SoftwareWars 200911 PDF

Draft: 1.
027, 11/07/2009
Words: 97474, Pages: 304
Copyright 2009 by Keith Curtis

I am making this book available as a free digital download. However, I
would prefer that you not give copies to other people, and instead tell them
to download it from lulu.com. This allows me to ensure all readers get the
latest version, and to keep track of the number of copies out there. Your
copy is yours, of course.
If you enjoyed a free version of this book, and you want to send me a
donation as thanks for my two years of labor, that would be appreciated!
You could purchase a paper copy, or go to keithcu.com and click on the PayPal button. Any amount donated over $5 will be given to worthy efforts in
free software. If you'd like to contribute money towards a particular area of
free software, but don't know how, I can help!
Several people advised me to use a Creative Commons license, but I did
not see anything that looked like this.
In general, I tried to get permission for the use of other's information.
However, I have over 100 images and it was hard to figure out where to get
permission for some of them. For those, I will claim permission under fair
use ;-) I am happy to remove any content if any owner objects.
Keith Curtis
[email protected]
twitter: @keithccurtis
ISBN 978-0-578-01189-9
TABLE
OF
CONTENTS
Free Software Battle..........................................................................1

Free Software Army.......................................................................3
iBio.................................................................................................5
Glossary.............................................................................................9
Wikipedia.........................................................................................10
Linux................................................................................................16
Distributed Development.............................................................20
Linux Kernel Superiority.............................................................24
The Feature Race.........................................................................35
Linux is Inexorably Winning........................................................38
Charging for an OS......................................................................39
Free Software Only Costs PCs.....................................................42
A Free Operating System............................................................43
Linux Distributions......................................................................49
AI and Google...................................................................................53
Deep Blue has been Deep-Sixed..................................................53
DARPA Grand Challenge..............................................................54
Software and the Singularity.......................................................59
Google..........................................................................................61
Conclusion...................................................................................69
Free Software..................................................................................70
Software as a Science..................................................................71
Definition of Free Software.........................................................74
Copyleft and Capitalism...............................................................75
Is Copyleft a Requirement for Free Software?............................76
Why write free software?.............................................................78
Should all Ideas be Free?............................................................89
Pride of Ownership......................................................................90
Where Does Vision Fit In?...........................................................91
Governments and Free Software.................................................92
Should all Software be GPL?.......................................................94
Microsoft's Responses to Free Software.....................................95
Just a Stab...................................................................................97
Patents & Copyright.........................................................................99
Software is math........................................................................103
Software is big...........................................................................105
Software is a fast-moving industry............................................106
Copyright provides sufficient protection...................................106
Conclusion.................................................................................107
4
Biotechnology Patents ..............................................................109
Openness in Health Care...........................................................112
The Scope of Copyright.............................................................114
Length of Copyright...................................................................114
Fair Use.....................................................................................116
Digital Rights Management (DRM)............................................117
Music versus Drivers.................................................................121
Tools...............................................................................................123
Brief History of Programming...................................................125
Lisp and Garbage Collection......................................................129
Reliability...................................................................................132
Portability..................................................................................140
Efficiency...................................................................................143
Maintainability...........................................................................147
Functionality and Usability........................................................149
Conclusion.................................................................................150
The Java Mess................................................................................152
Sun locked up the code..............................................................154
Sun obsessed over specs...........................................................156
Sun locked up the design...........................................................158
Sun fragmented Java.................................................................159
Sun sued Microsoft....................................................................160
Java as GPL from Day 0.............................................................160
Pouring Java down the drain......................................................162
Let's Start Today........................................................................164
The OS Battle.................................................................................169
IBM............................................................................................170
Red Hat......................................................................................172
Novell........................................................................................174
Debian.......................................................................................175
Ubuntu.......................................................................................179
Should Ubuntu Have Been Created?.........................................182
One Linux Distro?......................................................................187
Apple..........................................................................................190
Windows Vista...........................................................................201
Challenges for Free Software........................................................205
More Free Software...................................................................206
Cash Donations..........................................................................207
Devices......................................................................................209
Reverse Engineering.................................................................211
PC Hardware.............................................................................212
Fix the F'ing Hardware Bugs!....................................................214
5
Metrics.......................................................................................215
Volunteers Leading Volunteers..................................................216
Must PC vendors ship Linux?....................................................217
The Desktop...............................................................................219
Approachability..........................................................................220
Monoculture..............................................................................223
Linux Dev Tools..........................................................................225
Backward Compatibility............................................................226
Standards & Web...........................................................................228
Digital Images............................................................................229
Digital Audio..............................................................................229
The Next-Gen DVD Mess...........................................................230
MS's Support of Standards........................................................232
OpenDocument Format (ODF)...................................................234
Web............................................................................................240
Da Future.......................................................................................246
Phase II of Bill Gates' Career.....................................................246
Space, or How Man Got His Groove Back.................................249
The Space Elevator....................................................................254
21st Century Renaissance.........................................................266
Warning Signs From the Future................................................268
Afterword.......................................................................................270
US v. Microsoft..........................................................................270
Microsoft as a GPL Software Company.....................................272
The Outside World.....................................................................275
How to try Linux............................................................................295
Dedication......................................................................................296
Acknowledgments......................................................................296
Free Software Battle
FREE SOFTWARE
BATTLE
Some people think much faster computers are required for Artificial Intelligence, as well as new ideas. My own opinion is that
the computers of 30 years ago were fast enough if only we
knew how to program them.
John McCarthy, computer scientist, 2004
This IBM 305 RAMAC Computer, introduced in 1956, was the first computer
containing a (5 MB) hard drive on 24 huge spinning platters. Today you can
get 1000 times more memory in something the size of your thumb.
iven the technology that's already available, we should have

cars that drive us around, in absolute safety, while we lounge
in the back and sip champagne. All we need is a video camera on the roof, plugged into a PC, right? We have all the necessary
hardware, and have had it for years, but don't yet have robot-driven
cars because we don't have the software. This book explains how we
can build better software and all get our own high-tech chauffeur.
The key to faster technological progress is the more widespread
use of free software. Free versus proprietary (or non-free) software
is similar to the divide between science and alchemy. Before science, there was alchemy, where people guarded their ideas because
they wanted to corner the market on the means to convert lead into
gold. The downside of this strategy is that everyone would have to
learn for themselves that drinking mercury is a bad idea.1 The end of
the Dark Ages arrived when man started to share advancements in
math and science for others to use and improve upon. In fact, one
way to look at history is to divide it between periods of progress and
stagnation.
Computers are an advancement comparable to the invention of
movable type. While computers and the Internet have already
changed many aspects of our lives, we still live in the dark ages of
computing because proprietary software is still the dominant model.
One might say that the richest alchemist who ever lived is my former boss, Bill Gates. (Oracle founder Larry Ellison, and Google cofounders Sergey Brin and Larry Page are close behind.)
This book will discuss free software, but the question of whether
scientific research and resources of information such as libraries
should be free was answered long ago. In fact, the first (privately
funded) library in America was created by Ben Franklin in 1731, 45
years before the nation itself was founded. The library's motto was
the Latin phrase: To support the common good is divine. Ben
Franklin understood that sharing knowledge has no downside.
Human knowledge increasingly exists in digital form, so building
new and better models requires the improvement of software. People can only share ideas when they also share the software to display and modify them. It is the expanded use of free software that
will allow a greater ability for people to work together and increase
the pace of progress. The case studies examined in this book demonstrate that a system where anyone can edit, share, and review the
1
The digital version of this book has a number of hyperlinked words that take you
to references, like this video of writer Cory Doctorow at a Red Hat Summit.
body of work will lead not just to something that works, but eventually to the best that the world can achieve! Better cooperation
among our scientists will lead to, robot-driven cars, pervasive robotics, artificial intelligence, and much faster progress in biology, all of
which rely heavily on software.
A later chapter will describe the software freedoms in more
detail, and the motivations for programmers to use and write free
software, but it is important to clarify here that free software generally means that the source code is made available to its users.
Microsoft's Internet Explorer is not free because it requires a Windows license, but more importantly, you cannot download the source
code to learn how it works.
Today, proprietary software is considered more valuable than free
software because its owners charge for a black box, but that thinking is exactly backwards. Proprietary software is less valuable
because you cannot learn how it works, let alone improve it. It cannot make you better, and you cannot make it better. It is true that
not everyone will exercise the right to read and change their software, just as not everyone exercises their right to their freedom of
the press, but that doesn't make the freedom any less valuable!
Free Software Army

Justice officials argued that Microsoft's power was impregnable
because consumers were so dependent on Windows. His voice
rising, Gates exclaimed, You give me any seat at the table:
using Linux or Java I can blow away Microsoft!
World War 3.0: Microsoft and its enemies, Ken Auletta
Glenn Reynolds, in his book Army of Davids, talks about how

armies, like bloggers in pajamas, are changing journalism and other
aspects of our lives. This book will focus on the free software army,
created by Richard Stallman in 1985. The rank and file of this army
consists of loosely-knit programmers, who live in many countries,
speak different mother tongues, and either work for competing companies, or volunteer their time, to place their fingerprint on the
world's software knowledgebase.
Sourceforge.net, the largest free software repository, has
1,900,000 registered developers today. Even if we divide that number by 50, because many work part-time, we are still left with an
army of 38,000, three times bigger than the development teams of
Google and Microsoft combined. And SourceForge is just one free
software community; most of the bigger teams use their own servers
to manage and organize the development process.
The most important piece of free software is the Linux (pronounced Lin- ex) operating system, named after its founder Linus
Torvalds, who started coding it in college. While Linux is generally
not used on desktops today, it and other free software run on 60% of
all websites, an increasing number of cellphones, and 75% of the
world's top 500 fastest supercomputers:
IBM's Blue Gene (pun intended) supercomputer runs a lightweight Linux on

each compute node, and a complete Linux on its management nodes.
For its part, Microsoft has fiercely fought against Linux and the
trend towards free software by pretending it is just another proprietary competitor. With $28 billion in cash, dominant market share in
Windows, Office and Internet Explorer, and an army of thousands of
experienced programmers, Microsoft is a focused and enduring
competitor.
Microsoft is the largest proprietary software company, but others
have adopted its philosophy of hoarding all knowledge, no matter
how irrelevant to their bottom line or useful to others. Google, the
dominant player in Internet search, relies heavily on free software
and considers it an important part of their success, but they are very
secretive and protect nearly all the software they produce. They are
a black hole of free software: innovation enters but never leaves.
This is all perfectly legal and ethical, and the free market gives
everyone an unfettered right to innovate in any way, create any
license agreement, and charge anything for a product. But free software is not just a competitor, it is a different way of creating software.
The free software community has long threatened to take over the
world. Evangelist Eric Raymond once growled to a Microsoft VIP
that he was their worst nightmare. That was in the mid-1990s,
when Microsoft stock price was doing this:
Microsoft stock price, 1990 2000
A friend installed Linux in the mid-90s but he gave up because his

Backspace key didn't work. Free software has come a long way
since then, reaching technical critical mass, if not market dominance. This book will discuss the remaining technical challenges
preventing world domination, but it inertia and ignorance are its
biggest obstacles.
While this book presents a vision of the future, I believe we could
have had these advancements decades ago. Free software's paradoxical success should also cause us to question other assumptions
about copyright, patents, and other topics that this book will
address.
iBio
I first met Bill Gates at the age of twenty. He stood in the yard of
his Washington lake-front home, Diet Coke in hand, a tastefully
small ketchup stain on his shirt, which no one had the courage to
point out, and answered our questions, in-turn, like a savant. As a
college summer intern, I had planned for a potential encounter and I

approached him with questions that interested me but which would
be arcane to non-computer mortals.2
His answers demonstrated that he was one of the top software
experts on the planet and convinced me that I would be wise to start
off my career at Microsoft.
Writing software is a craft, like carpentry. While you can read books on programming languages and software algorithms, you can't learn the countless
details of a craft from a book. You must work with experts on real-world
problems. Before free software, you had to join a company like Microsoft.
I joined Microsoft in 1993 when it was hitting its stride. It had

recently released Windows 3.1 and Windows NT, setting itself on the
path of more than a decade of dominance in the PC operating system market, and the many other markets that flow from it. I worked
as a programmer for 11 years in a variety of different groups on
databases, Windows, Office, MSN, mobility, and research.
One day it just hit me I should quit. There were no big reasons,
only a lot of little ones. I had just launched v1 of the client and
server side of the Microsoft Spot watch, and while it contained
2
I asked him about the performance of Microsoft Exchange's database storage

engine as compared to the one inside Microsoft SQL Server, and about NetWare's
newly-announced clustering technology called SST Level 3.
sophisticated technologies, I didnt really believe it would take off in

the marketplace. I had gained lots of knowledge yet only understood
the Microsoft world. I was making decent money, but had no time to
enjoy it. Though my boss was happy with me, I was losing motivation to just keep doing the same thing I had been doing for over a
decade. When I looked around the company I saw a lot of ancient
codebases and unprofitable ventures.
Like many of my fellow employees, I was only vaguely familiar
with free software when I left and randomly decided to check out
this thing called Linux. At Microsoft, I got all the software I wanted
for free, and I always thought free software would be behind proprietary software. For 15 years I had made it a priority to learn about
many aspects of Microsoft technologies, and my office contained
rows of books on everything from Undocumented Windows to Inside
SQL Server. When running Windows I felt as comfortable as Neo in
the Matrix, without the bullets and leather, so while I was willing to
look around, I was half-forcing myself and didn't want this little
experiment to mess up my main computing environment.
Every technical decision was big for me: which version of Linux
should I try? Should I get an extra machine or can I try dual-boot?
Can I really trust it to live on the same hard drive as Windows? I got
some tips and assurance from a Microsoft employee who had
recently tried Linux, and with that, and the help of Google, I proceeded with the installation of Red Hat's Fedora Core 3.
While I came to not be all that thrilled with Fedora itself, I was
floored merely by the installation process. It contained a graphical
installer that ran all the way to completion, it resized my NTFS partition which I considered a minor miracle, setup dual boot, and
actually did boot, and let me surf the Web. I didnt have a clue what
to do next, but the mere fact that this all worked told me more about
the potential of Linux than anything I had read so far. You cannot, by
accident, build an airplane that actually flies.
Over time, what impressed me the most about Linux was the
power of it all. It came with tons of applications: Firefox, OpenOffice, GIMP, Audacity, Mono, MySQL, and many more for me to discover. The UI was simple, responsive, polished and customizable.
Installing the Apache web server took just a few seconds and gave
me access to a vast world of PHP. Installing the WordPress blog took
me 15 minutes the first time, but I knew when I became more profi-
cient at things, I could do it in one. I came to understand that

beyond its poorly debugged device drivers, a Windows computer is a
sad joke. By mid-2005, I was in love with computers again!
I've spent three years in diligent research on the key subjects of
this book, talking to hundreds of programmers, attending many conferences, and reading source code, magazines, websites and books.
This book isn't really about the death of Microsoft as much as it is
about the Microsoft proprietary development model that has pervaded or even infected computing. I have absolutely zero bitterness
towards Microsoft although I now believe they are toast. I loved
working there, learned an enormous amount, and enjoyed the privilege of working alongside many brilliant minds. Like many things in
life, it was fun while it lasted.
Glossary
GLOSSARY
Bit: A piece of information that can hold 2 values: 1 and 0. Bits are grouped
into bytes of 8, characters of 2 bytes (Unicode), 4-byte numbers and pictures with lots.1
Digitize: Process of converting something into 1s and 0s. Once something
is in a digital format, it can be infinitely manipulated by a computer.
Software: General term used to describe a collection of computer programs, procedures and documentation that perform tasks on a computer.
Function: The basic building block of software is a function, which is a discrete piece of code which accomplishes a task:
int SquareNumber (int n)
{
return n * n;
}
Machine language: At the lowest level, software is a bunch of bits that
represent an ordered sequence of processor-specific instructions to change
the state of the computer.
High-level language: A programming language that looks more like English.
Compiler: Software that (typically) converts a high-level language into a
machine language.
Kernel: The lowest level of an operating system that initializes and manages hardware.
Hardware: Physical interconnections and devices required to store and run
software.
Processor: Hardware that executes the programmer's instructions.
Hard drive: Spinning magnetic platters where bits are stored even after
the computer is turned off.
Memory: Hardware which provides fast access to bits of code and data for
the processor. A processor can only manipulate data after it has loaded
them into memory from the hard drive or network.
URL (Uniform Resource Locater): The textual location of a webpage, picture, etc. on the Internet. You can hand a URL to any computer in the world
that understands the Internet and it would return the same thing. (It
might notice that you prefer a version of the page in your language.) An email address is also a URL. The only thing everything on the Internet has is
a URL.
Like a number of places in this book, some of this text was taken from Wikipedia.
10
Wikipedia
WIKIPEDIA
A good friend of mine teaches High School in Bed-Stuy, Brooklyn pretty much the hood. Try to imagine this classroom; it
involves a lot of true stereotypes. But what does NOT fit the
stereotype is that he started a class wiki, and has all his students contribute to it. Instead of a total mess, instead of abuse,
graffiti and sludge, it's raised the level of ALL the students. It's
a peer environment: once it becomes cool to do it right, to be
right, abuse and problems dry up almost completely.
Slashdot.org commentator
My school blocks Wikipedia entirely. When asked why, the
answer is anybody can edit it. As opposed to the rest of the
Internet which is chock-full of nothing but the highest quality,
peer-reviewed content, written universally by the finest
experts, hand selected from across the world?
One of the great movements in my lifetime among educated
people is the need to commit themselves to action. Most people
are not satisfied with giving money; we also feel we need to
work. That is why there is an enormous surge in the number of
unpaid staff, volunteers. The needs are not going to go away.
Business is not going to take up the slack, and government cannot.
Peter Drucker, father of modern management
Graph of the number of entries in the Wikipedias of various languages. This

exponential growth is a confirmation of Metcalfe's law: the more users of
Wikipedia, the better it gets, so more use it.
Wikipedia
11
Encyclopedia Britannica ad from 1913
Compared to a paper encyclopedia, a digital edition has significant advantages. The biggest is cost, as printing and shipping a
50,000-page document represents an enormous expense in the production of an encyclopedia. The digital realm has other significant
advantages: the content can be constantly updated and multimedia
features can be incorporated. Why read about the phases of a 4stroke internal combustion engine when you can watch one in
action?
In the mid-1990s, Microsoft created Encarta, the first CD-ROM
based digital encyclopedia. CDs were a natural evolution for Microsoft because it was shipping its ever-growing software on an
increasingly large number of floppy disks. (Windows NT 3.1,
released in 1993, required 22 floppies. CDs quickly became more
cost-effective, as they hold 500 times more data, and are more reliable and faster, and Microsoft played an important role in introducing CD-ROM drives as a standard feature of computers.)
While CDs hold more data than floppies and are an important
technological advancement, this development was soon eclipsed by
the arrival of the web. Users could connect to a constantly-updated
encyclopedia of unlimited size from any computer without installing
it first.
Unfortunately for Microsoft, the Encarta team was slow in adopting the Internet because they felt some of the richness of its encyclopedia was lost on the web. However, with pictures, animations
12
Wikipedia
and text, even the early web was good enough and had substantial
advantages over a CD-ROM version. In the Internet realm, you only
need one Wikipedia, albeit running on hundreds of servers, for the
entire world; you don't even need to worry about the cost to make a
copy of an encyclopedia.
However, the biggest mistake the Encarta team made was not
realizing that the Internet could introduce feedback loops. The users
of an Internet encyclopedia can also become enhancers of it. If I
have a question about what I've read, or I think I've found a problem, I can post a question or fix the problem and report what I've
accomplished.
We will discuss later if the ability for anyone to edit, enhance or
add data will hurt quality, but it is important to remember that it
was the creation of the Internet that allows people in all the corners
of the world to work together and learn from each other; a completely new capability for man.
For any faults, Wikipedia became larger than the Encyclopedia
Britannica in just 2.5 years. The database now contains more than
15 times as many articles, and is already the best compendium of
human knowledge ever created. No corporation invested millions of
dollars in engineering or marketing either; it happened seemingly
on its own. Even if some of those articles are fluff about Star Trek
characters, many are not: Wikipedia's article on carbon nanotubes
and many other scientific topics is more detailed and more up to
date than Encyclopedia Britannica's. The vast depth of Wikipedia is
also a living refutation of perhaps the biggest criticism of free software and free content: that no one will work on the most arcane and
boring stuff. The lesson here is that different things are interesting
to different people.
Wikipedia is one of the 10 most popular websites on the Internet,
receiving 450 times the traffic of Encyclopedia Britannica, and with
an article collection that continues to grow at an exponential rate.
As Wikipedia has advanced, it has also added a multimedia collection, a dictionary, a compendium of quotes, textbooks, and a news
aggregator and they are just getting started.
In some ways, access to a search engine might seem to obviate
the need for an encyclopedia. But while search engines provide a
keyword index to the Internet, they do not replace the importance of
an encyclopedia: a comprehensive, coherent, neutral, compendium
of human knowledge.
Wikipedia
13
Imagine that you wanted to research a topic like nuclear power.

