Software Development Lifecycle Models

Download as rtf, pdf, or txt
Download as rtf, pdf, or txt
You are on page 1of 16

Software Development Lifecycle Models

1. (Spring 1997) Discuss the activities (steps or actions) that performed during the
requirements engineering phase of the software life cycle. Be sure to address the
outputs from each activity.

1. Software life cycle:

The software life cycle typically includes:

 Concept phase:
The initial phase of a software development project, in which the user needs are
described and evaluated through documentation
Input: Customer consultation
Output: Customer requirements

 Requirements phase:
The period of time in the software life cycle during which the requirements for a
software product are defined and documented
Input: Customer requirements
Output: Requirements specification

 Design:
The process of defining the architecture, components, interfaces, and other
characteristics of a system or component.
Input: Requirement document
Output: Design description

 Implementation:
All design decisions should have been made. The period of time in the software life
cycle during which a software product is created from design documentation.
Input: Design description
Output: Source code and executable program

 Test:
The period of time in the software life cycle during which the components of a
software product are evaluated and integrated, and the software product is
evaluated to determine whether or not requirements have been satisfied.
Input: Test plan
Output: Test summary report

 Installation and Checkout:


The period of time in the software life cycle during which a software product is
integrated into its operational environment and tested in this environment and
tested in this environment to ensure that it performs as required.

1
Input: Software system (including documentation)
Output: Customer acceptance and feedback

 Operation and Maintenance:


The period of time in the software life cycle during which a software product is
employed in its operational environment, monitored for satisfactory performance,
and modified as necessary to correct problems or to respond to changing
requirements.

 Retirement:
The period of time in the software life cycle during which support for a software
product is terminated.

2. Software development lifecycle models:

There are four common development tasks associated with all software lifecycle
models. These tasks are requirements (analysis and specification, design,
implementation and testing). Explain how these tasks are accomplished in the linear
sequential, incremental and spiral models (for each task, address when it begins,
when it ends, and whether or not the task is revisited). (Exam1 Fall 99)

Compare and contrast the linear sequential model with the incremental and spiral
life cycle models.

♦ Linear Sequential (waterfall model or classic life cycle):


Sometimes called the "classic life cycle" or the "waterfall model", the linear
sequential model suggests a systematic, sequential approach to software development
that begins at the system level and progresses through analysis, design, coding,
testing and maintenance.

Characterized by a sequential procession through the software process stages


(system/information engineering and modeling, software requirements analysis,
design, code generation, testing, maintenance). Some amount of feedback or overlap
may occur, but for the most part the model is strictly linear.
(picture)

Picture (page 31, textbook)

Pros:
-- Easy to identify which stage development is in
-- Each development team is working with "full" information
-- Good for systems in which all requirements are known at the outset
Cons:
-- Customer doesn't see the program until late in the life cycle
-- Requirements must be complete at the outset of the project
-- Doesn't accommodate change easily

♦ Prototyping Model:
Full system requirements are not always available at the outset of a project. Customer
may be unsure of some functionality. Developer may be unsure of the “best”
solution. There may be some human-computer interaction issues. The
prototyping model is design for these situations.
(Picture)

In most projects, the first system built is barely usable. It may be too slow, too big,
awkward to use, or all three. There is no alternative but to start again, smarting
but smarter, and build a redesigned version... The management question here is
whether to plan in advance to deliver a throwaway to customers.
Mock-ups can be story-boards (paper drawing), interactive (hyper-text, web pages),
or partially functional (visual tools).
Pros:
-- good for eliciting requirements for systems where customer is unsure of what
should/can be built
-- customer and developer have a baseline to work against
-- customers get early interaction with system
-- reduced likelihood of problems due to miscommunication
Cons:
-- the real system should then be built using a more formal life cycle model
-- customer may expect "look and feel" of the prototype
-- customer may want to have prototype delivered rather than waiting for full, well
engineered version
-- quality can suffer

♦ Rapid Application Development (RAD):


The primary phases of RAD are Business modeling, data modeling, process
modeling, application generation, testing and turnover. It is a "high-speed"
adaptation of the linear sequential model. Requirements must be well understood.
Requires enough people to form RAD teams. Requires customer commitment to
enable the process to operate at full speed.
RAD is a component-based technique that makes heavy use of reusable software
components. Each team works on separate program functionality.

