, [4] as shown in Figure 14-7. The starting
balance of 150 is shown. You can then exercise Deposit and Withdraw, and when you are
done, you can click Destroy.
[4]
Depending on your security settings, you may get a warning message about an ActiveX
control on the page. Click Yes to allow the interaction. If you have trouble running the
ActiveX control at all, check your security settings in Internet Explorer.
Figure 14-7. Accessing a late-bound COM object in Internet Explorer.

Pure Vtable Interface
Dual interfaces are very common. The default in an ATL wizard generated COM
component is dual interface. Visual Basic 6.0 also creates COM components with dual
interfaces. However, if there is no occasion for a COM component to be called by a late-
binding client, it is more efficient to implement only a pure vtable interface.
There is a slight issue in generating wrappers for COM components with a pure vtable
interface. To see the problem, consider the COM component in VtableComServer. As
with our LegacyComServer example, the top-level directory contains the DLL, the type
library file, batch files to register and unregister the server, and a client test program.
Source code for the COM server and client is provided in ServerSource and ClientSource
respectively. We want to implement a managed client program VtableNetClient.
First, verify that the COM client and server work. All you have to do is run the batch file
reg_bank.bat to register the server, and you can double-click on BankConsole.exe in
Windows Explorer to run the client.
Next, open up the solution VtableNetClient.sln in Visual Studio. Add a reference to the
COM type library "Bank 1.0 Type Library." You should get a clean build. But when you
run the program, you get an exception:
Unhandled Exception: System.InvalidCastException:
QueryInterface for interface BANKLib.IAccount failed.

at BANKLib.Account.GetBalance(Int32& pBalance)
at VtableNetClient.ShowBalance() in
C:\OI\NetCs\Chap14\VtableNetClient\Vtable
NetClient.cs:line 14
at VtableNetClient.Main() in
C:\OI\NetCs\Chap14\VtableNetClient\
VtableNetClient.cs:line 33
The problem is that the .NET client is in a separate apartment, and it needs marshaling.
You can use any of the following solutions:
1. Mark the IDL for the interface as dual.

2. Mark the IDL for the interface as oleautomation, and limit types used to
oleautomation friendly types.
3. Build and register the proxy/stub DLL for the interface.
4. Mark the Main method in the C# client with the [STAThread] or [MTAThread]
attribute (appropriate to the situation), to place it into the same threading model as
the COM server.
Examining the source code for VtableNetClient.cs, we see that we commented out the
attribute [STAThread] in front of Main. Uncomment, build, and run again. This time it
should work!
As an alternate solution, comment out [STAThread] again. Now in the server directory
VtableComServer run the batch file reg_bankps.bat to register the proxy/stub DLL. Build
and run the .NET client. Again, it should work!
Notice another feature of this .NET client program. Rather than calling methods on a
class object, we go through interface references. We obtain the interface references using
the C# as operator, as we discussed in Chapter 5. This use of the as operator is the analog
in .NET of QueryInterface in COM.
Calling Managed Components from COM Client
Obviously, it is much more likely that you will want to write new .NET applications that
make use of legacy COM components, however, there may be times when you need to go
in the opposite direction. For example, you may have an existing application that makes
use of one or more COM components, and you would like to eventually rewrite several of
those COM components as .NET components. However, in the mean time, you may want
to make use of those new .NET components in your existing COM client applications as
well.
COM client programs may use early binding (vtable interface) or late binding (IDispatch
interface) to access managed .NET components. Early binding requires that type library
information is available at compile time. Late binding does not require any type library
information at compile time, since binding takes place at runtime via the IDispatch
interface methods.
However, regardless of whether the client uses early or late binding, a bridge is required
between the unmanaged native execution environment of the COM client and the
managed execution environment of the .NET component. This bridge is known as the
COM Callable Wrapper (CCW), which acts as a proxy for the managed object as shown
in Figure 14-8. Only one CCW object is created for any given managed object created for
a COM client. The CCW manages object lifetime according to the reference counting
rules of IUnknown, and it also manages marshaling for the method calls made on the
object.
Figure 14-8. A COM callable wrapper between unmanaged and managed code.
While the RCW assembly is explicitly created as a file, the CCW is created dynamically
at runtime by the Common Language Runtime. The CLR creates exactly one CCW for a
managed object, regardless of the number of COM clients that request its services, and
both COM and .NET clients can make requests on the same .NET object simultaneously.
A Late Binding COM Client
There are many variations of a COM client calling an .NET component. We will illustrate
with just one scenario, a late binding COM client calling a managed component. We will
create an .NET component that can be called through VBScript on the bank.htm Web
page.
Looking at the VBScript code used in createobject on bank.htm, we see that the ProgId of
the COM object is "BankDual.Account2.1." We wish to create an .NET object that can be
used in place of this COM object, and that has the same ProgId. To avoid confusion with
the COM object, unregister it running the batch file unreg_bankdual.bat in the directory
LegacyComServer. Now if you access bank.htm in Internet Explorer and click the
"Create" button, you will get an error.
C# code for implementing a compatible bank account object is in the file Account.cs in
the directory NetServer.
// Account.cs
using System;
namespace NetServer
{
[ProgId("BankDual.Account2.1")]
{
static private int balance;
public Account()
{
balance = 1000;
}
public void Deposit(int amount)
{
balance += amount;
}
public void Withdraw(int amount)
{
balance -= amount;
}
public int Balance
{
get
{
return balance;
}
}
}
}
The code shown in bold enables us to assign "BankDual.Account2.1" as the ProgId,