Where would you go to get an unbiased opinion: the government?
Greenpeace? CNN? Some schools have banned Wikipedia, but secondary sources have long been disallowed. Even so, the article text
and links to primary sources can be a useful place to start research
on a topic.
While Wikipedia is a powerful resource, what is more amazing is
that it is built with the same surplus intellectual energy that others
spend on crossword puzzles or Sudoku. Wikipedia provides an additional outlet for people's energy, and something much greater than
any one person, or even one company, could accomplish.
A key element of Wikipedia's success is that its founders created
a community that people enjoyed working in, and this enjoyment
factor brought in even more people. This is a hard thing to do, and it
starts with an inspirational vision.
There is no monster multinational corporation behind Wikipedia.
There was no CEO who pounded the table and said he wanted to
create the biggest encyclopedia ever. Its annual budget is
$5,000,000, most of that goes to funding hardware, bandwidth and
the salary of the very tiny six-person operations team that keeps the
few hundred servers running. Maybe you haven't edited Wikipedia
yet, but millions of other registered members, and unregistered
users, have created and improved it over the years. I've made a few
fixes to a few places it is very easy!
Some may wonder about the susceptibility to inaccuracies and
vandalism of something as widely collaborative as Wikipedia. Fortunately, this digital graffiti does not threaten to spoil things for two
important reasons: accountability and pride. The good thing about
these shared efforts is that they provide a way to appreciate the
importance of one's fellow man.
Every change made to the encyclopedia is permanently logged
and publicly recorded in a version control mechanism similar to that
used in software; in fact, no changes are irreversible. Unlike graffiti,
which can take hours to clean up, dumping unwanted changes or
banning users takes mere seconds which is a great discouragement.
Ultimately, part of believing in the viability of a free encyclopedia
requires belief in the fundamental goodness of mankind. One must
trust that the amount of people in the world who gain satisfaction
from making a positive contribution to a product far outnumbers
14
Wikipedia
those who enjoy a few seconds of perverted pride in temporary

defacement. Moreover, with millions of registered users, there is a
virtual guarantee that problems will get noticed.
The obvious vandalism is easily caught and removed, but there
are more subtle forms of vandalism that are much harder to catch.
In fact, who is to say whether or not any edit is accurate?
Wikipedia has insulated its product from inaccuracies by implementing three content policies:
1. No original research: Articles should reference published,
reliable sources. The threshold for reliable is debatable, but
in practice, this is not a significant obstacle.
2. Neutral point of view: An article should fairly and, without
bias, represent all significant views that have been published by
reliable sources.
3. Verifiability: The threshold for inclusion in Wikipedia is
verifiability. Verifiable means that a reader should be able to
determine whether material added to Wikipedia has already
been published by a reliable source.
That the community accepts these concepts is a key to Wikipedia's success.
By making these policies an integral part of the culture,

Wikipedia created something not necessarily perfectly accurate, but
from reputable, verifiable resources which makes it good enough
that other people decided it is worth reading and contributing to.
There have been objective studies that have demonstrated that
Wikipedia is high quality, comparable to Encyclopedia Britannica. In
general, its greatest challenge is in political articles where emotions
run high, and most contributors would describe themselves as liberal. This is further complicated because many facts are disputed:
some scientists say Global Warming provides an imminent danger to
humanity, while others say it is a hoax, and Wikipedia cannot resolve
this contradiction between published, reliable sources.
Even for the cynics who believe that the vandals may still win,
consider that since its creation in January 2001, Wikipedia has
remained as much an encyclopedia as a self-organizing technological and social experiment. As Wikipedia evolves, tools are being created to detect and remove vandalism, and tag articles that don't
conform to style guidelines. Sometimes articles have various warnings about how it is a work in progress, which is a useful warning.
Every page also has a discussion page where issues are debated
before the content itself is updated.
Wikipedia
15
In short, Wikipedia is an evolving relationship between people

and their software. For example, should anonymous users be
allowed to make edits? Many believe they should not because
anonymity decreases accountability. This is an ongoing discussion.
Wikipedia is free to read, and a study suggested that it could generate up to $100 million per year in advertising revenue. One day,
they might choose to, and could use this money in any number of
ways: from acquiring proprietary content like maps, legal documents, and document templates, and making them free, to hiring
employees to charge ahead in areas underfunded by the community.
Eric Raymond, in his book The Cathedral and the Bazaar, analogizes the free software development model to a bazaar a disorganized conglomeration of input and ideas. That's an unsatisfactory
image, however, because it suggests something primitive and disorganized. Cathedrals took hundreds of years to build, but in less than
10 years, Wikipedia has produced a larger, and more comprehensive
product than its much older competitors. It is better to think of this
free software product as an already very polished cathedral in its
early years of development.
What else can independent, highly co-operative free software
communities build? The answer is a nearly infinite number of things.
Specifically, an entire software stack that is as free as Wikipedia and
uses zero Microsoft code. The software used to run Wikipedia is an
afterthought to many, but it is a significant piece of technology.
While Wikipedia and its software won't make a large dent in
Microsoft's profits, the Linux kernel is a mortal threat.
16
Linux
LINUX
Really, I'm not out to destroy Microsoft. That will just be a completely unintentional side effect.
Linus Torvalds, 2003
The Linux mascot, Tux, created by Larry Ewing
he kernel of an operating system (OS) is the central nervous

system of a computer. It is the first piece of software that the
computer executes, and it manages and mediates access to
the hardware. Every piece of hardware needs a corresponding kernel device driver, and you need all of your drivers working before
you can run any of your software. The kernel is the center of gravity
of a software community, and the battle between free software and
Windows is at its lowest level a battle between the Linux and Win-
Linux
17
dows kernels. Microsoft has said that it has bet the company on
Windows, and this is not an understatement! If the Windows kernel
loses to Linux, then Windows, and Microsoft, is also lost.1
The Linux kernel is not popular on desktops yet, but it is widely
used on servers and embedded devices because it supports thousands of devices and is reliable, clean, and fast. Those qualities are
even more impressive when you consider its size: printing out the
Linux kernel's 8,000,000 lines of code would create a stack of paper
30 feet tall! The Linux kernel represents 4,000 man-years of engineering and 80 different companies, and 3,000 programmers have
contributed to Linux over just the last couple of years.
That 30-foot stack of code is just the basic kernel. If you include a
media player, web browser, word processor, etc., the amount of free
software on a computer running Linux might be 10 times the kernel,
requiring 40,000 man-years and a printout as tall as a 30-story
building.
This 40 man-millennia even ignores the work of users reporting
bugs, writing documentation, creating artwork, translating strings,
and performing other non-coding tasks. The resulting Linux-based
free software stack is an effort that is comparable in complexity to
the Space Shuttle. We can argue about whether there are any motivations to write free software, but we can't argue it already exists!
One of the primary reasons I joined Microsoft was I believed their
Windows NT (New Technology) kernel, which is still alive in Windows Vista today, was going to dominate the brains of computers,
and eventually even robots. One of Bill Gates' greatest coups was
recognizing that the original Microsoft DOS kernel, the source of
most of its profits, and which became the Windows 9x kernel, was
not a noteworthy engineering effort. In 1988, Gates recruited David
Cutler from Digital Equipment Corporation, a veteran of ten operating systems, to design the product and lead the team to build the
Windows NT kernel, that was released as I joined in 1993.
While cloud computing, the movement of increasing number of applications and

services provided over the Internet, is one of the hot topics of today, it is unrelated to the Windows vs. Macintosh vs. Linux war that is going on. Even in a
future where applications like word-processing are done over the Internet, you
still need a kernel, a web browser, a media player, and so forth.
18
Linux
The kernel Cutler and his team developed looks like this:
50% of NT's code
Windows NT kernel architecture block diagram. Cutler had a Windows 95

doormat outside his office; you were encouraged to wipe your feet thoroughly before entering.
Unfortunately for Microsoft, the original kernel lived on through

Windows 95, Windows 98, and into Windows Me. (Microsoft also had
Windows CE, a small kernel for embedded devices. Microsoft had
three separate kernels for most of my tenure, whereas the same
Linux kernel is used on small and big devices.)
Windows has become somewhat popular for servers and devices,
but it never achieved the dominance it did on desktop PCs. Perhaps
the biggest reason is that its code wasn't available for others to
extend and improve upon. The Linux kernel took off because there
Linux
19
are people all over the world, from Sony to Cray, who tweaked it to
get it to run on their hardware. If Windows NT had been free from
the beginning, there would have been no reason to create Linux.
However, now that there is the free and powerful Linux kernel,
there is no longer any reason but inertia to use a proprietary kernel.
There are a number of reasons for the superiority of the Linux
kernel. But first, I want to describe the software development
process. When you understand how the Linux kernel is built, its
technical achievements are both more impressive and completely
logical.
20
Linux
Distributed Development
In Linux we reject lots of code, and that's the only way to create
a quality kernel. It's a bit like evolutionary selection: breathtakingly wasteful and incredibly efficient at the same time.
Ingo Molnar, Linux kernel developer
A portion of the speaker list for the 2006 Linux Kernel Symposium the
author attended. Linux kernel development is a distributed effort, which
greatly enhances its perspective.
Every 20th century management book I've read assumes that team
members work in the same building and speak the same language.
Linux
21
Microsoft's corporate culture was based on the theory that software

development was a collaborative effort that should be centralized so
that people could work together. As a result, most Microsofties,
especially the programmers, were based in Redmond because that
was where all the other engineers were located.
Microsoft had a very open development model inside the company: developers would periodically switch teams, collaborate freely
together on unreleased code, and join e-mail discussion groups with
engineers in similar product roles. These resources of collaboration
are some of the many which are unavailable to those outside.
A software program is basically a bunch of human-readable text
files, compiled together into a machine-specific binary. When just
one person is working on the codebase, things are easy to manage,
but when you have multiple people, you need three more tools:
1. A source control system which records all changes, analogous to the History tab that exists at the top of every
Wikipedia page. This provides many capabilities such as
going back to find out when and why a code change was
made.
2. A bug or issue tracker keeps a to-do list. This keeps track of
the problems over time, and as people leave.
3. A communications mechanism (in-person, email, chat or
forums) where programmers can discuss how to implement
features, and work together to troubleshoot problems.
Larger teams have developed more formal processes to, for example, get consensus on when someone is given permission to make
changes to source code, but the basic development process in the
free software world is little different than what takes place at Microsoft.
The Internet, which was born when Microsoft was a mature company, has changed countless aspects of our lives, including the way
software can be developed. Without the Internet, free software
could never exist because the developers wouldn't be able to work
together. (Microsoft uses the Internet to do development in other
places, although it is still primarily in Redmond.) Microsoft grew up
before the birth of the web, and thus has yet to fully embrace this
distributed process. For example, Microsoft's bug databases aren't
available on the Internet. Microsoft isn't taking the maximum benefit from the knowledge gained by its users because it doesn't have
as many feedback loops.
22
Linux
Linux has achieved enormous gains against Windows, even while,

from a 20th century manager's perspective, organizing the Linux kernel is a worst-case example for building a productive, prosperous
organization. Where are the team-building exercises? The three-year
planning retreats? It should be amazing that Linux releases anything at all, let alone dominates the supercomputer business.
Linux succeeds in spite of the cost of working remotely for several reasons. First, the distance between programmers forces people to formalize things, which helps quality: people get together at
conferences like the Linux Kernel Symposium and present papers
and get consensus and feedback on ideas from a wide variety of people. In addition, the ability to have anyone from around the world
contribute is a greater benefit than the cost of putting people
together. You couldn't create a building large enough to host the millions of Wikipedia contributors, let alone the thousands of Linux kernel contributors.
Ingo Molnar's quote above appears contradictory at first glance,
but it is not. Linux receives lots of different ideas; many get
rejected, but what remains incorporates the best of all the ideas.
The Internet allows you to quickly evolve towards optimal solutions
with feedback loops of discussions, tests, and bug reports. (Tests are
very important because they give you objective numbers. If you
want to see if the new disk cache code is faster, you can compile the
kernel, which is a very disk-intensive task, and time the result.)
In the free software movement, the battles aren't between
empires, but rather between engineers fighting over technical
details testosterone-laden VIPs are irrelevant. The Linux kernel
community has taken the idea of a meritocracy to the next level. All
changes to the official Linux kernel must go through Linus, and his
Lieutenant, Andrew Morton, and then the relevant subsystem maintainer but first, the proposed change has to go through everyone
else! All Linux changes are posted to a mailing list where anyone
can comment and give opinions. Linus wrote:
The contributors for any given project are self-selected. Someone pointed out that contributions are received not from a random sample, but from people who are interested enough to use
the software, learn about how it works, attempt to find solutions to problems they encounter, and actually produce an
apparently reasonable fix. Anyone who passes all these filters is
highly likely to have something useful to contribute.
Linux
23
Linus' primary job is to provide technical expertise. He once said

his job was to keep out bad code, and that might be enough. Let the
varied users of Linux take it to new places, while he will make sure
no one is screwing up the existing code along the way.
Linus was asked whether the unpolished nature of a large group
of programmers, with disparate backgrounds, created a dreary situation that made him want to go back and work in private, and he
said:
I actually like arguing (sometimes a bit too much), so the occasional flame-fest really does nothing but get me pumped up.
At the same time, I'm actually pretty good at just letting it go,
once I've argued enough and am bored with the argument. Part
of that is also having to occasionally just admit that you were
wrong, and have the ability to send out a mea culpa e-mail
just saying so.
I tend to care much more about improving the general development model than about the details of some particular subsystem. So that tends to make it easier for me to let go. I'll state
my opinions, but even if I'm convinced I'm right, if I'm not actually willing to write the code, in the end I'll happily be overridden by the people who do write the code.
This is obviously very much a matter of personality.
There are things that I tend to worry about, and that can be
really painful, but they are pretty rare. The classic example is
the old Linus doesn't scale argument, where it ended up
being the case that I really had to fundamentally change the
tools I use and how I work. And that was a lot more painful than
worrying about the actual code.
24
Linux
Linux Kernel Superiority

Here are the reasons Linux is superior to the Windows kernel:
1. Refactored Code (Reliability)

Here is a diagram of the Linux kernel:
Device
Drivers
Arch
(CPU-specific code)
Security
Network &
file
systems
Init &
Memory
Manager
Crypto
Layers of the Linux kernel onion. The Linux kernel is 50% device drivers,
and 25% CPU-specific code. The two inner layers are very generic.
Notice that it is built as an onion and is comprised of many discrete components. The outermost layer of the diagram is device
drivers, which is 50% of the code, and more than 75% of its code is
hardware-specific. The Microsoft Windows NT kernel diagram,
shown several pages back, puts all the device drivers into a little
box in the lower left-hand corner, illustrating the difference between
Linux
25
theory and reality. In fact, if Microsoft had drawn the kernel mode
drivers box as 50% of the Windows NT diagram, they might have
understood how a kernel is mostly hardware-specific code, and
reconsidered whether it was a business they wanted to get into.
Refactoring (smoothing, refining, simplifying, polishing) is done
continuously in Linux. If many drivers have similar tasks, duplicate
logic can be pulled out and put into a new subsystem that can then
be used by all drivers. In many cases, it isn't clear until a lot of code
is written, that this new subsystem is even worthwhile. There are a
number of components in the Linux kernel that evolved out of duplicate logic in multiple places. This flexible but practical approach to
writing software has led Linus Torvalds to describe Linux as Evolution, not Intelligent Design.
One could argue that evolution is a sign of bad design, but evolution of Linux only happens when there is a need unmet by the current software. Linux initially supported only the Intel 80386
processor because that was what Linus owned. Linux evolved, via
the work of many programmers, to support additional processors
more than Windows, and more than any other operating system ever
has.
There is also a virtuous cycle here: the more code gets refactored,
the less likely it is that a code change will cause a regression; the
more code changes don't cause regressions, the more code can be
refactored. You can think about this virtuous cycle two different
ways: clean code will lead to even cleaner code, and the cleaner the
code, the easier it is for the system to evolve, yet still be stable.
Andrew Morton has said that the Linux codebase is steadily improving in quality, even as it has tripled in size.
Greg Kroah-Hartman, maintainer of the USB subsystem in Linux,
has told me that as USB hardware design has evolved from version
1.0 to 1.1 to 2.0 over the last decade, the device drivers and internal
kernel architecture have also dramatically changed. Because all of
the drivers live within the kernel, when the architecture is altered to
support the new hardware requirements, the drivers can be
adjusted at the same time.
Microsoft doesn't have a single tree with all the device drivers.
Because many hardware companies have their own drivers floating
around, Microsoft is obligated to keep the old architecture around
so that old code will still run. This increases the size and complexity
of the Windows kernel, slows down its development, and in some
cases reveals bugs or design flaws that can't even be fixed. These
26
Linux
backward compatibility constraints are one of the biggest reasons