3
(picture)

Pros:
-- can deliver a full product in a short time period
-- promotes code reuse
Cons:
-- can require more people
-- requires a system that can be properly modularized
-- can fail if reusable components are not available

♦ Evolutionary:
• Incremental:
It combines elements of the linear sequential model and prototyping.
Requirements of the system are prioritized. The first increment is the delivery
of the core product, providing only the highest priority requirements.
Subsequent iterations expand on this core (enhancement maintenance).
(picture)

Picture (page 38, textbook)

Pros:
-- delivers a usable working product quickly
-- keeps all teams working
Cons:
-- can "design yourself into a corner"
-- works best with an even distribution of different priority features
-- requires good planning and design

• Spiral:
It incorporates iterative nature of prototyping with aspects of the linear sequential
model. Typical framework activities or task regions: customer
communication, planning, risk analysis, engineering, construction & release,
and customer evaluation. Development proceeds in increments. Each
increment contains all task regions. Early increment produces paper model or
prototype. Late increments produce more complete versions. All work
products are produced this way.
Picture (page 40, textbook)

Pros:
-- includes risk analysis
-- customer communication at all stages
-- entry points into the process for maintenance activities
Cons:
-- May require unrealistic overhead for small projects
-- requires considerable risk assessment expertise
-- hasn't been used as widely as linear sequential

• Component Assembly:
Essentially a spiral model. Relies on object technologies to provide reusable
software components. Uses object-oriented analysis to identify candidate
classes. Existing classes are used or new classes are created.
(Picture)

Pros:
 Highly reuse of code if available
 Process will be fast with reusable code
 Reusable codes have high reliability
Cons:
 have to decide which classes exist, otherwise, create new ones

• Concurrent Development:
Concurrent process model represents the major activities, tasks, and their states.
Each activity can be represented by a state chart. Defines a series of events
that trigger transitions from state to state for each activity. Often used as the
paradigm for development of client/server applications.
(Picture)

Pros:
-- Allows management to assess the status of a project in an evolutionary model.

5
-- Has been used in the development of client/server applications.
Cons:
 More of a state chart for activities than a development model.

3. Discuss the activities (steps or actions) that are performed during the
requirements engineering phase of the software life cycle. Be sure to address the
outputs from each activity. (Spring 97)

2. (Fall 1997) There are five principal phases to the linear sequential or "waterfall"
life cycle model. Please discuss each of these phases, addressing the activities that
are performed during each phase. Also identify the inputs and outputs
associated with each phase. Compare and contrast the linear sequential model
with the incremental and spiral life cycle models.

The five principal phase to the linear sequential or "waterfall" life cycle model:

 Software Requirements Analysis:


The period of time in the software life cycle during which the requirements for a
software product are defined and documented
Input: Customer requirements
Output: Requirements specification

 Design:
The process of defining the architecture, components, interfaces, and other
characteristics of a system or component.
Input: Requirement document
Output: Design description

 Implementation/Code Generation:
All design decisions should have been made. The period of time in the software life
cycle during which a software product is created from design documentation.
Input: Design description
Output: Source code and executable program

 Testing:
The period of time in the software life cycle during which the components of a
software product are evaluated and integrated, and the software product is
evaluated to determine whether or not requirements have been satisfied.
Input: Test plan
Output: Test summary report
• Maintenance
The period of time in the software life cycle during which a software product is
employed in its operational environment, monitored for satisfactory performance,
and modified as necessary to correct problems or to respond to changing
requirements.
Capability Maturity Model

The purpose of CMM is to assess the "maturity" level of a software development


organization.
• Characteristics of an immature organization
-- Processes are made up " on the fly"
-- Specified processes are not followed if they exist
-- Projects are routinely over budget
-- Projects are often behind schedule
• Characteristics of a mature organization
-- Processes are defined, understood and followed
-- Projects typically are delivered within budget
-- Projects are completed on time
-- Processes are monitored and improved

Five levels to CMM, the key activities are required at each level, and practices needed to
advance to the next level. Level one is the least mature level and level 5 is the most
mature level. An organization can’t skip a level or stage in its maturation.