making it compatible with the COM object we are replacing. If we left these lines out, we
would still be able to call the object through COM. The ProgId would be created from the
namespace and the class name, or "NetServer.Account." Other attributes would let us
assign various GUIDs, which would be useful in an early binding scenario. Note that to
distinguish our .NET component from the COM component it is replacing, we have
assigned the starting balance to be 1000.
We are going to deploy our component in the Global Assembly Cache, so we need to
create a strong name, as discussed in Chapter 7. We generate a public-private key pair
and place them in a file keypair.snk, using the command,
sn -k keypair.snk
In our Visual Studio project we reference this key file in AssemblyInfo.cs,
[assembly: AssemblyDelaySign(false)]
[assembly: AssemblyKeyFile("keypair.snk")]
[assembly: AssemblyKeyName("")]
Our project creates the target assembly NetServer.dll in the top-level source directory,
where we also have the keypair.snk file. We can run all the command-line programs from
the directory c:\OI\NetCs\Chap14\NetServer. We can then place our assembly in the
GAC using the command,
gacutil -i netserver.dll
You can use the .NET Admin Tool discussed in Chapter 7 to inspect the contents of the
GAC, verifying that NetServer has indeed been deployed there. See Figure 14-9.
Figure 14-9. Inspecting the GAC using the .NET Admin Tool.
In order to make our .NET component available to COM clients, we must provide
suitable entries in the Registry. This will enable the COM runtime to locate the
appropriate server path and so on. The Assembly Registration Utility, Regasm.exe, reads
the metadata within an assembly and adds these necessary entries to the Registry, which
allows COM clients to use the .NET assembly's components as if they were just old-
fashioned registered COM components (via the CCW proxy).
The syntax for using Regasm.exe is shown next. This allows COM client programs to
create instances of managed classes defined by in the assembly.
Regasm AssemblyPath [options]

Where the options may be any of the following.
/unregister Unregister types
/tlb[:FileName] Specified typelib
/regfile[:FileName] Specified output reg file name
/codebase Sets the code base in the registry
/registered Only refer to preregistered typelibs
/nologo Prevents displaying logo
/silent Prevents displaying of messages
/verbose Displays extra information
/? or /help Display usage help message
We run this utility on NetServer using the command shown in bold,
C:\OI\NetCs\Chap14\NetServer>regasm netserver.dll
RegAsm - .NET Assembly Registration Utility Version
1.0.2914.16
Copyright (C) Microsoft Corp. 2001. All rights reserved.
Types registered successfully
We can use the OLE/COM Object Viewer to inspect the entries made in the Registry.
Note that there is a special category of COM objects called ".NET Category." Figure 14-
10 shows the Registry entries for our "NetServer.Account" object. Note that the ProgId is
"BankDual.Account2.1," as specified by the attribute in our C# source code. Note also
that the InprocServer32 is mscoree.dll, which is the DLL implementing the CLR. As
previously mentioned, there is no file created for the CCW. Instead, when the wrapped
component is to be instantiated, the CLR creates the CCW on the fly.
Figure 14-10. OLE/COM Object Viewer shows Registry entries for an .NET
object.
A late-binding COM client can now call our .NET component. That is all there is to it!
You can double-click on bank.htm, and Internet Explorer will run the VBScript we
looked at before. Only this time, the .NET component NetServer.Account is invoked, as
you can tell by noticing that the starting balance is 1000, as shown in Figure 14-11.
Figure 14-11. Accessing an .NET object in Internet Explorer.
Platform Invocation Services (PInvoke)
Platform Invocation Services, also known as "PInvoke," makes unmanaged exported