Windows takes years to ship. The problem exists not just at the
driver layer, but up the entire software stack. When code isn't freely
available and in one place, it makes it hard to evolve.
While the internal logic of Linux has evolved a lot in the last ten
years, the external programmer interfaces have remained constant.
The key to a stable interface is incorporating the right abstractions.
One of the best abstractions that Linux adopted from Unix is the file
abstraction. In order to perform almost any function on a Linux computer, from reading a web page on a remote website to downloading
a picture from a camera, it is necessary to simply use the standard
file commands: open and close, read and write.
On my computer, in order to read the temperature of the CPU, I
just need to open the (virtual) text file
/proc/acpi/thermal_zone/THM0/temperature and the data I
request is inside:2
temperature:
49 C
In essence, the Linux kernel is a bundle of device drivers that

communicate with hardware and reveal themselves as a file system.
As new features, security issues, hardware requirements and scenarios confront the Linux kernel, the internal design evolves and
improves, but the file system abstraction allows code outside the
kernel to remain unchanged over longer periods of time.
Here is a random sample of the change log of the Linux kernel
from 2.6.14. As you can see, it is filled with all kinds of cleanup and
bugfix work:
spinlock consolidation
fix numa caused compile warnings
ntfs build fix
i8042 - use kzalloc instead of kcalloc
clean up whitespace and formatting in drivers/char/keyboard.c
s3c2410_wdt.c-state_warning.patch
[SCSI] Fix SCSI module removal/device add race
[SCSI] qla2xxx: use wwn_to_u64() transport helper
[SPARC64]: Fix mask formation in tomatillo_wsync_handler()
[ARCNET]: Fix return value from arcnet_send_packet().
Many of the Linux kernel's code changes are polish and cleanup. Clean
code is more reliable and maintainable, and reflects the pride of the free
software community.
2
This should arguably be expressed as XML, but because there is common code
that reads these values and provides them to applications, and because each file
contains only one value, this problem isn't very significant; the kernel's configuration information will never be a part of a web mashup.
Linux
27
If you look at the code changes required to make a bugfix, in the

vast majority of cases all that is needed is a revision of a few lines of
code in a small number of files. A general guideline Linux has for
bugfixes is this: if you can't look at the code change and prove to
yourself that it fixes the problem, then perhaps the underlying code
is confused, and this fix shouldn't be added near the end of a release
cycle.
According to Stanford University researchers, the Linux kernel
has .17 bugs per 1,000 lines of code, 150 times less than average
commercial code containing 20-30 bugs per 1,000 lines.3 Microsoft's
Windows bug databases aren't available on the Internet so it is
impossible to make comparisons, but even if Linux isn't more reliable already, it is setup to become so because the code is simple,
well-factored, and all in one place.
Within the free software community, different teams are disparate
entities, and so the idea of arbitrarily moving code from one part of
the system to another can't easily happen. Inside Microsoft there
are no boundaries, and so code is moved around for short-term performance gains at the cost of extra complexity.
These studies have limited value because their tools usually analyze just a few
types of coding errors. Then, they make the IT news, and get fixed quickly
because of the publicity, which then makes the study meaningless. However, these
tools do allow for comparisons between codebases. I believe the best analysis of
the number of Linux bugs is the 1,400 bugs in its bug database, which for 8.2 million lines of code is .17 bugs per 1,000 lines of code. This is a tiny number, though
it could easily be another 100 times smaller. Here is a link to the Linux kernel's
active bugs: http://tinyurl.com/LinuxBugs.
28
Linux
Here is a graph of all the function calls into the OS required to

return a simple web request. These pictures demonstrate a visual
difference in complexity that often exists between free and proprietary software:
System call graph in Microsoft's proprietary web server, IIS.
Diagrams provided by SanaSecurity.com
System call graph to return a picture in the free web server Apache.
Linux
29
2. Uniform Codebase (Reliability, Maintainability, and Usability)

My job is to say no, to some extent. If you keep things clean
enough and have some kind of requirement for what code can
look like, you're already ahead of the game.
Linus Torvalds
Linux engineers have found a way to run the same codebase on a

wide variety of processors, on devices from cellphones to supercomputers, an unprecedented achievement. Linux has been tuned to
first, run correctly, and then run efficiently on two, four, eight, and
now 1,000 processor machines. Software has infinite malleability, so
such a universal kernel has always been possible it just took a
bunch of different hardware companies working together to make it
happen.
Putting everything into one codebase helps reliability. Running
the Linux kernel on a 32-processor computer shakes out multithreaded bugs far more quickly than on a two-processor laptop.
Running on low-end machines keeps the code small and simple,
which makes it run faster on desktops. Features that first appear on
laptops and tablets eventually trickle their way down to even
smaller devices where the code undergoes even more testing and
enhancement. The many hardware and server developers who want
extreme reliability ensure that the kernel on my PC is as reliable as
Linux's most demanding customer.
Linux is more flexible than the Windows NT kernel, though both
are very clean and flexible. For example, the National Security
Agency (NSA) has created a free software component called
SELinux (Security Enhancements to Linux) that adds a strong security enforcement mechanism known as Mandatory Access Control.4
Making these mechanisms public helps ensure there are no back
This is a way to add additional security because the operating system can say, for
example: Because a media player has no reason to write files to disk, the system
can take away this permission. Before the kernel tries to do anything interesting,
it will ask the Mandatory Access System (MAC) whether such an operation is
allowed. The security checks in most other operating systems simply ask if the
person is allowed to do something.
Creating a default policy adds additional work for application writers, and by
itself doesn't entirely solve the problem. A word processor needs complete read
and write access, so how do you solve the problem of a virus in a document macro
opening all of your files and writing junk? SELinux doesn't deal with this situation
because it doesn't have this information. In GC programming languages, it is possible to walk the stack and determine more information about whether a macro,
or the word processor itself, is asking to open a file.
30
Linux
doors to the NSA's computers. I will discuss in a later chapter why

governments can adopt free software, even for high-security scenarios, but it appears the NSA already understands this concept.
Throughout the Linux world, one has many more choices for just
the right tool to do the job. Some might argue that too much choice
is a bad thing, but creating many components forces clear boundaries, and survival of the fittest whittles down the inferior choices
over time.
3. Frequent Ship Cycles (Maintainability

and Usability)
Microsoft had a motto: Ship early, ship often. This philosophy is
wise for software development because it forces teams to maintain a
high-quality product every day, and the earlier you release, the
sooner you can receive and incorporate feedback.
However, this philosophy only works when adopted. Unfortunately, Microsoft's two biggest products, Windows and Office, do not
follow this philosophy. Of course, paying $240 every year for the latest upgrade to Windows Ultimate wouldn't be acceptable either!
The Linux kernel ships every three months. For a product of its
size and complexity, Linux's rate of shipment is unprecedented. This
pace has allowed the Linux kernel to ship drivers before Windows,
and even before the hardware itself has shipped. Linux supported
USB 3.0 before Microsoft, which did not include it in Windows 7.
Because Linux is constantly near shipping, you can take any random
build from Linus' computer, put it into a rocket and feel quite confident that it won't crash.
A big part of the Department of Justice lawsuit against Microsoft
focused on the company's bundling of many software components
with their operating system. The government accused Microsoft of
excluding third-party software developers and stifling competition.
But Microsoft's tying has been both a blessing and a curse.
The blessing is that by having the pieces work together, they can
reuse code and be more integrated. The curse is that Microsoft has
created a situation whereby it revises and adds new features to all
of its interdependent components at the same time. As a consequence, its components take years to stabilize, and you can't ship
until the last component is ready.5
5
The alternative is for each component to use the previous version of all of its
dependent components, which means that the features in the latest Internet
Explorer wouldn't show up in various places that are using the old version. However, does the Help system need the latest version?
Linux
31
By contrast, in a free OS, software components depend only on

released versions. Every team doesn't try to ship on the same day, so
the OS contains the latest versions of all released components.6 Separate development organizations have enforced boundaries that
have simplified dependencies, and allows everyone to move ahead at
whatever pace they choose.
Many users wonder whether free software will ever be as good as
proprietary software because presumably the free software guys
can't afford things like usability studies. Having watched usability
studies at Microsoft, grabbing people off the street to give their bits
of feedback in a room with two-way mirrors is not necessary. The
Internet, and all its communications mechanisms, provides a continuous and richer feedback mechanism than you can get with any
usability study. In addition, usability studies don't matter if you can't
incorporate the changes quickly and easily. With clean codebases
and frequent ship cycles, usability will happen automatically. I've
spent a lot of time using Linux and find many applications are perfectly usable.
Shipping a new platform every five years in theory provides
Microsoft's partners a stable platform upon which to build. However, in reality, its results have flaws. For instance, I wasn't able to
install an HP Photodesk 7960 printer driver on Windows Server
2003, although the driver installed perfectly on XP. The subtle and
unintentionally undocumented differences between those operating
systems, which shipped two years apart, has created compatibility
headaches even on Microsoft's supposedly uniform platform.
4. Lower Development Costs (Maintainability)

It is much less expensive for hardware vendors to support Linux.
If you want to build a device driver, a great place to start is by looking at existing shipping device drivers, an opportunity that Linux
offers to everyone. A proprietary Device Driver Toolkit with its
sample code is never as good as production code. Those expensive
kits contain documentation, but not source code so you sometimes have to guess at what is happening down below.
We find in Windows today that hardware manufacturers have
duplicated a bunch of the functionality Windows provides but doesn't quite fit their needs. For example, IBM includes its own applet
and status icon for wireless Internet, so Windows XP on IBM hard6
Some components contain multiple versions to allow for a transition period.
32
Linux
ware has two. Presumably they weren't satisfied with the features
Windows provided, and weren't able to fix them. And so they had to
build new applets from scratch! This is also what gives Windows a
feeling of a jumble of components slapped together.
Here are five of the 100 applets IBM adds to Windows:
Windows XP with 5 of IBM's 100 extra applets. Notice the large number
of status icons on this almost-virgin installation.
Building all of these applets, designing multilingual user interfaces, providing the means to install and configure, etc. is ten times
more work than merely writing the device driver, leveraging other
shipping drivers, and uploading it to the official codebase.
The reason my Photodesk printer driver didn't work on Windows
Server 2003 was because of a crash in the installation code which
HP shouldn't even be bothering with in the first place.
Linux
33
5. Security (Reliability and Maintainability)

To mess up a Linux box, you need to work at it; to mess up your
Windows box, you just need to work on it.
Scott Granneman
Attempting to compare the security of operating systems is a

complicated endeavor because there are an infinite number of risks.
It is like asking whether one car is safer than another. Against what:
Getting hit from the front? Rolling over? Having the gas tank
pierced? Its ability to brake?
Furthermore, neither Windows nor Linux are perfectly secure
operating systems; both are prone to buffer-overrun viruses, an
issue I will discuss in the tools chapter. Furthermore, new threats
appear over time, and so each nefarious advancement requires new
logic to defend against it. Given these caveats, it is still possible to
make some comparisons.
Some of the previous advantages of Linux, such as its simplicity,
modularity, support for SELinux, etc. all help with its security. In
addition, the nature of how Linux is developed also helps. A document commissioned by the US Military said:
Open source software is potentially subject to scrutiny by
many eyes
Therefore bugs, security flaws, and poor design cannot hide for
long, at least when the software has a community of programmers to support it. And since fixing the code doesn't depend on
a single vendor, patches are often distributed much more
rapidly than patches to closed source software.
Can increase code quality and security
With closed source software, it's often difficult to evaluate the
quality and security of the code. In addition, closed source software companies have an incentive to delay announcing security
flaws or bugs in their product. Often this means that their customers don't learn of security flaws until weeks or months after
the security exploit was known internally.
Open Technology Development Roadmap
Another big difference between Linux and Windows is that Linux

was adapted from Unix, which had a multiuser design right from the
beginning. In Windows, users have historically had full Administrator access to the machine, including the ability to overwrite system
files. When the computer is attacked by a virus, the virus can gain
the same capabilities as the user and thereby hide itself inside sys-
34
Linux
tem files, which make it very difficult to remove. On Linux, I can

write only to my own files and have read-only access to all others.
Linux's multiuser nature, and therefore its focus on minimal permissions, minimizes any damage.
One study found that there are about 60,000 known viruses for
Windows, and only about 40 for Linux. Another study by Evans Data
showed that 8% of Linux developers say their machines have been
infected by malicious code, compared to 60% of Windows machines.
Brian Krebs of the Washington Post found that code to exploit
flaws in Internet Explorer 6 existed for 284 days in 2006, while Firefox was at risk for just 9. Computer security expert Bruce Schneier
recommended in December 2004 that people not run Internet
Explorer. Some argue that Linux and Firefox have fewer viruses
because they have fewer users, but Apache is well-respected for its
security.
For most of my tenure at Microsoft, we worried about features,
reliability, and performance, not security. Microsoft's Chief Research
and Strategy Officer, Craig Mundie, said in 2002:
Many of the products we designed in the past have been less
secure than they could have been because we were designing
with features in mind rather than security.
Microsoft has greatly increased its focus on security in the past

few years, and I am sure the security of every product improves with
every release, but the baggage of their codebases serve as an ongoing impediment.
Having browsed through the sources to a number of Linux applications, one can say the free codebases are typically cleaner than
their Windows counterparts, which makes them easier to secure.
The default server-oriented Debian 4.0 Linux distribution requires a
mere 170MB of disk space, whereas Windows Server 2003 requires
3 GB. All other things being equal, the fact that Debian is 17 times
smaller means it will be more secure. The free database MySQL is a
mere 26MB download; the code is clean and small, and therefore
much more likely to be reliable and secure.
Another advantage of Linux is that all the applications in a Linux
operating system receive security updates. In the Microsoft world,
only Microsoft's code is protected by Windows Update.
While Linux is free to acquire, it can also be cheaper to run and
maintain than Windows because of its better security. The city of
Manchester in England spent $2 million in 2009 to remove the Conficker worm from their computers.
Linux
35
6. Linux has learned from Windows

While the Windows NT kernel was state of the art at the time it
was released in 1993, most of its good ideas have been learned well
and absorbed, in spite of the fact that the code has never been
released.
For example, the Linux kernel supports asynchronous I/O
(input/output), an innovative way to do reads and writes without
tying up thread resources. This was an innovation first made widespread in Windows NT.
The ability to load code dynamically is another important feature
the Linux kernel adopted from NT and others. Plug and play and
suspend and hibernate was a collaboration between Microsoft and
hardware companies in the 1990s, and Linux now supports this feature.
Throughout the free software stack, developers have incorporated
good ideas from the outside world. There is no Not Invented Here
syndrome in free software; a good idea is a good idea, and existing
code is even better. In software today, the biggest impediment to
sharing ideas is not ego, but license agreements.
The Linux kernel has even learned from Microsoft's mistakes. For
example, one feature added to the Windows NT 4.0 kernel was to
put the code that draws widgets into the kernel itself. While this can
improve graphics performance, it also means that a bug in the code
of a button has the capacity to crash the entire system. The best way
to keep a system secure and reliable is to keep as much code as possible in user mode above the kernel, and Linux follows this strategy.7
The Feature Race