 Level 1 -- Initial:
• The key practices:
The software process is ad hoc (code & fix) or just chaotic. Few processes are
defined, fewer are followed. Requirements and Design may be nonexistent.
Success depends on individual effort
• Practices needed to advance to Level 2:
-- Project management. It is focused on control of commitments (schedule, resources,
cost).
-- Management oversight. It means review and approval of all major development
plans prior to commitment.
-- Quality assurance. It is responsible for making sure that work is being done the way
it is supposed to be done.
-- Change control. It applies to both management documents and software work
products. This includes changes to requirements documents, design and code.

 Level 2 -- Repeatable:
• The key practices:
Basic project management processes are in place to track cost, schedule and
functionality. Process discipline exists so that previous successes can be repeated
for similar projects. Plans and commitments are made in a controlled manner.
Everything depends on prior experience with similar work. Process tends to be
abandoned when schedule or budget gets tight. The organization has achieved a
stable process with a repeatable level of statistical control by initiating rigorous
project management of commitment, costs, schedules, and changes.

7
• Practices needed to advance to Level 3:
-- Establish a process group. A process group is a group of people that focus
exclusively on improving the software process.
-- Establish a software development process architecture. Software development
process architecture is synonymous to a development life cycle. It describes the
development phases and associated work products.
-- Introduce a family of software engineering methods and technologies. Software
engineering methods and technologies include design and code inspections,
formal design methods, configuration management tools, testing methods, and
prototyping and modern languages.

 Level 3 -- Defined
• The key practices:
Process for both management and engineering activities are documented and
standardized. All projects use a documented and approved version of the process
for developing and maintaining systems. It includes all activities defined for level
2. The organization has defined the process as a basis for consistent
implementation and better understanding. At this point advanced technology can
usefully be introduced.
• Practices needed to advance to level 4:
-- Establish a set of process measurements.
-- Create a process database to maintain a history.
-- Allocate resources to maintain this information and to train project members to use
it.

 Level 4 -- Managed:
• The key practices:
Detailed measures of software process and product quality are collected. Both process
and product are quantitatively understood. Quality control is based on
measurement. It includes activities defined for level 3. The organization has
initiated comprehensive measurement and analysis procedures. This is when the
most significant quality improvement begins.
• Practices needed to advance to level 5:
-- Automated gathering of process data to avoid error and omission.
-- Use of process data to analyze and modify the process to prevent problems and
improve efficiency.

 Level 5 -- Optimizing:
• The key practices:
Continuous process improvement based on quantitative feedback. Testing of
innovative ideas and technologies. It includes all activities defined for level 4.
The organization now has a foundation for continuous improvement and
optimization of the process.
1. Why are there five levels to the Capability Maturity Model (CMM)? (exam2 Fall
99)

-- Reasonably represent the actual historical phases of evolutionary improvement of real


software organizations
-- Represent a measure of improvement that is reasonable to achieve from the prior level
-- Suggest interim improvement goals and progress measures
-- Make obvious a set of immediate improvement priorities, once an organization's status
in this framework is known.

Testing --- Software Testing Strategies

Testing is the development phase that follows implementation and precedes installation
and maintenance. Th purpose of this phase is to assess the correctness of the system that
was produced.