DLL functions available to managed client code. PInvoke allows this to be done from
managed code written in any .NET programming language. Notice that PInvoke is not the
name of a class, or a method, but is just a nickname for Platform Invocation Services.
PInvoke allows marshaling between CLR data types and native data types, and bridges
other differences between the managed and unmanaged runtime environments. Although
PInvoke is primarily used to access the Win32 APIs, it can be used to call into your own
legacy DLLs that you may find are still useful. Unfortunately, PInvoke is in most
circumstances a one-way street. You can use it to call from managed code into
unmanaged DLL code and of course return back into managed code. PInvoke is used to
access global exported DLL functions, so even though it is possible for DLLs to export
class methods, they are currently not accessible via PInvoke.
If you are an experienced Windows programmer and have a good knowledge of the
Win32 API, you may be tempted, after learning about PInvoke, to call a familiar Win32
API function to perform a task. A secure .NET environment, however, will not give most
assemblies permission to call unmanaged code. Usually there will be a native .NET
Framework class method that can accomplish your aim, and you should use .NET
Framework classes whenever possible. Occasionally it will be necessary to drop down to
the underlying platform, and then PInvoke is invaluable.
A Simple Example
Let's begin with a very simple example of the use of PInvoke, to call the Windows
MessageBox function. Our sample program is in the directory SimplePInvoke.
// SimplePInvoke.cs
using System;
class SimplePInvoke
{
[DllImport("user32.dll", EntryPoint="MessageBoxA")]
public static extern int ShowMessage(int hWnd,
string text, string caption, int type);

{
ShowMessage(0, "Hello, World", "From PInvoke", 0);
}
}
The key step is to place a DllImport attribute before the prototype of the function we
want to call. The function must take ordinary C# data types as parameters, which have
natural mappings to the C data types of the native function. The function will be treated
as a static method in the class where it is defined. The one required parameter to the
DllImport attribute is the name of the DLL exporting the function. There are various
optional, named parameters, that can be used with DllImport. For a complete list, consult
the documentation of the DllImportAttribut class in the System.Runtime.InteropServices
namespace. In our example, we use the EntryPoint attribute to specify the name by which
the function is exported in the DLL. The name of the static method in the class can then
be different, and will be the name to be used in the C# code that calls the method. In our
example, the Win32 function has the name MessageBoxA and our C# code calls the
method under the name ShowMessage. Figure 14-12 shows the output from this little
program.
Figure 14-12. Calling the Win32 MessageBox function through PInvoke.

Marshaling out Parameters
The previous PInvoke example did not demonstrate how PInvoke automatically marshals
out parameters for you where there is a clear mapping between Win32 and the CLR
types. This is because the MessageBox takes only in parameters. The next example calls
the GetComputerName and GetLastError APIs via PInvoke. The code for this example is
in the directory PInvoke.
// PInvoke.cs
using System;
using System.Text;
public class Test

{
[DllImport("kernel32.dll", CharSet=CharSet.Ansi)]
public static extern bool GetComputerName(
StringBuilder name, out uint buffer);
[DllImport("kernel32.dll")]
public static extern uint GetLastError();