One of Microsoft's best arguments against free software over the
years has been to tout their new features and use that as proof
that free software will always be behind proprietary software. The
moment a Microsoft product comes out, it always offers features
that no one else has. However, most of the features are in fact ones
that others do have, and Microsoft is catching up. Nearly every new
feature is an evolutionary step forward in a place that Microsoft felt
needed work.
But like Microsoft, each team in the free software community is
improving their code every day. In fact, because Microsoft takes so
7
In Windows Vista, Microsoft moved some of the device drivers to user mode but
they should have kept the device drivers small, simple, and in the kernel and
instead moved the widgets and fluff to user mode.
36
Linux
long to ship, the free software community has often added features
before Microsoft. The website http://kernelnewbies.org displays the
latest list of the Linux features added since the previous release 3-4
months before, and it is typically 15 pages long! For example, here
is just the list of driver features added to the 2.6.26 version of the
Linux kernel, which had a 3-month dev cycle.
Linus 2.6.26 driver workitems
4.1. IDE/SATA
IDE
Add warm-plug support for IDE devices
Mark "idebus=" kernel parameter as obsoleted (take 2)
Remove ide=reverse IDE core
Add "vlb|pci_clock=" parameter
Add "noacpi" / "acpigtf" / "acpionboot" parameters
Add "cdrom=" and "chs=" parameters
Add "nodma|noflush|noprobe|nowerr=" parameters
Add Intel SCH PATA driver
Add ide-4drives host driver (take 3)
gayle: add "doubler" parameter
Remove the broken ETRAX_IDE driver
SATA
sata_inic162x: add cardbus support
libata: prefer hardreset
ata: SWNCQ should be enabled by default
Make SFF support optional
libata: make PMP support optional
sata_mv: disable hotplug for now, enable NCQ on SOC, add basic
port multiplier support
sata_fsl: Fix broken driver, add port multiplier (PMP) support
4.2. Networking
ssb: add a new Gigabit Ethernet driver to the ssb core
Add new qeth device driver,
Add new ctcm driver that reemplaces the old ctc one,
New driver "sfc" for Solarstorm SFC4000 controller.
Driver for IXP4xx built-in Ethernet ports
Add support the Korina (IDT RC32434) Ethernet MAC
iwlwifi: Support the HT (802.11n) improvements,,,, add default
WEP key host command, add 1X HW WEP support, add default
WEP HW encryption, use HW acceleration decryption by default,
hook iwlwifi with Linux rfkill, add TX/RX statistics to driver, add
debugfs to iwl core, enables HW TKIP encryption, add led
support, enables RX TKIP decryption in HW, remove
IWL{4965,3945}_QOS
ath5k: Add RF2413 srev values, add RF2413 initial settings,
identify RF2413 and deal with PHY_SPENDING, more RF2413
stuff, port to new bitrate/channel API, use software encryption for
now
pasemi_mac: jumbo frame support, enable GSO by default, basic
ethtool support, netpoll support
rt2x00: Add per-interface structure, enable master and adhoc
mode again, enable LED class support for rt2500usb/rt73usb
e1000e: Add interrupt moderation run-time ethtool interface, add
support for BM PHYs on ICH9
niu: Add support for Neptune FEM/NEM cards for C10 server
blades, add Support for Sun ATCA Blade Server.
gianfar: Support NAPI for TX Frames
ehea: Add DLPAR memory remove support
sfc: Add TSO support
b43: Add QOS support, add HostFlags HI support, use SSB blockI/O to do PIO
S2io: Multiqueue network device support implementation,, enable
multi ring support, added napi support when MSIX is enabled.
ixgbe: Introduce MSI-X queue vector code, introduce Multiqueue
TX, add optional DCA infrastructure, introduce adaptive interrupt
moderation
uli526x: add support for netpoll
fmvj18x_cs: add NextCom NC5310 rev B support
zd1211rw: support for mesh interface and beaconing
libertas: implement SSID scanning for SIOCSIWSCAN
ethtool: Add support for large eeproms
The scheduled bcm43xx removal
4.6. Video
cx88: Add support for the Dvico PCI Nano, add xc2028/3028
boards, add support for tuner-xc3028
saa7134: add support for the MSI TV@nywhere A/D v1.1 card,
add support for the Creatix CTX953_V.1.4.3 Hybrid
saa717x: add new audio/video decoder i2c driver
Support DVB-T tuning on the DViCO FusionHDTV DVB-T Pro
Add support for xc3028-based boards
ivtv: add support for Japanese variant of the Adaptec AVC-2410
Add basic support for Prolink Pixelview MPEG 8000GT
bttv: added support for Kozumi KTV-01C card
Add support for Kworld ATSC 120
CX24123: preparing support for CX24113 tuner
Added support for Terratec Cinergy T USB XXS
budget: Add support for Fujitsu Siemens DVB-T Activy Budget
Support for DVB-S demod PN1010 (clone of S5H1420) added
Added support for SkyStar2 rev2.7 and ITD1000 DVB-S tuner
em28xx-dvb: Add support for HVR950, add support for the HVR900
Add support for Hauppauge HVR950Q/HVR850/FusioHDTV7-USB
HVR950Q Hauppauge eeprom support
Adding support for the NXP TDA10048HN DVB OFDM
demodulator
Add support for the Hauppauge HVR-1200
pvrusb2-dvb: add DVB-T support for Hauppauge pvrusb2 model
73xxx
Add support for Beholder BeholdTV H6
cx18: new driver for the Conexant CX23418 MPEG encoder chip
s5h1411: Adding support for this ATSC/QAM demodulator
4.7. SCSI
zfcp: Add trace records for recovery thread and its queues, add
traces for state changes., trace all triggers of error recovery
activity,register new recovery trace., remove obsolete erp_dbf
trace, add trace records for recovery actions.
qla2xxx: Add support for host supported speeds FC transport
attribute., add FC-transport Asynchronous Event Notification
support., add hardware trace-logging support., add Flash
Descriptor Table layout support., add ISP84XX support., add
midlayer target/device reset support.
iscsi: extended cdb support, bidi support at the generic libiscsi
level, bidi support for iscsi_tcp
scsi_debug: support large non-fake virtual disk
gdth: convert to PCI hotplug API
st: add option to use SILI in variable block reads
megaraid_sas: Add the new controller(1078DE) support to the
driver
m68k: new mac_esp scsi driver
bsg: add large command support
Add support for variable length extended commands
aacraid: Add Power Management support
dpt_i2o: 64 bit support, sysfs
Firmware: add iSCSI iBFT Support
4.8. WATCHDOG
Add a watchdog driver based on the CS5535/CS5536 MFGPT
timers
Add ICH9DO into the iTCO_wdt.c driver
4.9. HWMON
thermal: add hwmon sysfs I/F
ibmaem: new driver for power/energy/temp meters in IBM System
X hardware
i5k_amb: support Intel 5400 chipset
4.10. USB
ISP1760 HCD driver
Linux
The scheduled ieee80211 softmac removal
The scheduled rc80211-simple.c removal
Remove obsolete driver sk98lin
Remove the obsolete xircom_tulip_cb driver
4.3. Graphics
radeon: Initial r500 support,,
intel_agp: Add support for Intel 4 series chipsets
i915: Add support for Intel series 4 chipsets
Add support for Radeon Mobility 9000 chipset
fb: add support for foreign endianness
pxafb: preliminary smart panel interface support,
Driver for Freescale 8610 and 5121 DIU
intelfb: add support for the Intel Integrated Graphics Controller
965G/965GM
Add support for Blackfin/Linux logo for framebuffer console
4.4. Sound
hda-codec - Allow multiple SPDIF devices, add SI HDMI codec
support, add support for the OQO Model 2, add support of Zepto
laptops, support RV7xx HDMI Audio, add model=mobile for
AD1884A & co, add support of AD1883/1884A/1984A/1984B, add
model for cx20549 to support laptop HP530, add model for alc883
to support FUJITSU Pi2515, add support for Toshiba Equium L30,
Map 3stack-6ch-dig ALC662 model for Asus P5GC-MX, support of
Lenovo Thinkpad X300, add Quanta IL1 ALC267 model, add
support of AD1989A/AD1989B, add model for alc262 to support
Lenovo 3000, add model for ASUS P5K-E/WIFI-AP, added support
for Foxconn P35AX-S mainboard, add drivers for the Texas
Instruments OMAP processors, add support of Medion RIM 2150,
support IDT 92HD206 codec
ice1724 - Enable AK4114 support for Audiophile192
ice1712: Added support for Delta1010E (newer revisions of
Delta1010), added support for M-Audio Delta 66E, add Terrasoniq
TS88 support
Davinci ASoC support
intel8x0 - Add support of 8 channel sound
ASoC: WM9713 driver
Emagic Audiowerk 2 ALSA driver.
Add PC-speaker sound driver
oxygen: add monitor controls
virtuoso: add Xonar DX support
soc - Support PXA3xx AC97
pxa2xx-ac97: Support PXA3xx AC97
4.5. Input
Add support for WM97xx family touchscreens
WM97xx - add chip driver for WM9705 touchscreen, add chip
driver for WM9712 touchscreen, add chip driver for WM97123
touchscreen, add support for streaming mode on Mainstone
wacom: add support for Cintiq 20WSX
xpad: add support for wireless xbox360 controllers
Add PS/2 serio driver for AVR32 devices
aiptek: add support for Genius G-PEN 560 tablet
Add Zhen Hua driver
HID: force feedback driver for Logitech Rumblepad 2, Logitech
diNovo Mini pad support
4.6. Video
V4L2 soc_camera driver for PXA270,,
Add support for the MT9M001 camera
Add support for the MT9V022 camera
Add support for the ISL6405 dual LNB supply chip
Initial DVB-S support for MD8800 /CTX948
cx23885: Add support for the Hauppauge HVR1400, add generic
cx23417 hardware encoder support
Add mxl5505s driver for MaxiLinear 5505 chipsets, basic digital
support.
37
pxa27x_udc driver
CDC WDM driver
Add Cypress c67x00 OTG controller core driver,,
Add HP hs2300 Broadband Wireless Module to sierra.c
Partial USB embedded host support
Add usb-serial spcp8x5 driver
r8a66597-hcd: Add support for SH7366 USB host
Add Zoom Telephonics Model 3095F V.92 USB Mini External
modem to cdc-acm
Support for the ET502HS HDSPA modem
atmel_usba_udc: Add support for AT91CAP9 UDPHS
4.11. FireWire
release notes at linux1394-user
4.12. Infiniband
IPoIB: Use checksum offload support if available, add LSO
support, add basic ethtool support, support modifying IPoIB CQ
event moderation, handle 4K IB MTU for UD (datagram) mode
ipath: Enable 4KB MTU, add code to support multiple link speeds
and widths, EEPROM support for 7220 devices, robustness
improvements, cleanup, add support for IBTA 1.2 Heartbeat
Add support for IBA7220,,,,,,,,,
mthca: Add checksum offload support
mlx4: Add checksum offload support, add IPoIB LSO support to
mlx4,
RDMA/cxgb3: Support peer-2-peer connection setup
4.13. ACPI and Power Management
ACPICA: Disassembler support for new ACPI tables
eeepc-laptop: add base driver, add backlight, add hwmon fan
control
thinkpad-acpi: add sysfs led class support for thinklight (v3.1),
add sysfs led class support to thinkpad leds (v3.2)
Remove legacy PM
4.14. MTD
m25p80: add FAST_READ access support to M25Pxx, add Support
for ATMEL AT25DF641 64-Megabit SPI Flash
JEDEC: add support for the ST M29W400DB flash chip
NAND: support for pxa3xx
NOR: Add JEDEC support for the SST 36VF3203 flash chip
NAND: FSL UPM NAND driver
AR7 mtd partition map
NAND: S3C2410 Large page NAND support
NAND: Hardware ECC controller on at91sam9263 / at91sam9260
4.15. I2C
Add support for device alias names
Convert most new-style drivers to use module aliasing
Renesas SH7760 I2C master driver
New driver for the SuperH Mobile I2C bus controller
Convert remaining new-style drivers to use module aliasing
4.16. Various
MMC: OMAP: Add back cover switch support
MMC: OMAP: Introduce new multislot structure and change
driver to use it
mmc: mmc host test driver
4981/1: [KS8695] Simple LED driver
leds: Add mail LED support for "Clevo D400P"
leds: Add support to leds with readable status
leds: Add new driver for the LEDs on the Freecom FSG-3
RAPIDIO:
Add RapidIO multi mport support
Add OF-tree support to RapidIO controller driver
Add serial RapidIO controller support, which includes MPC8548,
MPC8641
edac: new support for Intel 3100 chipset
Basic braille screen reader support
ntp: support for TAI
RTC: Ramtron FM3130 RTC support
Don't worry if you don't understand what these things mean as I don't
either. It is just important to understand that even the hardware of computers are too big and complicated for one company to oversee the development of.
38
Linux
This is just a portion of the code changes for that version of

Linux, and doesn't include work in the file systems, networking, performance, architecture-specific work, and so forth.
Free software has added countless features before Microsoft.
However, many of these features are so obscure that your eyes
would glaze over reading the list. Anyone who focuses on the highlytouted features in Microsoft's latest release is missing the larger
picture: even when these features are compelling, they often make
you more deeply tied into the proprietary world and make it harder
to migrate away.
Microsoft's new software always has new features, but usually
only one piece of code for a scenario. Linux supports 60 file systems,
including several clustering solutions, that allow a seamless scaleout of disk storage amongst a group of servers. Microsoft supports
just one, which is tied very tightly to its system, which in turn limits
its flexibility. Free software also approaches solutions from different
directions. There are several ways to attack virtualization and Linux
is working on them in parallel and finding what is common between
them.8
Linux is Inexorably Winning

The fact that the Linux kernel has these many advantages over
Windows means two things. First, the argument that free software
engineers are unable to innovate and only copy the work of others is
not true.
The head of Windows Server recently said that free software, by
its very nature, does not allow intellectual property to be built. This
statement is factually incorrect: the Linux kernel is doing many
things no kernel has ever done before, just as Wikipedia is doing
things no encyclopedia has ever done before.
Second, even if Microsoft gave away the source code to the Windows kernel, the existing free software community would dismiss it.
A worldwide community has pounded Linux into the shape it should
be. Linus calls Windows witchcraft, so why work on an inferior
codebase whose details up till now have been opaque?
Two of the biggest differences in strategy is User Mode Linux (UML) which
changes Linux to run as an application on top of another instance of Linux, and
the standard virtualization, which runs the guest OS in kernel mode, though it
doesn't actually talk to the hardware. The Linux kernel is evolving towards figuring out the architecture, and what is shared between the different strategies.
Linux
39
Charging for an OS
A Linux operating system is an entirely different beast compared
to a Microsoft operating system. Microsoft was constantly torn
about how much value to invest in Windows, and how much to set
aside for extra licensing revenue in other products. Windows Vista
has five different versions, (originally they announced eight!), each
with basically the same code, but with dramatically different prices:
Product
(Amazon.com)
Upgrade
New
Windows Vista Ultimate
$243
$350
Windows Vista Business
$176
$260
Windows Vista Home Premium
$140
$219
$85
$157
Windows Vista Home Basic

Windows Vista Enterprise
Custom volume license
Microsoft charges $85 to $350 for Windows Vista, but the code in each version is 99% the same.
Microsoft's natural goal is to put as many new and compelling

features into high-end, high-profit versions even though the prices
are not correlated to the work required to build the various Vista
versions.
Creating multiple versions is tricky because if third-party applications depend on features that aren't in a particular version of Windows, then the applications won't run, and the Windows brand is
weakened. Therefore, Microsoft would sometimes add logic to cripple the performance of high-end features on low-end versions. In the
free software world, no one deliberately cripples their own creation.
When I worked at Microsoft, there were numerous turf wars. For
example, the Word team fought with the Windows team over
whether WordPad, a tiny Windows applet, should have the ability to
read Word's DOC files. The Windows team wanted to create an operating system that had the ability to display the hundreds of millions
of DOC files, but the Word team didn't want to create a reason for
people not to purchase Word. There were also running battles
between the teams responsible for Outlook and Outlook Express,
Exchange and SQL, Works and Word, FoxPro and Access, Access
and VB, SQL Server and Access, PC and XBox. Each team was paranoid about another team adding functionality that would discourage
someone from buying their product.
40
Linux
Microsoft also was torn between bundling important features in

the operating system, like they did with web browsers, instant messaging and multimedia, and leaving features out in order to reap
more money later. Windows doesn't ship with a dictionary because
Office has one. Windows doesn't come with development tools
because those are part of Microsoft's Visual Studio business. With
Linux, anything free is welcomed into the operating system.
In addition to Microsoft's strategic decisions to exclude certain
features, there are also cases of benign neglect. The Sound
Recorder in Windows XP lets you only record for 1 minute, a limitation that exists from the days of 16-bit Windows that no one bothered to fix. Microsoft's official solution is for the customer to
purchase Office OneNote.
Applets, command-line tools, and many other important but
unsexy parts of an operating system were always allocated very limited resources. Ask Steve Ballmer for resources for the unimportant
Sound Recorder, and you would receive a dirty, polystyrene fork
hurtling in your direction. In the free software model, anyone, on
their time frame, can improve a piece of code that is useful or interesting to them, whether strategic or not. Eric Raymond calls this
phenomenon a developer scratching their own itch. Wikipedia is
built almost entirely from this mechanism, and only free software
can capture every little advancement.
Complexity of License Agreements

It would be possible to fund the construction of all roads with
tolls. This would entail having toll booths at all street corners.
Such a system would provide a great incentive to improve
roads. It would also have the virtue of causing the users of any
given road to pay for that road. However, a toll booth is an artificial obstruction to smooth drivingartificial, because it is not
a consequence of how roads or cars work.
Richard Stallman
Even if you believe it is perfectly okay to charge for software, it is

hard to know what to charge consumers for software. Microsoft
used to charge a fixed amount for a server product. However, one
day they realized that customers who had 10 users connected to a
server should be paying less than those who had 100. This observation resulted in the creation of client access licenses (CALs); a concept that required payment for each individual user of server
software, but which is an additional burden on the customer. If you
have 1,000 employees accessing 1,000 servers, you need to fill out a
lot of paperwork.
Linux
41
This was all before the invention of the Internet whereby the
number of users of a server could easily be in the thousands and
which made many usages of CALs expensive and unsustainable.
Therefore, Microsoft moved towards a model where the cost was
based on the number of processors in the computer so that little
boxes would cost less than big boxes.
This model worked until Intel introduced the concept of hyperthreading, which fools a computer into thinking there are two processors inside the computer, but which adds only 15-30% more
performance. Microsoft's customers would of course be unhappy at
the thought of purchasing a bunch of new licenses for such a small
performance improvement, so Microsoft ended up giving free
licenses for hyperthreaded processors.
Then, virtualization was created:
Intel virtualization marketing diagram: Virtualization allows the ability to

run multiple operating systems on one computer: each instance thinks it
controls the machine. The best uses for virtualization are web hosting,
inside corporate data centers, and for software developers.
Virtualization allows different applications to be fully isolated

from each other inside different complete instances of an operating
system, but at the same time share the CPU and other hardware
resources. Isolation is important because even within a single com-
42
Linux
pany's data center, different departments don't want to run their

code on the same machine. If Hotmail went down, they didn't want
it to be the fault of the Microsoft Bob web page.
I was told that inside the cavernous datacenters of a Fortune 500
company, the computers used on average only 15% of their CPU's
capacity. The different departments all maintained their own hardware, which they'd built out to handle the maximum possible load;
this is an idea as silly as utilizing only 15% of an office building.
Virtualization gives you software isolation but allows you to share
hardware. However, when you install more proprietary software,
licensing issues arise. If I put three copies of a database server in
separate virtualization instances on a four-processor computer,
under many of Microsoft's licensing models I would have to purchase enough licenses for 12 processors even though the computer only has four. Like hyperthreading, virtualization is another
technology that wasn't conceived of when Microsoft created their
per-processor licensing model.
In a free software environment, you can add new hardware and
add or remove applications without paying or keeping track of anything. I will write more about economic aspects later, but for now it
is only important to understand that free software sidesteps these
hassles, which are often very onerous for enterprises.
Free Software Only Costs PCs

Creating a unified industrial software base built around free software will not only make PCs more powerful, but it will also allow us
to push their intelligence everywhere, from cars to medical equipment. Free software will be built because it is so valuable to businesses; everyone else will come along for the ride and get it for free.
The proprietary software model has hurt many of the smaller software markets. In fact, except for gaming business which is as much
about artistry as software, Microsoft is nearly the last proprietary
software company standing. When I was at Microsoft, our biggest
competitors were companies such as Borland, WordPerfect, Corel,
Lotus, Netscape and Sybase names you never hear anymore.
There is no Microsoft of educational software selling products
used in every school. There are several companies selling proprietary products, perhaps receiving enough revenue to stay in business, but the proprietary software model has prevented them from
achieving critical mass.
Linux
43
Recently, I saw an ad for a trivial software applet capable of converting a DVD to the iPod video format. Microsoft managed to convince everybody that each little software product is worth selling.
Now you see why Stallman's analogy of proprietary software as tollbooths is apt they serve as permanent, pervasive obstacles holding up progress for software that was written years earlier. Free
software flows with much less friction. In fact, in the free software
world, the definition of a PC operating system completely changes.
A Free Operating System

Our most potent operating system competitor is Linux and the
phenomena around open source and free software. The same
phenomena fuels competitors to all of our products. The ease of
picking up Linux to learn it or to modify some piece of it is very
attractive. The academic community, start-up companies, foreign governments and many other constituencies are putting
their best work into Linux.
Bill Gates
The One Laptop Per Child has as much CPU power as a workstation of 15
years ago, but would be just a shiny box without free software and content.
The biggest difference between Windows and Linux is that free

software contains thousands of applications, installable with one
click, and managed as one set. A Linux operating system includes all
the obvious stuff like a spreadsheet, web browser, and instant mes-
44
Linux
saging, but it also includes tools for making pictures and music, children's applications, server software, the Bible, development tools,
and much more.
Audacity is the most popular free audio editor on Linux. It doesn't
have a talking paper clip: It looks like you're trying to add echo.
Would you like some help? But it does provide a well-rounded set of
features, and has many effects for the manipulation of sound.
A rich and reliable free audio manipulation tool
Audacity, of course, doesn't have a one-minute limitation on its

ability to record as Windows XP does. Its most important feature is
its plugins for import, export, and effects extensibility is one of
free software's best features.
Linux
45
Audacity Audio Editor Feature List

Recording
Audacity can record live audio
through a microphone or mixer, or

digitize recordings from cassette
tapes, vinyl records, or minidiscs.
With some sound cards, it can also
capture streaming audio.
Record from microphone, line input,
or other sources.
Dub over existing tracks to create
multi-track recordings.
Record up to 16 channels at once
(requires multi-channel hardware).
Level meters can monitor volume
levels before, during, and after
recording.
Import and Export

Import sound files, edit them, and
combine them with other files or

new recordings. Export your recordings in several common file formats.
Import and export WAV, AIFF, AU,
and Ogg Vorbis files.
Import MPEG audio (including MP2
and MP3 files) with libmad.
Export MP3s with the optional
LAME encoder library.
Create WAV or AIFF files suitable for
burning to CD.
Open raw (headerless) audio files
using the Import Raw command.
Note: Audacity does not currently
support WMA, AAC, or most other
proprietary or restricted file formats.
Editing
Easy editing with Cut, Copy, Paste,
and Delete.
Use unlimited Undo (and Redo) to go
back any number of steps.
Very fast editing of large files.
Edit and mix an unlimited number of
tracks.
Use the Drawing tool to alter individual sample points.
Fade the volume up or down
smoothly with the Envelope tool.
Effects
Change the pitch without altering
the tempo, or vice-versa.