Two main tasks in software testing:


1) Verification--refers to the set of activities that ensure that software correctly
implements a specific function, conforms the design and the process was followed.
( Are we building the product right ?)
2) Validation-- refers to a different set of activities that ensure that software that has
been built is what the customer intended and is traceable to customer requirements.
( Are we building the right product?)
*The reasons that a developer should not do the software testing:
1) Analysis and design are constructive tasks, while testing is a destructive task;
2) If they test well they find errors, if they program well they do not make errors;
3) Developers tend to overlook some errors as trivial, fix them and move on;
4) Developers and the ITG (independent Test Group) should work closely through out a
software project.

*What gets tested?


Source code and documentation.

Software testing can be seen as a bottom up process.


• At the lowest level, individual modules or units are tested independently.
• These units are then combined into sub-systems and are tested further.
• Finally, all the sub-systems are aggregated and the whole system is tested.

The previous strategy involves four different types of testing:


• Unit test
• Integration test
• Validation test

9
• System test

These four levels of testing reflect back to four different software development products:
• system requirements ----system test
• software requirements ----validation test
• design documentation ----integration test
• source code --- -unit test

Unit testing
--Unit testing focuses on the source code of a specific software unit(module).
--The purpose of unit testing is to verify that a particular software unit performs its
assigned function correctly.
--This is the lowest level of testing.

Two tools in unit testing


1) Stubs-- Some units may not have been written and tested yet. In order to test a unit
that makes use of these other units, stubs may need to be created. A stub is a stand-in
for a real unit, which simulates the actual unit being accessed.
2) Drivers-- In order to test a unit, something has to cause it to execute. A driver is just a
dummy "main program" that is responsible for executing the unit to be tested. The
driver must supply necessary parameters as well as memory allocation when needed.

Integration Testing
Integration testing follows unit testing and it combines the tested units into subsystems
and verifies that, when the units work together, the subsystem operates properly.

Integration testing approaches:


• Incremental
• non-incremental

Incremental integration typically proceeds in one of these manners


• Top-Down integration-- The system is integrated beginning with the " main" control
module. Modules subordinate to the main control module are incorporated next.
These modules are incorporated either depth-first or breadth-first. Stubs are created
for lower level modules.
• Bottom-Up integration-- Bottom level modules are integrated into clusters with their
calling module. This cluster represents a sub-system. These sub-systems are
subsequently integrated into larger sub-systems. Drivers are created for higher level
modules during integration testing.

Validation Testing
Once the entire system is integrated, using the design as a blueprint, it must be validated.
At this point, the goal is to determine if the system that was constructed satisfies the
user's requirements. The input to validation testing is the software requirements
document.
System Testing
If the software is to become part of a larger system, system testing must be performed.
The purpose of system testing is to verify and validate that the software product operates
correctly within the confines of the larger system.

Other concepts about testing

 Regression Testing is to re-test the changed modules at the unit level, and then
integration testing is performed to ensure that the sub-system still works (usually
bottom-up).

 Statistical Testing ( traditional ) is the name for one approach to finding out how
many faults are likely to be present in a software system. The first step in
statistical testing is to randomly inject errors into the software prior to testing it. The
software is then tested and faults are found and reported. The intentionally injected
faults and actual faults are categorized. A ratio is calculated that represents how
many real faults are found per injected fault. This is used to predict the number of
real faults.

 Mean Time To Failure is a metric hold over from hardware testing, where a
hardware device can fail as an artifact of stress from the physical environment as
well as design limitations. Software does not fail as an artifact of time, failure is
instead related to how the software is used. The software failure may occur: when
software transitions among its modules at blinding speed; in different level of
granularity when time must be measured in; in modules that are not themselves
faulty.

 Black box testing is the term applied to the process of testing a software system, or
subsystem, without looking at its internal representation. Testing involves providing
the software some inputs, collecting the outputs, and verifying that the outputs are
correct.

 White Box Testing ( clear box testing) involves looking at the internal construction
of the software. It involves testing all statements, decisions, function calls, etc.