{
bool result = true;
uint error = 0;
StringBuilder name = new StringBuilder(128);
uint length = 128;
result = GetComputerName(name, out length);
if (result == true)
else
{
error = GetLastError();
Console.WriteLine("Error: {0:x}", error);
}
return 0;
}
}
Translating Types
Since GetComputerName returns a name, StringBuilder was used instead of string. [5] For
input only arguments you can use string. An out attribute was placed on the length
attribute because the second argument to GetComputerName is a pointer. Unsigned types
were used because DWORD is an unsigned 32 bit quantity. For comparison, here are the
prototypes of the corresponding Win32 functions:
[5]
Instances of string are immutable, so we use the StringBuilder class, which was
discussed in Chapter 3.
BOOL GetComputerName(
LPTSTR lpBuffer, // computer name
LPDWORD lpnSize // size of name buffer
);
DWORD GetLastError(VOID);
Some CLR types do not map directly into unmanaged types. You have to tell the
Execution Engine (mscoree.dll) how to translate to a BSTR. You do that by annotating
the declaration with the MarshalAs attribute:
[MarshalAs(UnmanagedType.BStr)] public string foo
The UnmanagedType enumeration lists all the translatable types.
Summary
In this chapter we studied mixing managed and unmanaged code running under
Windows. We saw how to call legacy COM components from within the managed .NET
environment using a Runtime Callable Wrapper or RCW. We also looked at the use of a
COM Callable Wrapper (CCW) to enable a COM client to call a .NET component.
Finally, we looked at using Platform Invocation Services (PInvoke), and saw how
automatic marshaling is provided for both in and out parameters.
We have come to the end of a long journey, which we hope will be the first of many
journeys in the world of .NET. We hope you enjoyed the trip. Good luck on your .NET
programming projects!
Appendix A. Visual Studio.NET Top
Although it is possible to program .NET using only the command line compiler, it is
much easier and more enjoyable to use Visual Studio.NET. In this chapter we cover the
basics of using Visual Studio to edit, compile, run, and debug programs. You will then be
equipped to use Visual Studio in the rest of the book. This chapter covers the basics to get
you up and running using Visual Studio. We will introduce additional features of Visual
Studio later in the book as we encounter a need. This book was developed using beta
software, and in the final released product you may encounter some changes to the
information presented here. Also, Visual Studio is a very elaborate Windows application
that is highly configurable, and you may encounter variations in the exact layout of
windows, what is shown by default, and so on. As you work with Visual Studio, a good
attitude is to see yourself as an explorer discovering a rich and varied new country.
Overview of Visual Studio.NET
Open up Microsoft Visual Studio.NET 7.0 and you will see a starting window similar to
what is shown in Figure A-1.
Figure A-1. Visual Studio.NET main window.
What you see on default startup is the main window with an HTML page that can help
you navigate among various resources, open or create projects, and change your profile
information. (If you close the start page, you can get it back anytime from the menu Help
| Show Start Page.) Clicking on My Profile will bring up a profile page on which you can
change various settings. There is a standard profile for "typical" work in Visual Studio
("Visual Studio Developer" profile), and special ones for various languages. Since Visual
Studio.NET is the unification of many development environments, programmers used to
one particular previous environment may prefer a particular keyboard scheme, window
layout, and so on. For example, if you choose the profile "Visual Basic Developer," you
will get the Visual Basic 6 keyboard scheme. In this book we will use all the defaults, so
go back to the profile "Visual Studio Developer" if you made any changes. See Figure A-
2.
Figure A-2. Visual Studio.NET profile page.

To gain an appreciation of some of the diverse features in Visual Studio.NET, open up the
Bank console solution in the AppA directory for this Appendix (File | Open Solution...,
navigate to the Bank directory, and open the file Bank.sln). You will see quite an
elaborate set of windows. See Figure A-3.
Figure A-3. A console project in Visual Studio.NET.

Starting from the left are icons for the Server Explorer and the Toolbox, followed by the
main window area, which currently is just a gray area. Underneath the main window is
the Output Window, which shows the results of builds and so on. Continuing our tour, on
the top right is the Solution Explorer, which enables you to conveniently see all the files
in a "solution," which may consist of several "projects." On the bottom right is the
Properties window, which lets you conveniently edit properties on forms for Windows
applications. The Properties window is very similar to the Properties Window in Visual
Basic.
From the Solution Explorer you can navigate to files in the projects. In turn, double-click
on each of Account.cs and Bank.cs, the two source files in the Bank project. Text editor
windows will be brought up in the main window area. Across the top of the main window
are horizontal tabs to quickly select any of the open windows. Visual Studio.NET allows
you to select the window to show from the Windows menu. Figure A-4 shows the open
source files with the horizontal tabs.
Figure A-4. Horizontal tabs for open source files.