Remove static, hiss, hum, or other
constant background noises.
Alter frequencies with Equalization,
FFT Filter, and Bass Boost effects.
Adjust volumes with Compressor,
Amplify, and Normalize effects.
Other built-in effects include:
Echo
Phaser
Wahwah
Reverse
Sound Quality
Record and edit 16-bit, 24-bit, and
32-bit (floating point) samples.

Record at up to 96 KHz.
Sample rates and formats are converted using high-quality resampling
and dithering.
Mix tracks with different sample
rates or formats, and Audacity will
convert them automatically in real
time.
Plug-Ins
Add new effects with LADSPA plug
ins.
Audacity includes some sample plugins by Steve Harris.
Load VST plugins for Windows and
Mac, with the optional VST Enabler.
Write new effects with the built-in
Nyquist programming language.
Analysis
Spectrogram mode for visualizing
frequencies.
Plot Spectrum command for
detailed frequency analysis.
Free and Cross-Platform

Licensed under the GNU General
Public License (GPL).

Runs on Mac OS X, Windows, and
GNU/Linux.
Developers of free software applications tend to build extensibility plugins as a fundamental way of writing their software because
they know their tool will never by itself be able to do all the things
people will want. A plugin provides a boundary between things that
46
Linux
manage data, and things that manipulate it. The most popular plugins eventually become a part of the base system, but by being built
separately, they have forced clean boundaries and modularity.9
Every application that Linux has that Windows doesn't is a feature Windows is missing:
Richard Stallman's free software vision realized: A free Linux operating system contains an entire store of free applications available with one click,
and built to work together. Having so many tools at your disposal makes
computers more personal, powerful, productive, and enjoyable. Your computing experience becomes limited only by your creativity.
A free operating system is where Metcalfe's law meets software:

the more people using free software, the more applications will be
built for it. Just one piece of free software isn't useful, but with an
entire stack, we can enter a shining age.
9
I argue in another place in the book that software has no clear boundaries. What I
meant was that one never really knows precisely what the interface between manager and manipulator should be. For audio files, the boundary seems clear: here is
some audio data, chew on it. However even there you must ask: what DSP APIs
are available to the plugins? Otherwise, each plugin will need lots of duplicate
code that the manager already likely has! It is the new hardware capabilities that
create a need for a change at this boundary. The lesson here is to keep your
boundaries simple, but assume you may need to change them.
Linux
47
Going from today's 20 million Linux users to the anticipated one

billion means the potential for 50 times more resources. The free
software stack has several challenges I will discuss throughout this
book, but it is important to mention here that few of the applications
in the dialog box are as polished or reliable as Firefox. However,
many are very powerful, and more than good enough to depend on
in a business.
While Linux still needs work, Windows is no day at the beach.
Here is an email from Bill Gates' describing his experience of
installing Microsoft MovieMaker on Windows:
From:
Sent:
To:
Cc:
Bill Gates
Wednesday, January 15, 2003 10:05 AM
Jim Allchin
Chris Jones (WINDOWS); Bharat Shah (NT); Joe
Peterson; Will Poole; Brian Valentine; Anoop
Gupta (RESEARCH)
Subject:
Windows Usability degradation flame
I am quite disappointed at how Windows Usability has been
going backwards and the program management groups don't
drive usability issues.
Let me give you my experience from yesterday.
I decided to download (Moviemaker) and buy the Digital Plus
pack ... so I went to Microsoft.com. They have a download place
so I went there.
The first 5 times I used the site it timed out while trying to
bring up the download page. Then after an 8 second delay I got
it to come up.
This site is so slow it is unusable.
It wasn't in the top 5 so I expanded the other 45.
These 45 names are totally confusing. These names make stuff
like: C:\Documents and Settings\billg\My Documents\My Pictures seem clear.
They are not filtered by the system ... and so many of the things
are strange.
I tried scoping to Media stuff. Still no moviemaker. I typed in
movie. Nothing. I typed in movie maker. Nothing.
So I gave up and sent mail to Amir saying - where is this
Moviemaker download? Does it exist?
So they told me that using the download page to download
something was not something they anticipated.
They told me to go to the main page search button and type
movie maker (not moviemaker!)
I tried that. The site was pathetically slow but after 6 seconds
of waiting up it came.
48
Linux
I thought for sure now I would see a button to just go do the
download.
In fact it is more like a puzzle that you get to solve. It told me to
go to Windows Update and do a bunch of incantations.
This struck me as completely odd. Why should I have to go
somewhere else and do a scan to download moviemaker?
So I went to Windows update. Windows Update decides I need
to download a bunch of controls. (Not) just once but multiple
times where I get to see weird dialog boxes.
Doesn't Windows update know some key to talk to Windows?
Then I did the scan. This took quite some time and I was told it
was critical for me to download 17megs of stuff.
This is after I was told we were doing delta patches to things
but instead just to get 6 things that are labeled in the SCARIEST possible way I had to download 17meg.
So I did the download. That part was fast. Then it wanted to do
an install. This took 6 minutes and the machine was so slow I
couldn't use it for anything else during this time.
What the heck is going on during those 6 minutes? That is
crazy. This is after the download was finished.
Then it told me to reboot my machine. Why should I do that? I
reboot every night why should I reboot at that time?
So I did the reboot because it INSISTED on it. Of course that
meant completely getting rid of all my Outlook state.
So I got back up and running and went to Windows Update
again. I forgot why I was in Windows Update at all since all I
wanted was to get Moviemaker.
So I went back to Microsoft.com and looked at the instructions.
I have to click on a folder called WindowsXP. Why should I do
that? Windows Update knows I am on Windows XP.
What does it mean to have to click on that folder? So I get a
bunch of confusing stuff but sure enough one of them is
Moviemaker.
So I do the download. The download is fast but the Install takes
many minutes. Amazing how slow this thing is.
At some point I get told I need to go get Windows Media Series
9 to download.
So I decide I will go do that. This time I get dialogs saying
things like "Open" or "Save". No guidance in the instructions
which to do. I have no clue which to do.
The download is fast and the install takes 7 minutes for this
thing.
So now I think I am going to have Moviemaker. I go to my
add/remove programs place to make sure it is there.
It is not there.
Linux
49
What is there? The following garbage is there. Microsoft

Autoupdate Exclusive test package, Microsoft Autoupdate
Reboot test package, Microsoft Autoupdate testpackage1.
Microsoft AUtoupdate testpackage2, Microsoft Autoupdate Test
package3.
Someone decided to trash the one part of Windows that was
usable? The file system is no longer usable. The registry is not
usable. This program listing was one sane place but now it is all
crapped up.
But that is just the start of the crap. Later I have listed things
like Windows XP Hotfix see Q329048 for more information.
What is Q329048? Why are these series of patches listed here?
Some of the patches just things like Q810655 instead of saying
see Q329048 for more information.
What an absolute mess.
Moviemaker is just not there at all.
So I give up on Moviemaker and decide to download the Digital
Plus Package.
I get told I need to go enter a bunch of information about
myself.
I enter it all in and because it decides I have mistyped something I have to try again. Of course it has cleared out most of
what I typed.
I try (typing) the right stuff in 5 times and it just keeps clearing
things out for me to type them in again.
So after more than an hour of craziness and making my programs list garbage and being scared and seeing that Microsoft.com is a terrible website I haven't run Moviemaker and I
haven't got the plus package.
The lack of attention to usability represented by these experiences blows my mind. I thought we had reached a low with
Windows Network places or the messages I get when I try to
use 802.11. (don't you just love that root certificate message?)
When I really get to use the stuff I am sure I will have more
feedback.
Linux Distributions
With Linux, each OS distribution carves out a niche to meet its
users' needs. There are specialized versions of Linux containing educational software, tools for musicians, versions dedicated to embedded or low-end hardware, and regional versions of Linux produced
in places like Spain and China.
The various distributions have much in common, including the
Linux kernel, but use different free software and installation mechanisms. One distribution called Gentoo downloads only one binary, a
50
Linux
bootstrapping compiler. The rest of its deliverables are the source

code to the components they offer. This gives the user the ability to
build a system highly optimized for his hardware.
Some distributions are optimized to run well on old hardware and
fit on CDs the size of a credit card:
Damn Small Linux is the most popular Linux for old computers and ships on
80x60 mm CDs.
Linux is very popular on servers, which require an additional

focus on performance, reliability, and security. One of the simplest
ways to decrease bloat and security risks is to remove the graphical
interface:
top - 12:54:17 up 62 days, 20:14, 2 users, load
Tasks: 127 total, 1 running, 126 sleeping, 0
Cpu(s): 0.2% us, 0.3% sy, 0.0% ni, 98.5% id,
si
Mem:
514248k total, 489360k used,
24888k
Swap: 1020088k total,
18416k used, 1001672k
PID
2041
1
2
3
4
USER
root
root
root
root
root
average: 0.16, 0.42, 0.43

stopped, 0 zombie
0.2% wa, 0.7% hi, 0.2%
free,
free,
PR NI VIRT RES SHR S %CPU %MEM

10 -5
0
0
0 S
0 0.0
16 0 1564 528 460 S
0 0.1
RT 0
0
0
0 S
0 0.0
34 19
0
0
0 S
0 0.0
RT 0
0
0
0 S
0 0.0
79128k buffers
177528k cached
TIME+
0:11.74
0:01.09
0:00.01
0:14.63
0:00.00
COMMAND
usb-storage
init
migration/0
ksoftirqd/0
watchdog/0
Screenshot of a text-based Linux process viewer. Just because a computer

displays only text doesn't mean there isn't a modern, rock-solid operating
system underneath.
It is as hard to write a graphical user interface (GUI) for a server

as it is to write the server itself: every new feature in the server
needs a corresponding improvement in the administration interface.
If your server can generate 1,000 records per second, then your
user interface had better be able to handle that without choking.10
While most devices do not need a GUI, desktop users do, and like
everything in the free software world today, there are several good
choices.
10 Many embedded scenarios don't include a GUI because there is no standard one
yet. Most GUIs are too complicated and slow for embedded scenarios: the brandnew cash register at my local Starbucks has a touch display with a text interface.
Linux
51
The following map shows the Debian branch of the Linux distribution family tree; they derive from each other, just like in a biological
ecosystem:
A portion of the Linux family tree showing the Debian branch, the biggest
free software distribution.
Here is a chart showing the relative popularity of Linux distros:
50%
Linux distributions, sorted by popularity. The line shows the divide between
both halves of the popularity curve.
52
Linux
What you see here is an almost perfectly smooth curve that illustrates a relatively new idea called the Long Tail. One way to think
about this idea is to look at the English language. Words like the
are used with great frequency, but many more words like teabag
are used infrequently. There is a long tail of infrequently used English words, and to just ignore them would be to throw away much of
what makes our language so special.
The lesson of the long tail in business is the importance of catering to customers with special interests. The long tail of Linux distributions means that the creation of a free software ecosystem
doesn't mean the end of the free market, or of competition.
Wikipedia and the Linux kernel are two of the best examples of
the fact that free software and the free exchange of ideas can create
superior products without licensing fees. The mere existence of
these premier products, without a gigantic company behind them, is
proof that the proprietary development model is doomed.
AI and Google
AI
AND
53
GOOGLE
The future is open source everything.

Linus Torvalds
That knowledge has become the resource, rather than a

resource, is what makes our society post-capitalist.
Peter Drucker, 1993
magine 1,000 people, broken up into groups of five, working on

two hundred separate encyclopedias, versus that same number
of people working on one encyclopedia? Which one will be the
best? This sounds like a silly analogy when described in the context
of an encyclopedia, but it is exactly what is going on in artificial
intelligence (AI) research today.1
In early drafts of this book, I had positioned this chapter after the
one explaining economic and legal issues around free software.
However, I now believe it is important to discuss artificial intelligence separately and first, because AI is the holy-grail of computing,
and the reason we haven't solved AI is that there are no free software codebases that have gained critical mass. Far more than
enough people are out there, but they are usually working in teams
of less than five, and very often in proprietary codebases.
Deep Blue has been Deep-Sixed

Some people worry that artificial intelligence will make us feel
inferior, but then, anybody in his right mind should have an
inferiority complex every time he looks at a flower.
Alan Kay, computer scientist
The source code for IBM's Deep Blue, the first chess machine to
beat then-reigning World Champion Gary Kasparov, was built by a
team of about five people. That code has been languishing in a vault
at IBM ever since because it was not created under a license that
would enable further use by anyone, even though IBM is not
attempting to make money from the code or using it for anything.
The second best chess engine in the world, Deep Junior, is also
not free, and is therefore being worked on by a very small team. If
One website documents 60 pieces of source code that perform Fourier transformations, which is an important software building block. The situation is the same
for neural networks, computer vision, and many other advanced technologies.
54
AI and Google
we have only small teams of people attacking AI, or writing code

and then locking it away, we are not going to make progress any
time soon towards truly smart software.
Today's chess computers have no true AI in them; they simply
play moves, and then use human-created analysis to measure the
result. If you were to go tweak the computer's value for how much a
queen is worth compared to a pawn, the machine would start losing
and wouldn't even understand why. It comes off as intelligent only
because it has very smart chess experts programming the computer
precisely how to analyze moves, and to rate the relative importance
of pieces and their locations, etc.
Deep Blue could analyze two hundred million positions per second, compared to grandmasters who can analyze only 3 positions
per second. Who is to say where that code might be today if chess AI
aficionados around the world had been hacking on it for the last 10
years?
DARPA Grand Challenge

Proprietary software developers have the advantages money
provides; free software developers need to make advantages for
each other. I hope some day we will have a large collection of
free libraries that have no parallel available to proprietary software, providing useful modules to serve as building blocks in
new free software, and adding up to a major advantage for further free software development.
What does society need? It needs information that is truly available to its citizensfor example, programs that people can
read, fix, adapt, and improve, not just operate. But what software owners typically deliver is a black box that we cant study
or change.
Richard Stallman
The hardest computing challenges we face are man-made: language, roads and spam. Take, for instance, robot-driven cars. We
could do this without a vision system, and modify every road on the
planet by adding driving rails or other guides for robot-driven cars,
but it is much cheaper and safer to build software for cars to travel
on roads as they exist today a chaotic mess.
At the annual American Association for the Advancement of Science (AAAS) conference in February 2007, the consensus among
the scientists was that we will have driverless cars by 2030. This
prediction is meaningless because those working on the problem are
not working together, just as those working on the best chess soft-
AI and Google
55
ware are not working together. Furthermore, as American cancer

researcher Sidney Farber has said, Any man who predicts a date for
discovery is no longer a scientist.
Today, Lexus has a car that can parallel park itself, but its vision
system needs only a very vague idea of the obstacles around it to
accomplish this task. The challenge of building a robot-driven car
rests in creating a vision system that makes sense of painted lines,
freeway signs, and the other obstacles on the road, including dirtbags not following the rules.
The Defense Advanced Research Projects Agency (DARPA), which
unlike Al Gore, really invented the Internet, has sponsored several
contests to build robot-driven vehicles:
Stanley, Stanford University's winning entry for the 2005 challenge. It

might not run over a Stop sign, but it wouldn't know to stop.
Like the parallel parking scenario, the DARPA Grand Challenge of

2004 required only a simple vision system. Competing cars traveled
over a mostly empty dirt road and were given a detailed series of
map points. Even so, many of the cars didn't finish, or perform confidently. There is an expression in engineering called garbage in,
garbage out; as such, if a car sees poorly, it is helpless.
What was disappointing about the first challenge was that an
enormous amount of software was written to operate these vehicles
yet none of it has been released (especially the vision system) for
others to review, comment on, improve, etc. I visited Stanford's
56
AI and Google
Stanley website and could find no link to the source code, or even
information such as the programming language it was written in.
Some might wonder why people should work together in a contest, but if all the cars used rubber tires, Intel processors and the
Linux kernel, would you say they were not competing? It is a race,
with the fastest hardware and driving style winning in the end. By
working together on some of the software, engineers can focus more
on the hardware, which is the fun stuff.
The following is a description of the computer vision pipeline
required to successfully operate a driverless car. Whereas Stanley's
entire software team involved only 12 part-time people, the vision
software alone is a problem so complicated it will take an effort
comparable in complexity to the Linux kernel to build it:
Image acquisition: Converting sensor inputs from 2 or more
cameras, radar, heat, etc. into a 3-dimensional image sequence
Pre-processing: Noise reduction, contrast enhancement
Feature extraction: lines, edges, shape, motion
Detection/Segmentation: Find portions of the images that
need further analysis (highway signs)
High-level processing: Data verification, text recognition,
object analysis and categorization
The 5 stages of an image recognition pipeline.
A lot of software needs to be written in support of such a system:
The vision pipeline is the hardest part of creating a robot-driven car, but
even such diagnostic software is non-trivial.
AI and Google
57
In 2007, there was a new DARPA Urban challenge. This is a sample of the information given to the contestants:
It is easier and safer to program a car to recognize a Stop sign than it is to

point out the location of all of them.
Constructing a vision pipeline that can drive in an urban environment presents a much harder software problem. However, if you
look at the vision requirements needed to solve the Urban Challenge, it is clear that recognizing shapes and motion is all that is
required, and those are the same requirements as had existed in the
2004 challenge! But even in the 2007 contest, there was no more
sharing than in the previous contest.
Once we develop the vision system, everything else is technically
easy. Video games contain computer-controlled drivers that can race
you while shooting and swearing at you. Their trick is that they
already have detailed information about all of the objects in their
simulated world.
After we've built a vision system, there are still many fun challenges to tackle: preparing for Congressional hearings to argue that
these cars should have a speed limit controlled by the computer, or
telling your car not to drive aggressively and spill your champagne,
or testing and building confidence in such a system.2
2
There are various privacy issues inherent in robot-driven cars. When computers
know their location, it becomes easy to build a black box that would record all
58
AI and Google
Eventually, our roads will get smart. Once we have traffic information, we can have computers efficiently route vehicles around any
congestion. A study found that traffic jams cost the average large
city $1 billion dollars a year.
No organization today, including Microsoft and Google, contains
hundreds of computer vision experts. Do you think GM would be
gutsy enough to fund a team of 100 vision experts even if they
thought they could corner this market?
There are enough people worldwide working on the vision problem right now. If we could pool their efforts into one codebase, written in a modern programming language, we could have robot-driven
cars in five years. It is not a matter of invention, it is a matter of
engineering. Perhaps the world simply needs a Linus Torvalds of
computer vision to step up and lead these efforts.
this information and even transmit it to the government. We need to make sure
that machines owned by a human stay under his control, and do not become controlled by the government without a court order and a compelling burden of proof.
AI and Google
59
Software and the Singularity