 Gray Box Testing is the term used to describe everything between white box and
black box testing. Primarily it involves using some amount of white box
information to guide black box testing. Software measurements can be used to
decide how much black box testing to perform on various subsystems.

11
Object-Oriented Testing Strategies

• Object-Oriented Testing
Assessing the needs of OO Testing requires that three things be done
1) The definition of testing must be broadened to include error discovery techniques
applied to OOA and OOD Models.
2) The strategy for unit and integration testing must change significantly
3) The design of test cases must account for the unique characteristics of OO
software.

Testing OOA and OOD Models:


1) Correctness: semantic correctness, class definition and class hierarchy for omission
and ambiguity, class relationship;
2) Consistency: relationships, class-responsibility-collaboration (CRC) model and
object-relationship diagram can be used to facilitate consistency checking.

• Object-Oriented Testing Strategies.

1) Unit testing of OO software -- Class testing for OO software is the equivalent of


unit testing for conventional software.
• Unit testing of conventional software focuses on algorithmic detail and data flow
across the module interface.
• Class testing for OO software is driven by the operations encapsulated by the class
and the state behavior of the class.

2) Integration Testing of OO Software –


• Thread- Based testing: integrates the set of classes required to respond to one
input or event for the system. Each thread is integrated and tested
individually.
• Use-based testing:
a. begins the construction of the system by testing those classes
(independent classes) that use very few server classes.
b. Next, the dependent classes (those that use the independent classes ) are
tested.
c. This process iterates until entire system is constructed.
d. The use of drivers and stubs as replacement operations is to be avoided
when possible.
• Cluster testing: represents one step in the integration testing of OO software.
A cluster of collaborating classes is executed by designing test cases that
attempt to uncover errors in the collaborations.
3) Validation Testing of OO Software –
• Focuses on user visible actions and user recognizable outputs from the system.
• Tester draws upon use cases that are part of the analysis model.
• Conventional black-box testing methods can be used to drive validation tests.
• Object-behavior method and even flow diagrams can also be used.

*Approach to OO Test Case Design


1)Each test case should be uniquely identified and associated with the class to be tested;
2)The purpose of the test should be stated;
3)A list of testing steps should be developed for each test.

*Contents of Testing Steps:


1) A list of specified states for the object that is to be tested;
2) A list of messages and operations that will be exercised;
3) A list of exceptions that may occur;
4) A list of external conditions;
5) Supplementary information.

3. (Spring 1998) Contrast structural analysis and design methods and object-
oriented analysis and design methods. Focus your discussion on the different
views (models) of the problem domain and the ability of each of these methods to
express these views.

The structural analysis:


• Structural analysis takes a distinct input-processing-output view of requirements.
Data are considered separately from the processes that transfer the data.
• The structural analysis approach makes heavy use of functional decomposition.
• Structural analysis focuses attention away from the inherent properties of objects
during the modeling process and lead to a model of the problem domain that is
orthogonal to the three essential principles of object-orientation: Encapsulation,
classification of objects and inheritance.
• Tools:
The Structural Analysis consists of three models:
1. Functional Model -- Data flow diagram (DFD) serve as a basic to the functional
Model. DFD provides an indication of how data are transformed as they move
through the system and depicts the functions that transform the data flow.
2. Behavior Model -- The state transition diagram (STD) serves as a basis for
behavior model. The STD represents the behavior of a system by depicting its
states and the events that cause the system to change state.
3. Data Model -- Data model consists of data object, attributes and relationships.
Data object is described by attributes. Relationships can be represented using

13
Entity-Relationship Diagram(ERD)
The Object-Oriented analysis:
• The intent of object-oriented analysis is to define all class (and the relationships and
behavior associated with them) that are relevant to the problem to be solved.
• Object-oriented analysis focuses on encapsulation, classification and inheritance of
objects.
• Tools:
Data Model -- Data model consists of data object, attributes and relationships. Data
object is described by attributes. Relationships can be represented using Entity-
Relationship Diagram (ERD).