Toolbars
Visual Studio comes with many different toolbars. You can configure which toolbars you
wish displayed, and you can drag toolbars to position them to where you find them most
convenient. You can also customize toolbars by adding or deleting buttons that
correspond to different commands.
To specify which toolbars are displayed, bring up the menu View | Toolbars. You can also
right-click in any empty area of a toolbar. There will be a check mark next to the toolbars
which are currently displayed. By clicking on an item on this menu you can make the
corresponding toolbar button appear or disappear. For your work in this book add the
toolbars,
• Build
• Debug
Customizing a Toolbar
We want to make sure that the "Start Without Debugging" command is available on the
Debug toolbar. If it is not already on your Debug toolbar (it is a red exclamation point),
you can add it by the following procedure, which can be used to add other commands to
toolbars.
1. Select menu Tools | Customize... to bring up the Customize dialog.
2. Select the Commands tab.
3. In Categories, select Debug, and in Commands select Start Without Debugging. See
Figure A-5.
Figure A-5. Adding a new command to a toolbar.
4. Drag the selected command onto the Debug toolbar, positioning it where you desire.
Place it to the immediate right of the wedge-shaped Start button.
5. Close the Customize dialog.
Creating a Console Application
As our first exercise in using Visual Studio, we will create a simple console application.
Our program Bytes will attempt to calculate how many bytes there are in a kilobyte, a
megabyte, a gigabyte, and a terabyte. If you want to follow along on your PC as you read,
you can use the Demos directory for this chapter. The first version is in Bytes\Step1. A
final version can be found in Bytes\Step3.
Creating a C# Project
1. From Visual Studio main menu choose File | New | Project.... This will bring up the
New Project dialog.
2. For Project Types choose "Visual C# Projects" and for Templates choose "Empty
Project."
3. Click the Browse button, navigate to Demos, and click Open.
4. In the Name field, type Bytes. See Figure A-6. Click OK.
Figure A-6. Creating an empty C# project.
Adding a C# File
At this point you will have an empty C# project. We are now going to add a file Bytes.cs,
which contains the text of our program.
1. In Solution Explorer right-click over Bytes and choose Add | Add New Item.... This
will bring up the Add New Item dialog.
2. For Categories choose "Local Project Items" and for Templates choose "Code File."
3. For Name type Bytes.cs. See Figure A-7. Click Open.
Figure A-7. Adding an empty C# file to a C# project.
Using the Visual Studio Text Editor
In the Solution Explorer double-click on Bytes.cs. This will open up the empty file
Bytes.cs in the Visual Studio text editor. Type in the following program, and notice things
like color syntax highlighting to indicate reserved words as you type.
// Bytes.cs
using System;
public class Bytes
{
{
int bytes = 1024;
Console.WriteLine("kilo = {0}", bytes);
bytes = bytes * 1024;
Console.WriteLine("mega = {0}", bytes);
Console.WriteLine("giga = {0}", bytes);
Console.WriteLine("tera = {0}", bytes);
return 0;
}
}
Besides the color syntax highlighting, other features include automatic indenting. All in
all, you should find the Visual Studio editor friendly and easy to use.
Building the Project
You can build the project by using one of the following:
• Menu Build | Build

• Toolbar
• Keyboard shortcut Ctrl + Shift + B
Running the Program
You can run the program by using one of the following:
• Menu Debug | Start Without Debugging

• Toolbar
• Keyboard shortcut Ctrl + F5
You will see the following output in a console window that opens up:
kilo = 1024
mega = 1048576
giga = 1073741824
tera = 0
We will investigate the reason for the strange output later. If you press any key, as
indicated, the console window will close.
Running the Program in the Debugger
You can run the program in the debugger by using one of the following:
• Menu Debug | Start