Futurists talk about the Singularity, the time when computational capacity will surpass the capacity of human intelligence. Ray
Kurzweil predicts it will happen in 2045.3 The flaw with any date
estimate, other than the fact that they are always prone to extreme
error, is that our software today has no learning capacity, because
the idea of continuous learning is not yet a part of the foundation.
Even the learning capabilities of an ant would be useful.
I believe the benefits inherent in the singularity will happen as
soon as our software becomes smart. I don't believe we need to
wait for any further Moore's law progress for that to happen. Computers today can do billions of operations per second, like add
123,456,789 and 987,654,321. Even if you could do that calculation
in your head in one second, it would take you 30 years to do the billion that your computer can do in that second.
Even if you don't think computers have the necessary hardware
horsepower to be smart today, understand that in many scenarios,
the size of the input is the driving factor to the processing power
required. In image recognition, for example, the amount of work
required to interpret an image is mostly a function of the size of the
image. Each step in the image recognition pipeline, and the processes that take place in our brain, dramatically reduce the amount
of data from the previous step. At the beginning of the analysis
3
His prediction is that the number of computers, times their computational capacity, will surpass the number of humans, times their computational capacity, in
2045. Therefore, the world will be amazing then.
This calculation is flawed for several reasons:
1. We will be swimming in computational capacity long before 2040. Today, my
computer is typically running at 2% CPU when I am using it, and therefore
has 50 times more computational capacity than I need. An intelligent agent
twice as fast as the previous one is not necessarily more useful.
2. Many of the neurons of the brain are not spent on reason, and so shouldn't
be in the calculations.
3. Billions of humans are merely subsisting, and are not plugged into the global
grid, and so shouldn't be measured.
4. There is no amount of continuous learning built in to today's software.
Each of these would tend to push Singularity forward and support the argument
that the benefits of singularity are not waiting on hardware. Humans make computers smarter, and computers make humans smarter, and this feedback loop
makes 2045 a meaningless moment.
Who in the past fretted: When will man build a device that is better at carrying
things than me? Computers will do anything we want, at any hour, on our command. A computer plays chess or music because we want it to. Robotic firemen
will run into a burning building to save our pets. Computers have no purpose
without us. We should worry about robots killing humans as much as we worry
about someone stealing an Apache helicopter and killing humans today.
60
AI and Google
might be a one million pixel image, requiring 3 million bytes of

memory. At the end of the analysis is the data that you are looking at
your house, a concept that requires only 10 bytes to represent. The
first step, working on the raw image, requires the most processing
power, so therefore it is the image resolution (and frame rate) that
set the requirements, values that are trivial to change. No one has
shown robust vision recognition software running at any speed, on
any sized image!
While a brain is different from a computer in that it does work in
parallel, such parallelization only makes it happen faster, it does not
change the result. Anything accomplished in our parallel brain could
also be accomplished on computers of today, which can do only one
thing at a time, but at the rate of billions per second. A 1-gigahertz
processor can do 1,000 different operations on a million pieces of
data in one second. With such speed, you don't even need multiple
processors! Even so, more parallelism is coming.4 Once we build
software as smart as an ant, we will build software as smart as a
human the same day, because it is the same software.
Most computers today contain a dual-core CPU and processor folks promise that
10 and more are coming. Intel's processors also have limited 4-way parallel processing capabilities known as MMX and SSE. Intel could add even more of this
parallel processing support if applications put them to better use. Furthermore,
graphics cards exist to do work in parallel, and this hardware could also be
adapted to AI if it is not usable already.
AI and Google
61
Google
One of the problems faced by the monopoly, as its leadership
now well understands, is that any community that it can buy is
weaker than the community that we have built.
Eben Moglen
In 1950, Alan Turing proposed a thought experiment as a definition of AI in which a computer's responses (presumed to be textual)
were so life-like that, after questioning, you could not tell whether
they were made by a human or a computer. Right now the search
experience is rather primitive, but eventually, your search engine's
response will be able to pass the Turing Test. Instead of simply
doing glorified keyword matching, you could ask it to do things like:
Plot the population and GDP of the United States from 1900
2000.5 Today, if you see such a chart, you know a human did a lot of
work to make it.
The creation of machines that can pass the Turing Test will make
the challenge of outsourcing seem like small potatoes. Why outsource work to humans in other countries when computers nearby
can do the task?
AI is a meaningless term in a sense because building a piece of
software that will never lose at Tic-Tac-Toe is a version of AI, but it
is a very primitive type of AI, entirely specified by a human and executed by a computer that is just following simple rules.
Fortunately, the same primitive logic that can play Tic-Tac-Toe can
be used to build arbitrarily smart software, like chess computers
and robot-driven cars. We simply need to build systems with enough
intelligence to fake it. This is known as Weak AI, as opposed to
Strong AI, which is what we think about when we imagine robots
that can pass the Turing Test, compose music, get depressed.
In Strong AI, you wouldn't give this machine a software program
to play chess, just the rules. The first application of Strong AI is
Search; the pennies for web clicks will pay for the creation of intelligent computers.
The most important and interesting service on the Internet is
search. Without an index, a database is useless imagine a phone
directory where the names were in random order. There is an enormous turf war taking place between Google, Yahoo!, and Microsoft
for the search business. Google has 200,000 servers, which at 200
5
Of course, there are some interesting complexities to the GDP aspect, like
whether to plot the GDP in constant dollars and per person.
62
AI and Google
hits per second gives them the potential for three trillion transactions per day. Even at fractions of pennies per ad, the potential revenue is huge. Right now, Google has 65% of the search business,
with Yahoo! at 20% and Microsoft at 7%. Bill Gates has said that
Microsoft is working merely to keep Google honest, which reveals
his acceptance that, unlike Windows and Office, MSN is not the
leader. (Note that Microsoft's search and other online efforts have
an inherent advantage because they get as much software as they
want for free. Any other company which wanted to build services
using Microsoft's software would have much higher costs.)
Furthermore, to supplant an incumbent, being 10% better is
insufficient. It will take a major breakthrough by one of Google's
competitors to change the game Microsoft's Bing is not one of
those. I use Google because I find its results good enough and
because it keeps a search history, so that I can go back in time and
retrieve past searches. If I started using a different search provider,
I would lose this archive.
Google depends heavily on free software, but very little of their
code is released to outsiders. One can use many of Google's services
for free, as Google makes most of its money on advertising, but you
cannot download any of their code to learn from it or improve it or
re-use it in ways not envisioned by them. Probably 99% of the code
on a typical server at Google is free software, but 99% of the code
Google itself creates is not free software.6 Google's source code is
not only not freely available, it is not for sale.
In fact, Google is an extremely secretive and opaque company.
Even in casual conversation at conferences, its engineers quickly
retreat to statements about how everything is confidential. Curiously, a paper explaining PageRank, written in 1998 by Google cofounders Sergey Brin and Larry Page, says, With Google, we have a
strong goal to push more development and understanding into the
academic realm. It seems they have since had a change of heart.
6
Although Google doesn't give away or sell their source code, they do sell an appliance for those who want a search engine for the documents on an internal
Intranet. This appliance is a black box and is, by definition, managed separately
than the other hardware and software in a datacenter.
It also doesn't allow tight integration with internal applications. An example of a
feature important to Intranets is to have the search engine index all documents I
have access to. The Internet doesn't really have this problem as basically everything is public. Applications are the only things that know who has access to all
the data. It isn't clear that Google has attacked this problem and because the
appliance is not extensible, no one other than Google can fix this either. This feature is one reason why search engines should be exposed as part of an application.
AI and Google
63
Google has sufficient momentum and sophistication to leave its

competitors in the dust. Here is a list of Google's services:
Google is applying Metcalfe's law to the web: Gmail is a good product, but
being a part of the Google brand is half of its reason for success.
64
AI and Google
Even with all that Google is doing, search is its most important
business. Google has tweaked its patented PageRank algorithm
extensively and privately since it was first introduced in 1998, but
the core logic remains intact: The most popular web pages that
match your search are the ones whose results are pushed to the
top.7
Today, PageRank can only rank what it understands. If the database contains words, it ranks the words. PageRank lets the wisdom
in millions of web sites decide what is the most popular, and therefore the best search result because the computer cannot make
that decision today. PageRank is an excellent stopgap measure to
the problem of returning relevant information, but the focus should
be on putting richer information into the database.
I believe software intelligence will get put into web spiders, those
programs that crawl the Internet and process the pages. Right now,
they mostly just save text, but eventually they will start to understand it, and build a database of knowledge, rather than a database
of words. The rest is mostly a parsing issue. (Some early search
engines, treated digits as words: searching for 1972 would find any
reference to 1, 9, 7 or 2; this is clearly not a smart search algorithm.)
The spiders that understand the information, because they've put
it there, also become the librarians who take the search string you
give it, and compare that to its knowledge.8 You need a librarian to
build a library, and a librarian needs the library she built to help
you.
Today, web spiders are not getting a lot of attention in the search
industry. Wikipedia documents 37 web crawlers, and it appears that
the major focus for them is on performance and discovering spam
websites containing only links which are used to distort rank.9
7
8
9
Some of Google's enhancements also include: freshness, which gives priority to

recently-changed web pages. It also tries to classify queries into categories like
places and products. Another new tweak is to not display too many results of one
kind: they try to mix in news articles, advertisements, a Wikipedia entry, etc.
These enhancements are nice, but are far from actually understanding what is in
the articles, and it appears to be applying smarts to the search query rather than
the data gathered by the spiders.
One might worry about how a spider that has only read a small portion of the
Internet can help you with parts it has not seen? The truth is that these spiders all
share a common memory.
Focusing on the spider side means continually adding new types of information
into the database as it starts to understand things better. Let's say you build a spider that can now understand dates and guess the publication date of the page.
(This can be tricky when a web page contains biographical information and therefore many dates.) Spiders will then start to tag all the web pages it reads in the
AI and Google
65
The case for why a free search engine is better is a difficult one to
make, so I will start with a simpler example, Google's blogging software.
Blogger
While Google has 65% of the billion-dollar search business, it has
10% or less of the blog business. There exists an enormous number
of blog sites, the code for which is basically all the same. The technology involved in running Instapundit.com, one of the most influential current-events blogs, is little different than that running
Myspace, the most popular diary and chatboard for Jay-Z-listening
teenage girls.
Google purchased the proprietary blogging engine Blogger in
2000 for an undisclosed amount. Google doesn't release how many
users they have because they consider that knowledge proprietary,
but we do know that no community of hundreds of third party developers is working to extend Blogger to make it better and more useful.
The most popular free blogging engine is WordPress, a core of
only 40,000 lines of code. It has no formal organization behind it,
yet we find that just like Wikipedia and the Linux kernel, WordPress
is reliable, rich, and polished:
WordPress, the most popular free blogging engine

future with this new information. All of this tagging happens when the data is
fetched, so that is where the intelligence needs to go.
66
AI and Google
WordPress is supported by a community of developers, who have

created plug-ins, written and translated documentation, and
designed many themes to customize the look. Here are the categories of plug-ins available for WordPress:
Administration
Administration Tools
Advertisement
Anti-Spam
Comments
Meta (tagging)
Restrictions
Statistics
Syntax Highlighting
Syndication
Translation and Languages
Tweaking
Monetizing
Design, Layout and Styles
Archive
Calendar - Event
Navigation
Randomness
Styles
Widgets
Links
3rd-parties services
Graphics, Video, and Sound

Audio
Images
Multimedia
Video
Odds and Ends
Financial
Forums
Geo
Miscellaneous
Mood
Time
Weather
Outside Information
Del.icio.us
Technorati
Posts
Audio Posts
Editing Posts
Formatting Posts
Miscellaneous Post Plugins
There are hundreds of add-ons for WordPress that demonstrate the health
of the developer community and which make it suitable for building even
very complicated websites. This might look like a boring set of components,
but if you broke apart MySpace or CNN's website, you would find much of
the same functionality.
Google acquired only six people when it purchased Pyra Labs, the
original creators of Blogger, a number dwarfed by WordPress's hundreds of contributors. As with any thriving ecosystem, the success of
WordPress traces back to many different people tweaking, extending and improving shared code. Like everything else in the free software community, it is being built seemingly by accident.10
10 In fact, WordPress's biggest problem is that the 3rd party development is too rich,
and in fact chaotic. There are hundreds of themes and plugins, many that duplicate each other's functionality. But grocery stores offer countless types of toothpaste, and this has not been an insurmountable problem for consumers. I talk
more about this topic in a later chapter.
AI and Google
67
In addition to blogging software, I see other examples where

Google could have worked more closely with the free software community. Recently I received many e-mails whose first words were:
Your cr. rating doesn't matter that I dutifully marked as spam. It
took weeks before Gmail's spam filter caught on. Spam is a very
hard problem and cooperating with others could help improve
Google faster, and lower their R&D costs.
Search
Google tells us what words mean, what things look like, where
to buy things, and who or what is most important to us.
Google's control over results constitutes an awesome ability
to set the course of human knowledge.
Greg Lastowka, Professor of Law, Rutgers University
And I, for one, welcome our new Insect Overlords.
News Anchorman Kent Brockman, The Simpsons
Why Google should have built Blogger as free software is an easier case to make because it isn't strategic to Google's business or
profits, the search engine is a different question. Should Google
have freed their search engine? I think a related, and more important question is this: Will it take the resources of the global software
community to solve Strong AI and build intelligent search engines
that pass the Turing Test?
Because search is an entire software platform, the best way to
look at it is by examining its individual components. One of the most
fundamental responsibilities for the Google web farm is to provide a
distributed file system. The file system which manages the data
blocks on one hard drive doesn't know how to scale across machines
to something the size of Google's data. In fact, in the early days of
Google, this was likely one of its biggest engineering efforts. There
are (today) a number of free distributed file systems, but Google is
not working with the free software community on this problem. One
cannot imagine that a proprietary file system would provide Google
any meaningful competitive advantage, nevertheless they have built
one.
Another nontrivial task for a search engine is the parsing of PDFs,
DOCs, and various other types of files in order to pull out the text to
index them. It appears that this is also proprietary code that Google
has written.
It is a lot easier to create a Google-scaled datacenter with all of
its functionality using free software today than it was when Google
was formed in 1998. Not only is Google not working with the free
68
AI and Google
software community on the software they have created, they are

actually the burdened first-movers. What you likely find running on
a Google server is a base of Linux and other free software, upon
which Google has created their custom, proprietary code. Google
might think their proprietary software gives them an advantage, but
it is mostly sucking up resources, and preventing them from leveraging advancements from outside developers.
And like Microsoft's Windows NT kernel, even if Google were to
release their infrastructure code, much of it would not be picked up
because the free software community has developed their own solutions. In fact, in late 2006, Google began to release tiny bits and
pieces of their most boring software, but when I looked at the codebases, there didn't appear to be much contributions from the outside because it isn't nearly as interesting to the world as it would
have been ten years earlier. Furthermore, as these codebases have
lived inside Google for a long time they probably have lots of dependencies on other Google technologies which make it hard for it isolated and used in the outside world.
What about the core of Google's business, the code that takes
your search request and attempts to make sense of it so that it can
pass the Turing Test? Google has not even begun to solve this problem, and even many simpler problems, so it makes one wonder if it
is something a single company can solve by itself.
There are two kinds of engineering challenges for Google:
Those necessary, non-strategic, and at best loosely correlated to
their profits, like blogging, language translation, and spam detection, none of which Google is cooperating with the community on.
Then there is the daunting problem of building software with
Strong AI which Google had better be working on with the rest of
the world. The idea of Google owning Strong AI is at least as scary
as Microsoft owning Windows and Office. Google has publicly stated
that Microsoft's proprietary software model has been bad for the
industry, but doesn't recognize that it is trying to do the exact same
thing!
Google is one of the few new, large, and fast-growing software
businesses in America and few people are publicly arguing that the
company give away the farm by sharing their core technology with
the free software community. This is especially scary because it is an
irreversible step. However, software is not a datacenter or a relationship with customers and advertisers. Most of the users of
Google's code would not be in competition with Google, but would
AI and Google
69
be taking it to new places that they hadn't considered. Furthermore,