Structural design (conventional design):


Software Design ----The process of applying various techniques and principles for the
purpose of defining a device, a process or a system in sufficient detail to permit its
physical realization.
Input --- SRS
Output --- All potential solutions
• Differences between Specification and Design:
1. Specification answers the question of "What the system will do." Prototypes and
mock-ups here address user interaction with the system; business processes and
constrains, requirements of process interaction, and other product environment issues
and interfaces.
2. Design answers the question of "How will it be done". Prototypes and mock-ups here
address how data will be represented, how data will be transformed, how control will
flow through the system, how the system will react to stimuli, and other
implementation specific details.

There are four layers in structural design:


1. Data design: Transforms information from analysis into the data structures that will
be required to implement the software. The information can be gathered from entity-
relationship diagram or a data dictionary provided in the specification.
2. Architecture design: defines the relationship among major structural elements
(processes or functionalities). It depicts interactions of subsystems defined in the
analysis model. It is typically a top-down decomposition of the program into
subsystems.
3. Interface design: Includes communication among subsystems, communication to
external systems (other software products), and interaction with humans that use the
program.
4. Procedural design: takes the components/subsystems identified in architectural design
and produces a procedural description of each, and provide the low-level design for
the system, which can be translated into a programming language implementation.

* Characteristics of "good" design:


• Must implement all of the explicit requirement in the analysis model
• Must accommodate all of the implicit requirements
• Must be readable and understandable by those who generate the code and those who
subsequently maintain the system
• Must provide a complete picture of the system, the view of the system includes: data,
functional, and behavioral

* Cohesion module:
performs single well-defined task, have little interaction with other modules when
it comes to accomplishing its specified task, becomes a self-contained unit that
can be reused wherever its particular service is needed.
Types: coincidental, logical, temporal, procedural, communicational (highest).
* Coupling:
A measure of the interconnection between module, determining the dependencies
that a module has on other modules, indicates how difficult the module is to
understand, reuse, and modify. Types: data, stamp, control, external, common,
content (highest).

* A well design module strives to achieve high cohesion and low coupling for easier
reuse, understand and maintain.

* Fan-in: the number of connections coming into a module.


* Fan-out: the number of modules coming out of a module.

Object-oriented design:
Object-oriented design (OOD) transforms the analysis model created using OOA into a
design model that serves as a blueprint for software construction. OOD draws on four
important design concepts: abstraction, information hiding, functional independence, and
modularity.
There are four design layers: subsystem layer, class and object layer, message layer, and
responsibility layer.
1. Subsystem layer represents each of the subsystems that enable the software to achieve
its customer-defined requirements.
2. The class and object layer contains the class hierarchies that enable the use of
generalization and specifications; it contains the design representations of each
object.
3. The message layer contains the details of object communications and establishes the
external and internal interfaces.
4. The responsibilities layer contains the data structure and algorithmic design for all
attributes and operations for each object.

* Difference between conventional design and OO design:


1. Conventional design focuses on process. Problems are broken into the functions that
must be performed to process inputs into outputs. Data are passed to these functions,
transformed, and passed along to other functions. The design exhibits a hierarchical
control structure.
2. OO design focuses more on collaborations among objects. An object is consisted of
both data and the methods that manipulate the data. OOD considers abstraction,
information hiding, functional independence, and modularity. It has 4 layers:

15
subsystem, class & object, message, and responsibilities layers. The Modeling
components for OOD methods include: representation of hierarchy of modules,
specification of data definitions, specification of procedural logic, indication of end-
to-end processing sequences, representation of object states and transitions, definition
of classes and hierarchies, assignment of operations to classes, detailed definition of
operations, specification of message connections, identification of exclusive services.
The design issues of OOD include decomposability, composability, understandability,
continuity, and protection. The OOD models include Booch's, Coad & Yourdon's,
Jacobson's, Rambaugh's, and Wirfs-Brock's. The four design components of OOD are
problem domain, Human interaction, Task management, and data management.

You might also like