• Toolbar
• Keyboard shortcut F5
A console window will briefly open up and then immediately close. If you want the
window to stay open, you must explicitly program for it, for example, by asking for
input. You can set a breakpoint to stop execution before the program exits. We will
outline features of the debugger later in the chapter.
Project Configurations
A project configuration specifies build settings for a project. You can have several
different configurations, and each configuration will be built in its own directory, so you
can exercise the different configurations independently. Every project in a Visual Studio
solution has two default configurations, Debug and Release. As the names suggest, the
Debug configuration will build a debug version of the project, where you can do source
level debugging by setting breakpoints, and so on. The bin\Debug directory will then
contain a program database file with a .pdb extension that holds debugging and project
state information.
You can choose the configuration from the main toolbar . You can also
choose the configuration using the menu Build | Configuration Manager..., which will
bring up the Configuration Manager dialog. From the Active Solution Configuration
dropdown, choose Release. See Figure A-8.
Figure A-8. Choosing Release in the Configuration Manager.
Build the project again. Now a second version of the IL language file Bytes.exe is
created, this time in the bin\Release directory. There will be no .pdb file in this directory.
Creating a New Configuration
Sometimes it is useful to create additional configurations, which can save alternate build
settings. As an example, let's create a configuration for a "checked" build. If you build
with the /checked compiler switch, the compiler will generate IL code to check for
integer underflow and overflow. In Visual Studio you set compiler options through dialog
boxes. The following steps will guide you through creating a new configuration called
CheckedDebug that will build a checked version of the program.
1. Bring up the Configuration Manager dialog.
2. From the Active Solution Configuration: dropdown, choose <New...>. The New
Solution Configuration dialog will come up.
3. Type CheckedDebug as the configuration name. Choose Copy Settings from Debug.
Check "Also create new project configuration(s)." See Figure A-9. Click OK.
Figure A-9. Creating a new configuration.
Setting Build Settings for a Configuration
Next we will set the build settings for the new configuration. (You could also set build
settings for one of the standard configurations, if you wanted to make any changes from
the defaults provided.) Check the toolbar to verify that the new CheckedDebug is the
currently active configuration.
1. Right-click over Bytes in the Solution Explorer and choose Properties. The "Bytes
Property Pages" dialog comes up.
2. In Configuration Properties, select Build. Change the setting for "Check for overflow
underflow" to True (see Figure A-10). Click OK.
Figure A-10. Changing the build settings for a configuration.
Debugging
In this section we will discuss some of the debugging facilities in Visual Studio. To be
able to benefit from debugging at the source code level, you should have built your
executable using a Debug configuration, as discussed previously. There are two ways to
enter the debugger:
1. Just-in-Time Debugging. You run normally, and if an exception occurs you will be
allowed to enter the debugger. The program has crashed, so you will not be able to
run further from here to single step, set breakpoints, and so on. But you will be
able to see the value of variables, and you will see the point at which the program
failed.
2. Standard Debugging. You start the program under the debugger. You may set
breakpoints, single step, and so on.
Just-in-Time Debugging
Build and run (without debugging) the Bytes program from the previous section, making
sure to use the CheckedDebug configuration. This time the program will not run through
smoothly to completion, but an exception will be thrown. A "Just-In-Time Debugging"
dialog will be shown (see Figure A-11). Click Yes to debug.
Figure A-11. Just-In-Time Debugging dialog is displayed in response to an
exception.
Click OK in the "Attach to Process" dialog and then click Break in the "Microsoft
Development Environment" dialog. You will now be brought into a window showing the
source code where the problem arose, with an arrow pinpointing the location.
To stop debugging you can use the toolbar button or the menu Debug | Stop
Debugging.
Standard Debugging
Breakpoints
The way you typically do standard debugging is to set a breakpoint and then run using the
debugger. As an example, set a breakpoint at the first line:

The easiest way to set a breakpoint is by clicking in the gray bar to the left of the source
code window. You can also set the cursor on the desired line and click the "hand" toolbar
button to toggle a breakpoint (set if not set, and remove if a breakpoint is set). Now
you can run under the debugger, and the breakpoint should be hit. A yellow arrow over
the red dot of the breakpoint shows where the breakpoint has been hit. See Figure A-12.
Figure A-12. A breakpoint has been hit.
When you are done with a breakpoint, you can remove it by clicking again in the gray bar
or by toggling with the hand toolbar button. If you want to remove all breakpoints, you
can use the menu Debug | Clear All Breakpoints, or you can use the toolbar button .
Watching Variables
At this point you can inspect variables. The easiest way is to slide the mouse over the
variable you are interested in, and the value will be shown as a yellow tool tip. You can
also right-click over a variable and choose Quick Watch (or use the eyeglasses toolbar
button ). Figure A-13 shows a typical Quick Watch window. You can also change the
value of a variable from this window.
Figure A-13. Quick Watch window shows variable, and you can change it.
When you are stopped in the debugger, you can add a variable to the Watch window by
right-clicking over it and choosing Add Watch. The Watch window can show a number of
variables, and the Watch window stays open as the program executes. When a variable
changes value, the new value is shown in red. Figure A-14 shows the Watch window
(note that the display has been changed to hex, as described in the next section).
Figure A-14. Visual Studio Watch window.
Debugger Options
You can change debugger options from the menu Tools | Options, and select Debugging
from the list. Figure A-15 illustrates setting a hexadecimal display. If you then go back to
a Watch window, you will see a hex value such as 0x400 displayed.
Figure A-15. Setting hexadecimal display in Debugging Options.