Google would still have a significant first-mover advantage of the
code they created.
In addition, if someone else eventually creates a free search
engine that a worldwide community of researchers coalesce around,
where will Google be then? Perhaps Microsoft could flank Google by
building a free search engine that scientists and researchers around
the world could tinker in. There is an interesting free codebase
called Lucene, run by the Apache foundation, which is steadily gaining use in Enterprises who want to run their own search engine. It
seems quite possible that this is the codebase and community that
will provide a threat to Google in five to ten years.
Conclusion
Comic from xkcd.com
There is reason for optimism about the scientific challenges we

confront because the global community's ability to solve problems is
greater than the universe's ability to create them. The truth waiting
for us on the nature of matter, DNA, intelligence, etc. has been
around for billions of years.
There are millions of computer scientists sitting around wondering why we haven't yet solved the big problems in computer science.
It should be no surprise that software is moving forward so slowly
because there is such a lack of cooperation.
70
Free Software
FREE SOFTWARE
If you have an apple and I have an apple and we exchange
these apples then you and I will still each have one apple. But if
you have an idea and I have an idea and we exchange these
ideas, then each of us will have two ideas.
George Bernard Shaw
Inscription found on a wall in Edinburgh.
uch of man's existence until the late 20th century involved

an economy focused primarily on the manipulation of
scarce, unmalleable atoms. Copyright law was created to
protect writers, the true creators, and to regulate the few who had
the means of making copies the publishers. In a digital world, we
can all be creators and publishers, so we need to revisit many fundamental questions, from the means we use to protect ideas, to the
ways in which we purchase them.
This and the next chapter discusses details of free software, copyright and patent laws, but let's start by remembering that a shift
towards a presumption of free digital intellectual property is actually a moral question. The Internet makes transmitting knowledge
essentially free, so as free software evangelist Eben Moglen asks: If
Free Software
71
you could feed the world for free, would you? Likewise, if you could
provide every child access to a library of human knowledge they
would never outgrow, would you? It is the Internet that makes this
question possible to ask, and necessary to answer!
Without the right software to decode and manipulate it, a digital
idea is just a blob of bits to your computer. With the Internet, we can
exchange bits, but with free software, we can exchange ideas. While
free knowledge and free software are not any direct goal of the free
market, they provide tremendous benefits to a free market because
they allow anyone to create further value. If the larger goal to
encourage as many programmers as possible to write software, then
the free software approach has already demonstrated its superiority,
even though it is only on 1% of desktops. In fact, the proprietary
world was always destined to have just one company dominate; a
clone of Bill Gates who came along later would have been unable to
learn from and improve upon the innovations of the first.
Free software brings the libertarian benefit of allowing information to be used in unlimited new ways, combined with the communitarian benefit of ensuring that no one is left behind by the access
cost of knowledge. Because free software is better for the free
market than proprietary software, and an important element of a
society characterized by the free exchange of ideas, I think it is a
better name than open source, although both represent the same
basic idea. (Another reason to call it free software is that there is
an academic tradition that the person who discovers or defines
something has the right to give it a name, and Richard Stallman
defined free software long before others created open source.)
Software as a Science
In any intellectual field, one can reach greater heights by
standing on the shoulders of others. But that is no longer generally allowed in the proprietary software fieldyou can only
stand on the shoulders of the other people in your own company.
The associated psychosocial harm affects the spirit of scientific
cooperation, which used to be so strong that scientists would
cooperate even when their countries were at war. In this spirit,
Japanese oceanographers abandoning their lab on an island in
the Pacific carefully preserved their work for the invading U.S.
Marines, and left a note asking them to take good care of it.
Richard Stallman
Software is a science: you reason, you create a hypothesis in the

form of source code, and then you test the hypothesis by running it
72
Free Software
on a computer. Some think software is an art because well-written

software has a certain elegance to it. But one could say the same
about many of the best mathematical proofs or formulas in physics.
Elegance does not make science into an art, it is simply an embodiment of Einstein's As simple as possible, but not simpler.
Linus Torvalds summarized the similarity between free software
and science as thus:
Science may take a few hundred years to figure out how the
world works, but it does actually get there, exactly because
people can build on each others' knowledge, and it evolves over
time. In contrast, witchcraft/alchemy may be about smart people, but the knowledge body never accumulates anywhere. It
might be passed down to an apprentice, but the hiding of information basically means that it can never really become any better than what a single person/company can understand.
And thats exactly the same issue with open source (free) vs
proprietary products. The proprietary people can design something that is smart, but it eventually becomes too complicated
for a single entity (even a large company) to really understand
and drive, and the company politics and the goals of that company will always limit it.
Even the word university, man's place for shared study, derives
from the Latin universitas magistrorum et scholarium, meaning a
community of teachers and scholars. Universities were long understood to be places where people were placed together to learn from
each other.
Unfortunately, today, proprietary software has spread from the
corporate world to universities and other public institutions. If corporations want to hoard their scientific advancements, that is fine,
albeit short-sighted, but our public institutions should not be following suit! Not only Stanford's robot-driven car, Stanley, but also a ton
of other proprietary software is written by public institutions today.
Just unleashing our public institutions towards free software will
greatly increase the pace of progress, without even accounting for
the software funded by corporations.
Some think of free software as a Marxist idea, but science has
always been public because it was understood that giving the knowledge away would spurn further innovation, and because the scientist needed some shoulders to stand on in the first place.
Corporations were created not to hoard knowledge but to take the
advancements in science and apply them to practical uses. There
still is plenty of opportunity for competition and free markets, even
if all of the advancements in science are freely available.
Free Software
73
Scientists' work is reviewed by peers, published in journals, and

discussed at conferences; the exchange of new snippets of knowledge between participants is fundamental to further progress. This
even allows for competition to be the first to discover something, or
put the advancement to commercial use.
With proprietary software, we've created something outside the
classically accepted model of scientific research, and even worse, it
has become a very dominant model in private and public institutions. Letting one company own an innovation in science might provide it an incentive to sell better products, but it limits the many
other people who could use that advancement in other ways.
Science is not all knowledge. Science is not a business, a service,
or a product. And to the extent that it took science to make a product, patents were created to protect ideas that were easily copied
but not easily invented or discovered.
One could even argue that patents are not necessary to stimulate
progress because the challenge someone has when going up against
GE's light bulb idea is the infrastructure to produce and distribute
such a product, and the knowledge gained in doing all of this. An old
line in business is that the one who wins is the firstest with the
mostest. Large companies have economies of scale and satisfied
customers that any newcomers would have to overcome, even if they
built a better product.
Economies of scale are great for the free market because they are
a powerful driver of increased efficiency and quality, but they also
mean that to defeat someone you need to be more than 10% better.
The need for breakthroughs to defeat incumbents is good for the
free market because it forces newcomers to think big. Fortunately,
transformative technologies come along frequently enough that no
incumbent is ever truly safe.
Because software is a science, making it freely available may hurt
proprietary software companies, but it will help every other type of
company.
74
Free Software
Definition of Free Software

Making Linux use the GPL license was definitely the best thing
I ever did.
Linus Torvalds
Because software is a science, we need to create license agreements which allow, and even encourage, cooperation among programmers. Computers scientists need software to be freely available
for them to do their work.
Richard Stallman has defined the four basic software freedoms:
1.
2.
3.
4.
The freedom to run the program, for any purpose. (You, not
your software, are in control of what is happening.)
The freedom to study how the program works and adapt it
to your needs.
The freedom to give a copy of the program to your neighbor. Sharing ideas is impossible without sharing the program to create and display them.
The freedom to improve the program, and release your
improvements to the public, so that the whole community
benefits.
The GNU General Public License (GPL) is the copyright mechanism he came up with to protect those freedoms, which will allow
maximal re-use of advancements. The goal is to protect the freedom
of the user of free software. Without the ability to study and manipulate the software, you are greatly limited in your ability to use it and
further advance it.
Copyright was created to protect the creators, from the publishers (those with the means to make copies), by granting the creator
exclusive rights. The GNU GPL is sometimes called copyleft,
because it grants the same expansive rights to everyone, creator,
and user.
It sounds backwards to protect users rather than creators, but
protecting users also helps creators. All Linux programmers except
Linus started off as users. Linus was the first user of Linux as well
as being the first contributor to Linux. He fully owned his code and
could fix any problem he found. Copyleft ensures that code protected with this license provides those same guarantees to future
users and creators.
Copyleft helps Linus because it encourages, even requires users
and programmers of Linux to give back to his invention. Linux only
Free Software
75
ran on a 80386 CPU when first released because that is what Linus
owned. All the improvements that it took to run on the other processors got put into Linux, to the benefit of Linus, and everyone else.
Whatever you think of free software today, using it is a choice. In
fact, creating it is a charitable act, and we should be grateful to
Linus Torvalds for releasing his work, just like we should be grateful
to Einstein for making his E=MC2 theory publicly available.
Microsoft, Apple, Google, and many of the other blue-chip computer companies do not yet accept the idea that software should be
free in the ways that Stallman defines above. According to them,
you can generally run the code for whatever purpose, but not copy,
study, or enhance it.
Copyleft and Capitalism

With the GPL freedoms comes one important responsibility:
ensuring future enhancements to free code have to stay as free as
the original code. This encourages a scientific community to stay
that way, but is a responsibility that only applies to the users who
choose to join the set of creators.
Free software is free for everyone but for the developers who will
take it into new places. Software protected by copyleft is called free
software, but it is not truly free because of the assumption that
future advancements will also be made free. There is no free lunch
in this world, and copyleft ensures that people are giving back to
this free software. This copyleft obligation is necessary but not
expensive.
Necessary
The reason it is necessary to have copyleft is that only 100% free
software is something someone else can further improve. Improving
a piece of free software, and making the enhancements proprietary,
effectively makes that entire piece of software proprietary. You need
access to an entire codebase to be able make changes to anywhere
in it. Denying free availability of code enhancements creates a new
boundary between science and alchemy.
Not Expensive
Free software is not expensive because, in practical terms,
advancements in software are nearly always based on existing
knowledge and are very small in scope. What you must give back is
much smaller than what you have freely received.
76
Free Software
When Linus first released Linux, it was a mere 10,000 lines of

code. He has since received 8.2 million lines in return, worth 4,000
man-years, or $400 million. Free software, like science, is a pyramid
scheme that works. The idea that someone gives you $400 million
worth of technology for free means that it is likely you will not need
to do much to get it to work for you, and your copyleft obligation
will be a very small fraction of what you have freely received.
A change in software is usually 10-100 hundred lines of code,
which is .001% to .01% of a million-line codebase. Technology
advances in a steady, stepwise fashion. Therefore, it is important
that once a piece of software is free, every little advancement to it is
also made free. Wikipedia and Linux both require the copyleft obligation, and I believe this is an important part of their success. They
also demonstrate that the copyleft obligation is enough to sustain
development of free software, without involving license fees.
Free software positions other people as the beneficiaries of your
solutions, so they can then focus on other problems, to your benefit.
Free software allows for an optimal division of labor, performed by
people who don't necessarily understand the whole picture, but are
just trying to make the software work better for them, or a customer
of theirs.
A free license doesn't take away your ability to benefit from your
work. And as long as it doesn't take away your ability to benefit,
there will always be motivation to write free software. Just as the
public availability of scientific advancements has not decreased the
motivation to do science, the free flow of software innovations does
not threaten to undermine the software market. If people only make
changes to software when they want to, and those changes are captured for others to use, no one is worse off. It helps others, and it
likely helps the creator because writing software is hard work.
Is Copyleft a Requirement for

Free Software?
An equitable and effective intellectual property system must
take into account both first creators and those who come later
to build upon their work.
Committee for Economic Development
Richard Stallman's copyleft, the idea of ensuring that free ideas

stay free, is one of the most important concepts in the digital age. It
is such a radical idea that it is not universally accepted, even within
the free software community.
Free Software
77
In fact, source code is considered free by today's software community if it supports the first three freedoms (run, study, copy), but
not copyleft (make those enhancement freely available to all.)
Two very popular free licenses, the MIT and BSD licenses, are
considered free but simply say: Please include this copyright message at the top of the source code. You can use this code, copy it
and study it, but you can also make it proprietary again. This sort of
free software does not require anyone to contribute anything back if
they enhance it.
Stallman considers these lax licenses; while they sound reasonable and are, strictly speaking, more free than copyleft, the problem is that this once-free software frequently becomes proprietary
again. Keith Packard has told the story of how the Unix windowing
system was initially created under a lax license, but was rewritten
multiple times because it got hijacked and made proprietary multiple times. An enormous amount of programming work was wasted
but because the codebase was not GPL right from the beginning.
One of the reasons why Unix was never much competition for
Windows is that many of its vendors did not work together. Linux's
GPL license nudges people into working together to save money in
total development costs and speed progress.
Some argue the lax licenses are less scary to organizations
which don't understand or truly appreciate free software. This issue
can be solved by better education of the computing community, not
by encouraging license agreements for ignorant people. As Eben
Moglen points out, things in public domain can be appropriated
from in freedom-disrespecting ways. In general, once people understand that software is a science, the idea of enabling proprietary science will not be interesting.
Further adoption of copyleft will increase the efficiency of the
free software community and help it achieve world domination
faster. Software protected by copyleft is not totally free software,
and that is why it is taking off.
78
Free Software
Why write free software?

There is only one valid definition of a business purpose: to create a customer.
Peter Drucker
What the free software movement showed was that by proving
a concept within reach and offering some working model, no
matter how defective, partial or bad, in the presence of a community sharing the objective, the achievement of the outcome is
simply a matter of allowing people to work freely with one
another.
In the world of software, which is low capital intensive in every
sense, that's all that's required for the output to take form. We
built tens, and then hundreds of billions of dollars, in valuable
software using almost no venture capital inputs.
Eben Moglen
While there are many reasons to write software, it is still worth

asking whether it will remove future motivations when the work
becomes available for others to use and build upon.
The problem with this question is that it's a trick: a person is only
motivated to write software when there is a need to improve what
currently exists. If software doesn't work properly for someone, that
is a motivation to make it better. Having a license fee for an
enhanced version is one motivation, but there can be others.
Windows and Office deliver half of Microsoft's revenues but are
less than half of their engineering. Microsoft uses their dominance
in those two markets to fund a number of other much less profitable
ventures pretty much the rest of the company. Even worse,
because Windows and Office expertise and code are locked up, the
entire industry must wait for Microsoft to improve its products on
whatever shipping cycle they choose.
In a free software world, resources can flow where they are
needed, in the quantity they are needed. Code to support a feature
gets written if the feature is important. Free software enables a
freer flow of resources than proprietary software, and therefore
enables higher quality.
The job of a business is to make satisfied customers, and free software increases possibilities for service businesses, as I will discuss
later. While law libraries and LexisNexis contain millions of freely
available legal pleadings and decisions, no one has said this has
decreased the motivation to become a lawyer! In fact, imagine trying to be a lawyer without access to these resources.
Free Software
79
1. Laws of Supply and Demand Say So

Microsoft was the company that turned the software industry
on its head by introducing lower-cost solutions years ago to
undermine the Unix businesses of IBM and Hewlett-Packard,
and the database businesses of Oracle and IBM. If anyone
knows the importance of pitching the market on low-cost, highvalue software, it's Microsoft.
Matt Asay
There are many answers to the question of motivation, but it is

important to start by mentioning that analyzing motivations is a secondary level of analysis. To understand the free market of free software you have to start with the laws of supply and demand:1
Conventional models of economics do not apply to products with zero marginal cost. The model above assumes there can be supply shortages that
would tend to increase prices.
Copying and other manipulation of bits is a computational task

whose marginal cost is the cost of the electricity, which is dropping
according Moore's law because smaller transistors use less electric1
This analysis doesn't consider the elasticity of demand, etc. but while those considerations add complications, they do not overturn the basic laws.
80
Free Software
ity. When you assume the computational cost is zero, you no longer
can have supply shortages of bits. The law of demand dictates that
as prices are lowered, demand will increase. Consequently, a product with zero cost should, in principle, have infinite demand. One
answer to why people will write free software in the future is
because there will be infinite demand for it. Wikipedia, Linux, FireFox, and many other free software products, have user bases which
are growing every year, and are even taking away marketshare from
proprietary products, as predicted.2
The law of supply says that higher prices give producers an incentive to supply more in the hope of making greater revenue. The supply curve seems to suggest that if the price of a product is zero,
producers will have no incentive to make anything. However, the
supply curve never envisioned that the marginal cost of production
would be zero. It is this difference which upends the conventional
economic rules.
Something which requires no marginal cost to make, and is
acquired by consumers for no cost, should have infinite supply and
infinite demand. Free software will take off because the most basic
laws of economics say it should. If Wikipedia charged $50 for the
privilege of browsing their encyclopedia, they would not have had
the millions of contributors and tens of millions of enhancements
they have received so far. Wikipedia's infinite supply of free bits is
creating an infinite demand.
There is the total cost of ownership (TCO), but that is a comparative measure of
software quality. If you use some free software to help you make a website, the
cost of that software is the amount of time you spend with it to accomplish your
task. If one tool builds a website in half the time, or builds one twice as nice, then
the TCO of those software packages is different, even if both are free.
If a car company gave away their top of the line cars for free, the demand would
be very high, even though the owners still had to purchase gas and had other
ongoing costs.
Even if two pieces of software have the same TCO, there is an additional cost: the
cost of switching from one to the next. Software is not interchangeable in the way
that the laws of supply and demand envision. In fact, the switching costs between
software often dwarfs the difference in their TCO.
Free Software
81
2. Services & Support

Free enterprise cannot be justified as being good for business.
It can be justified only as being good for society.
This new knowledge economy will rely heavily on knowledge
workers. The most striking growth will be in knowledge technologists: computer technicians, software designers, analysts
in clinical labs, manufacturing technologists, and paralegals.
Peter Drucker
There are many opportunities for volunteers and public institutions to create free software, but commercial enterprises need software to run their business, and they are an enormous potential
source of funding. For-profit service and support organizations will
provide the persistence that corporations will demand of free software.
As with other sciences, there should be many avenues for corporations to make money via the use and production of freely-available
advances in computer science. Service companies will write free
software when a customer pays them, just as a physicist or lawyer
does work whenever a customer pays them. In fact, by some estimates, 75% of software is written for internal use inside a corporation, without any thought of selling it to others. This corporate
software is free to its customers inside a corporation. Software
companies making licensing revenue is already a very small part of
the software industry today.
Free software is much more conducive to creating robust software services business because all of the relevant information is
publicly available. In the world of proprietary software, the company
that wrote the software is typically the only one capable of providing
support. Microsoft has created a huge software ecosystem, but the
free software service ecosystem has the potential to be much larger.
Of course, there is no guarantee of quality of service providers, but
this same issue exists today with car mechanics.
Today, many free software projects have thriving service and support communities around them. While others are not particularly
healthy yet, this is a function of the small overall marketshare of
free software, not any fundamental flaw in the business model.
In fact, the proprietary model creates fundamental limitations in
the service business. When I left Microsoft, I took on a consulting
job helping a team build a website which used Microsoft Passport as
the authentication mechanism. However, as I ran into problems,
even Google wasn't able to help because the knowledge I needed to
82
Free Software
fix my problems was locked up behind the Microsoft firewalls. Even

though I was an expert in many Microsoft technologies, I was
stymied and needed to use my old contacts to get the code working.
If I hadn't been a former employee, I would not have been able to
solve my problem. Fixing a problem in proprietary software can
sometimes feel like performing witchcraft you have to try lots of
random incantations because you can't know what is really going
on.
I spoke with an IT employee in a hospital who told me that after
her hospital purchased some software, their vendor became unresponsive. After making their money on the sale, they had no motivation to help the hospital anymore. The hospital fought with their
software supplier about whether the enhancements they were
requesting were bugs, which would be fixed quickly and for free, or
features, which would cost money and weren't guaranteed to be
made for a year or more. To the hospital, this distinction was irrelevant: lives were on the line, and they needed improvements, and
they didn't care how their vendor categorized them!
In a services business, the vendor gets paid to make the customer
happy, not before the product is in use. Restructuring the software
industry into a services business based on free software will
increase customer satisfaction and strengthen the relationship
between vendor and customer.
Free Software
83
There is an enormous market in software for running a business.