Single Stepping
When you are stopped in the debugger, you can single step. You can also begin execution
by single stepping. There are a number of single step buttons. The most
common are (in the order shown on the toolbar):
• Step Into
• Step Over
• Step Out
There is also a Run to Cursor button .
With Step Into you will step into a function, if the cursor is positioned on a call to a
function. With Step Over you will step to the next line (or statement or instruction,
depending on the selection in the dropdown next to the step buttons ). To
illustrate Step Into, build the Bytes\Step2 project, where the multiplication by 1,024 has
been replaced by a function call to the static method OneK. Set a breakpoint at the first
function call, and then Step Into. The result is illustrated in Figure A-16. Note the red dot
at the breakpoint and the yellow arrow in the function.
Figure A-16. Stepping into a function.

When debugging, Visual Studio maintains a Call Stack. In our simple example the Call
Stack is just two deep. See Figure A-17.
Figure A-17. The call stack.

Summary
Visual Studio.NET is a very rich integrated development environment (IDE), with many
features to make programming more enjoyable. In this appendix we covered the basics of
using Visual Studio to edit, compile, run, and debug programs, so that you will be
equipped to use Visual Studio in the rest of the book. Nonetheless, it is worth spending
time to become familiar with many more of the Visual Studio features, because
understanding how to use them will make your development work much easier. A project
can be built in different configurations, such as Debug and Release. Visual Studio.NET
has a vast array of features for building database applications, Web applications,
components, and many other kinds of projects. We discuss some of these additional
features in the chapters where they are pertinent.

Application Development Using C# and

Uploaded by

Copyright:

Available Formats

Application Development Using C# and

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Application Development Using C# and

Uploaded by

Copyright:

Available Formats

Application Development Using C# and .

The Integrated .NET Series From Object Innovations Top

Introduction to C# Using .NET

Application Development Using C# and .NET

Introduction to Programming Visual Basic Using .NET

Application Development Using Visual Basic and .NET

.NET Architecture and Programming Using Visual C++

Fundamentals of Web Applications Using .NET and XML

Programming PERL in the .NET Environment

Microsoft .NET is an advance in programming technology that greatly simplifies

The Web site for the book series is:

Additional information about .NET technology is available at:

The Integrated .NET Series from Object Innovations

About this Series

Robert J. Oberg, Series Editor

Introduction to C# Using .NET

Introduction to Programming Visual Basic Using .NET

Intermediate .NET Framework Books

Application Development Using C# and .NET

Application Development Using Visual Basic .NET

Fundamentals of Web Applications Using .NET and XML

Chapter 1. What Is Microsoft .NET? Top

Microsoft .NET provides:

• The Common Language Runtime, a robust runtime platform.

Microsoft and the Web

Applications in the Internet Age

Open Standards and Interoperability

Windows on the Desktop

Problems with Windows

Maintaining a Windows PC is a chore, because applications are quite complex. They

The Glass House and Thin Clients

A New Programming Platform

• Code can be validated to prevent unauthorized actions.

There are several important features to the .NET platform:

Modern programming relies heavily on reusable code provided in libraries. Object-

Examples of classes in the .NET Framework include Windows programming, Web

Common Language Runtime

A runtime provides services to executing programs. Traditionally there are different

Managed Code and Data

Managed data is allocated and deallocated automatically. This automatic deallocation is

Microsoft and ECMA

Multiple Language Development

A practical key to success in software development is a set of effective tools. Microsoft

• VS.NET provides a very high degree of functionality for creating applications in

The Importance of Tools

The importance of tools should not be underestimated. Ada, a very powerful

The Role of XML

• XML can be used to model data in coordination with ADO.NET datasets.

Success Factors for Web Services

Chapter 2. .NET Fundamentals Top

Problems of Windows Development

Deployment of applications can be a chore. Critical entries have to be made in a Registry

Applications of the Future

The Magic of Metadata

[Serializable] class Customer

Customer cust = new Customer();

cust = new Customer();

foreach (Customer x in list)

Console.WriteLine("Saving Customer List");

Console.WriteLine("Restoring to New List");