One of the biggest markets is known as Enterprise Resource Planning (ERP), an umbrella term for software used to manage the back
office of an enterprise. ERP covers everything from payroll to customers to inventory:
Category
Features
Manufacturing
Engineering, Bills of Material, Scheduling, Capacity, Workflow Management, Quality Control, Cost
Management, Manufacturing Process, Manufacturing Projects, Manufacturing Flow
Supply Chain Management
Inventory, Order Entry, Purchasing, Product Configuration, Supply Chain Planning, Supplier Scheduling, Inspection of goods, Claim Processing,
Commission Calculation
Financials
General Ledger, Cash Management, Accounts

Payable, Accounts Receivable, Fixed Assets
Projects
Costing, Billing, Time and Expense, Activity Management
Human Resources
Human Resources, Payroll, Training, Time & Attendance, Benefits
Customer Relation- Sales and Marketing, Commissions, Service, Customer Contact and Call Center support
ship Management
Data Warehouse
and various Self-Service interfaces for Customers,

Suppliers, and Employees
The major modules of Enterprise Resource Planning (ERP).
ERP is a multi-billion dollar industry, and today it is dominated by

proprietary software. This is ironic because ERP solutions, unlike a
word processor, are specifically customized to a particular business.
The need for customization suggests that ERP software could be
free, but with a robust service business behind it: customizing it,
helping import and manage the data, providing training and technical support, etc. Enhancements made by various service providers
could be fed back into the core product, and service providers could
themselves be trained and certified, thus providing revenue for the
core development.
One of the perceived weaknesses of free software is that there is
no single owner of the work, and therefore no one has skin in the
game, or a throat to choke when something goes wrong. This thinking is faulty because it assumes the computing environment is simple and homogeneous and makes about as much sense as having
one doctor for your dermatology and proctology needs. Computing
84
Free Software
systems today are not only large and complicated, they are also heterogeneous. In the 1970s, companies like IBM provided all the hardware, software, and services necessary to run a business, but
today's computing environments are very different. Even in a homogeneous Microsoft shop where all of the servers are running Windows, SQL Server, and .Net, you still might use HP hardware and
administration tools, Intel chips and drivers, an EMC disk farm, etc.
Computer software is not smart yet, but don't let that fool you
into thinking that it is not large and complicated. A trained engineer
can become an expert at 100,000 lines of code, but because a modern computer contains 100 million lines of code, you need at least
1,000 different people to help with all possible software problems.
In other words, it is a fact of life in a modern IT department that,
whether using free or proprietary software, support will require
relationships with multiple organizations.
In fact, free software can allow service and support teams to better help you because they can build expertise in more than one area
because all of the code and other relevant information is out there.
Service companies can even build a hierarchy of relationships. You
might call HP to get help with your server, and HP might have an
escalation relationship with MySQL if they track it down to a database problem they can't fix. These hierarchies can provide one
throat to choke.
3.
Lowers Hardware and Software Costs
All hardware companies have a compelling reason to use and support free software: it lowers their costs. IBM and Cray are happy to
give you a Linux OS for free, so you can put your money toward the
supercomputer they are selling. The Playstation 3 runs Linux, with
Sony's blessing, because it is another reason to buy their hardware
and take their product to new places they have yet to exploit.
Free software lowers the cost of hardware, and its greater usage
will stimulate new demand for computers and embedded devices. If
a complete, free software stack were magically available today that
enabled computer vision and speech, our toys would have them
tomorrow. A world of rich free software is a world with amazing
hardware.
Free software levels the playing field and makes the hardware
market richer and more competitive. One of the reasons an MRI
machine is expensive is because the software is proprietary. When a
hardware vendor controls the software, it can set the price at the
cost of the hardware plus the cost to develop the software, rather
Free Software
85
than something approximating their hardware costs. If MRI software were free, the hardware cost would drop, more people could
afford an MRI, and the quality would increase faster.
In fact, there already is free, high-quality scientific software suitable for building an MRI machine (in products like SciPy), but the
current manufacturers build their products using proprietary software. They aren't colluding with each other, but it reminds me of the
old days of databases where your only choice was whether to pay
many thousands for Oracle or DB2. The healthcare hardware companies had better watch their backs!
Even proprietary software companies have an incentive to use
free software, to lower their costs. It is ironic that Microsoft could
make higher profits, and build better products, by using free software.
4. Educational uses
I once asked some of my computer science lecturers why they
didn't get students to do something useful, like work on free
software, instead of assigning them pointless busy work
projects. Two main answers:
1. It's too hard to grade. (Why?)
2. It's seen by many to be exploitative. (As opposed to
busy-work?)
Dear Ken Starks (founder of Helios Project), I am sure you
strongly believe in what you are doing but I cannot either support your efforts or allow them to happen in my classroom. At
this point, I am not sure what you are doing is legal. No software is free and spreading that misconception is harmful. I
admire your attempts in getting computers in the hands of disadvantaged people but putting Linux on these machines is holding our kids back. This is a world where Windows runs on virtually every computer and putting on a carnival show for an
operating system is not helping these children at all. I am sure
if you contacted Microsoft, they would be more than happy to
supply you with copies of an older version of Windows and that
way, your computers would actually be of service to those
receiving them.
Karen, middle school teacher
Students in every field use software in their studies; this is free

brainpower! In the computer world, there are two levels of programming: the boring work that engineers do, and the fancy coding
that researchers do.
Given that software PhD-types follow the lead of other fields and
release their ideas publicly in papers, and want people to use their
86
Free Software
ideas, you'd think making their code freely available would be a part
of the computer science research culture today, but it isn't, even in
universities. There is a paper for Standford's Stanley, but no code.
Releasing software with the paper is the exception rather than the
rule today.
Even though all the key ideas are available in a paper, re-using
the ideas in such a document takes a lot more time than working
with the software directly. You can reuse software without fully
understanding it, but you can't re-implement software without fully
understanding it!
At Microsoft, both a researcher's time and his code could get allocated to a product group if anyone found their work interesting.
Researchers at Microsoft wanted to ship their code and many PhDs
joined Microsoft because they knew their work had the potential to
become widely used.3
In the future, when we get a complete set of GPL codebases, it
will get interesting very fast because researchers will realize that
the most popular free codebase is also the best one for their
research.
5.
New Sources of Revenue
Firefox made $52 million in 2005 by sharing ad revenue with

Google. Firefox made this money when their browser was configured to use the Google search service. Free programs can receive a
cut of the business they create for a service provider. Your free
photo management tool could charge for the right to be the default
printing service of their application. Widespread use of free software will create new opportunities to extract value.
6.
Fame
The self-satisfaction and adulation that people receive from producing things that others use and enjoy should not be misunderestimated, even in the world of software. It was a common feeling
among my peers at Microsoft that we should pinch ourselves
because we were getting paid to write code that millions of people
used. It is almost as much fun to be appreciated in the software
world as it is in other endeavors. I once played music for 200 people
and I didn't even care that I wasn't getting paid when a girl who
liked my music put my business card in her bra.
3
Ideally, researchers would do work directly in a product group's codebase. Unfortunately, too many product group codebases were so big, old, and complicated
that researchers typically couldn't work in them directly.
Free Software
87
7. Man's Surplus Energy
Goom is the only software project I know admired by myself, the wife, my
three-year-old son, and the mother-in-law. Dave Prince
The Linux kernel and the rest of the free software stack have
spots of brilliance in many places, but not many you can visualize.
My epiphany that Linux was going to win on the desktop happened
when I cranked up the default Linux media player. Totem doesn't
contain a web browser or try to sell you anything, and it plays a very
wide variety of formats, but what impressed me was its elegant visualization engine, not-so-elegantly named Goom.
Goom is visual proof of the proposition that with the free flow of
ideas, a clever engineer from Toulouse who you've never heard of,
was never hired by anyone, never signed a non-disclosure agreement was able to write some beautiful code that now graces millions
of computers. For all we know, he did this work in his pajamas
88
Free Software
like the bloggers who took out Dan Rather in the Rathergate4 scandal. If you want to read the code to learn its secrets, post a question
in a support forum, or send a donation to say thanks, the software
repository SourceForge enables this, and Goom is one of its 100,000
projects.
No one edits Wikipedia for fame or swooning girls. But this
energy spent is an example of the surplus intelligence of millions of
people that can be harnessed and put to work for interesting things.
The surplus intelligence of computer scientists is enough to write all
of the software we need. Programmers employed by businesses are
just icing on the cake, or to answer the phone at 3AM when a computer is sick.
8.
Increased Demand for Content
Content is king. In the digital world, content is encoded in digital

format. If that format is open, and manipulated by free software,
then the format can spread to every device and potential supporter
of the arts. Every free software tool that manipulates a format can
become a content platform. We think of music and movies as content
platforms, but they are just the final outputs of two pieces of software. Musicians could start selling the data files they use to make
their songs. This would allow people to enjoy the music of their
favorite musicians, and re-mix it in new ways.
I would like to see a huge selection of document templates for
OpenOffice, the free competitor to MS Office. Some of the template
creators could charge, or even request donations. Nothing about a
software license agreement says that the content you produce using
it must also be without cost. Free software doesn't concern itself at
all with the cost of things built with the software. We might develop
a world with lots of totally free content, but that is a completely different set of choices we would make.
If we have to choose between free software and free content, free
software is a much better choice because it will allow us to understand each other's content, which is the first step towards collaboration. A world full of free software is a world full of many different
forms of content.
The Rathergate scandal was sometimes written that way because the documents
that Dan Rather broadcast, which were supposedly from the 1970s, had centered
text, proportional fonts, and the letters 187th written with superscript. Typewriters did not have superscript back then, so they were clearly forged!
Free Software
89
Should all Ideas be Free?

Science's complexity forces people to work together. Many therefore wonder if the success of Wikipedia and YouTube suggest that all
forms of intellectual property, such as books and music, should also
be free. This will be just one of the interesting economic questions
for us to ponder in the early 21st century. As someone writing a
book, this topic became very relevant to my wallet.
Fortunately, I have concluded that there is an important distinction between static and dynamic intellectual property. A song, once
created, is not typically edited by a user. Software is often edited by
a user because it often won't work until it is fixed to support an individual scenario.
Linus created Linux, but thousands of other people have put their
stamp on it over the last 15 years, at a pace that is only increasing,
improving on his humble creation to make it work better for themselves. Books, music and movies are often made by one person, or
small teams of people, and are meant to be used as-is without the
need for modification. (There may be a need to translate or re-dub a
movie or song into other languages, but this doesn't require the
same creative intellectual energy, and isn't something that requires
further improvements over time.)
Writing has been a way to make a living for many hundreds of
years, and I am not ready to propose that we tweak society in such a
fundamental way and remove longstanding incentives to create nonscience. There is no such thing as a proprietary book, as by definition you can see what it contains and build on top of the ideas in it.
Richard Stallman talks about a world where the music is free, but
music players have a button where you can click to send $1 directly
to the artist when you hear a song that you like. Such a world could
provide a sufficient incentive to create free music, and it could be
the choice of artists to create such an arrangement. We can create
arbitrary rules, so let's experiment!
Video Games and New Media

While books, music and movies are mostly static creations, video
games are a mix of software and static multimedia. Video games
contain the game engine that models the virtual world, but also contains 3-D models, music, a storyline, which are meant to be enjoyed
as-is. Therefore, video games should continue to be sold for money
to pay for the art, but the game engines could be GPL.
90
Free Software
Making game engines free will allow for much more user-created
content. iD Software, creators of the popular game Doom, did
release their game engine, though it is a very old version of their
code which they no longer use. Even so, a free software community
has surrounded it! This version of Doom has never been used in any
Xbox 360 game, but you can run it on an iPod.
Game creators today keep their game engine private because
they consider it a competitive advantage but it is also a huge part
of their costs. The Halo series of video games, the most popular on
the Xbox platform, spends three years between releases. A big part
of the holdup is the software engine development.
A game engine is a piece of technology comparable in size to the
Linux kernel because developers need lots of features:
the ability to model a world, create the rules of the game,
and efficiently send updates between computers
text to speech, means to deal with cheaters, and other forms
of AI.
If the thousands of game programmers around the world started
working together on the software, I shudder to think what they
might build! Hollywood might increase their investments in the
game business if there were a free standard codebase: movies, video
games and virtual reality are closely-related technologies.
Pride of Ownership
While much free software is not paid for or organized by a company, the quality of it is just as good because the programmers do so
with the same pride as they do a day job.
When I joined Microsoft, I casually signed away the right to use
my work anywhere else. Despite the fact that I no longer owned my
work, and would never see the code again after I left the company, I
didn't treat it with any less importance. It is in the nature of man to
do his best whether or not he exclusively owns an idea.
In fact, unpaid volunteer programmers might be more tempted to
do it right because there isn't a deadline. A lot of lesser-known free
software in use today was started purely for the enjoyment of the
programmer. Linus wrote Linux to learn about his computer, and
even today, 25% of contributors to the kernel do not work for a company.
A programmer might have a boring job building websites something that pays the bills, but writing fun code in his free time might
Free Software
91
provide more of a personal reward. There is a free physics engine

used in video games called Open Dynamic Engine maintained by a
PhD in his free time, while he works for Google during the day.
Gcompris is a free educational program created by Bruno Coudoin
for his two children.5 He didn't need to make it perfect, because
other programmers who recently became fathers showed up to help!
In proprietary software, the pride motivation to do the right thing
is actually one step removed because you write software in the hope
that you can ship it and sell it. And that urge to get it out there may
encourage you to cut corners.
In a free software environment, you write code because there is a
need an itch to scratch. This encourages you not to overbuild
something that is of no use to you like the Microsoft Office paperclip
character.
Where Does Vision Fit In?

Ideas are somewhat like babies they are born small, immature, and shapeless. They are promise rather than fulfillment.
In the innovative company executives do not say, This is a
damn-fool idea. Instead they ask, What would be needed to
make this embryonic, half-baked, foolish idea into something
that makes sense, that is an opportunity for us?
Peter Drucker
In a world of free software where there is no strong leadership

from the top, where does vision fit in? The answer is that there is
vision in the big, and vision in the small. Vision in the big is when
someone has an idea for a new project, describes this vision to others, creates a proof of concept, and an environment that encourages
people to join the project. However, these new, big efforts like
Wikipedia and the Linux kernel are very infrequently launched.
In actuality, on a daily basis, team members struggle to refine a
product to make it better achieve that original vision. This refining
activity requires lots of little work items with each step requiring a
tiny bit of visionary insight moving the product forward in support of
that big idea. In creating free products that anyone can contribute
to, progress toward big visionary efforts will happen faster.
Note, free software doesn't imply that obstacles like ego and stubbornness will disappear, but it is also no different than what exists
everywhere else.
5
I did check to verify that Bruno had children, but like changing diapers, voluntarily writing free educational software for children is something only a parent would
do!
92
Free Software
Governments and Free Software
Comic from xkcd.com
Documents produced as official duties of an employee of the US

government are assumed to be in the public domain, unless classified. While creation of free software for profit-seeking corporations
might still be up for debate, governments and other publicly funded
institutions can embrace free software.
That works sponsored by a government should be placed in the
public domain reminds us of their public nature. Copyleft is the 21st
century equivalent of public domain. It is in the spirit of public
domain, but provides better protection to the intellectual property.
The U. S. government is starting to understand this: NASA, the
Department of Defense, and the NSA are making steady movement
towards free software, which should also demonstrate to everyone
that free software can meet even the most extreme requirements of
security and reliability.
Free Software
93
Why is free software secure? A common principle in software is

that security through obscurity is a bad idea. For example, the
most important security encryption algorithms today are designed
and analyzed in public and have free software implementations:
Math applied during encryption of data. The security of encryption algorithms is guaranteed by the many smart eyeballs who have analyzed it. Ultimately, if the mathematicians prove an encryption algorithm is secure,
there are no back doors, and a password is the only key to the data.
Would the GPL require that the U.S. government give away its top
secret code? The GPL's goal is to ensure that all users of software
have the right to inspect and make changes. The user of military
software is the military itself, so the conditions are met.
The U.S. government might not want to give away certain source
code to other countries, but this is also solvable. Since enforcement
of copyright is the right granted to the U.S. Congress, they can create a law that says any GPL software stamped Top Secret is
removed from copyleft obligations, something that should apply only
to a small amount of code.
A question to ponder: If the military were to create a vision system to help it build more discriminate bombs less likely to hit civilians, would the military be inclined to give this away or would they
be afraid that the enemy would use it to build more discriminate
94
Free Software
IEDs? Software has so many different applications and purposes, it

will be quite difficult for the government to make the correct tradeoffs.
Because GPL should be usable by all of the U.S. government's
code, I would like to see them endorse it for all publicly-funded software development. This would also help to ensure that the U.S. is
not left behind when free software achieves world domination.
Should all Software be GPL?

The era of procrastination, of half-measures, of soothing and
baffling expedients, of delays, is coming to a close. In its place
we are entering a period of consequences.
Winston Churchill
I don't envision a world where all software need be free, but I do

believe that all the world's large, interesting software projects
require a collaborative effort and should therefore be built by a community and protected by a copyleft license.
There is a lot of generic mass in the lower layers of software.
Eventually as you move up the stack, you reach areas that are specific to a particular business. At this point the importance of cooperation decreases, but it is also something which isn't even useful to
competitors. In other words, whether in the lower or upper layers,
all software may as well be free.
It is still the early days in the coming gold rush era of free software, and those who understand this fact first will gain all the users,
and get all the help. While the Linux has locked in its spot, most of
the rest of the free software stack isn't in this position yet, although
many, like Firefox and OpenOffice, are close.
One other relevant point is that, the sooner the free software
community is brought in, the better off the codebase will be. There
are a number of products that started out non-free and were only
given to the community much later. By this time, one usually finds
that the codebase is a complicated mess, and it takes a while for the
community to understand it and clean things up. Mozilla's Firefox
and IBM's Eclipse Java IDE are examples of products that took a
while to take off after being liberated.
Free Software
95
Microsoft's Responses to Free

Software
1. Windows + Office = $3
Its easier for our software to compete with Linux when theres
piracy than when theres not.
Bill Gates
One of Microsoft's responses to free software has been to lower

prices. Microsoft has created student and community editions of
certain products that were low in cost but also often crippled.
Microsoft also announced a bundle, available to qualifying thirdworld countries, that includes a low-end Windows, a low-end Office,
and a few other items for just $3. There are fewer than one billion
people with PCs, and so presumably a cheap bundle will make it easier for the remaining five billion people to get PCs. For potential customers there are several things wrong with this approach:
1. A Linux OS comes with a much larger set of applications than
is included this bundle.
2. The source code is not available, so engineers in these
emerging markets cannot build up a service industry of companies who support and customize this software.
3. Customers become locked into Microsoft file formats and
other proprietary technologies, and the countries become
dependent on Microsoft for future progress.
4. The prices of the rest of Microsoft's software, from servers to
developments tools, have not been lowered. This three-dollar
bundle only gives users access to a small fraction of
Microsoft's technologies.
Of course the $3 bundle is a very smart defensive move by Microsoft. It is charitable of them to lower their price, however, the move
might be considered dumping, which Wikipedia defines as an act
of a manufacturer in one country exporting a product to another
country at a price that is either below the price it charges in its
home market or is below its costs of production.
The low-cost bundle is a danger to the free software movement. I
can think of nothing better than having millions of programmers
around the world build their industrial base around a free license.
Emerging markets should start off their economies on the right foot
with software that's completely free.
96
Free Software
If it weren't for piracy, Linux would have likely taken over the
world already. I have been told that more than 90% of users in China
run pirated software, and as no one can truly know, it could very
well be closer to 99%. If it weren't for these illegal copies, China
would have been forced to develop a suitable free software stack.
And once something existed for a country such as China, it would
have been usable for the rest of the world as well.
The United States needs to move rapidly towards free software if
it is to be relevant in building the future. The U.S. invented the transistor, but it has also spawned a ton of old, proprietary systems that
are a drag on future productivity. America is the widespread purveyor of non-free software which means the transition to free software will be much more difficult than for other countries which
don't have this baggage.
2.
CodePlex
Microsoft has also created a number of websites where developers can use free code and collaborate, and the latest is called CodePlex. While it does demonstrate that Microsoft understands the
benefits of free software, this website mostly contains tiny add-ons
to proprietary Microsoft products. CodePlex may serve mostly to kill
off Microsoft's community of partners who previously sold add-ons
to Visual Basic and other products. While these serve as a bulwark
against the free software movement and provide a way for Microsoft
to claim that they get this new way of developing software, it ultimately undermines their business.
3.
Interop
Another way Microsoft has responded to requests for increased

open-ness is by publishing specifications for certain file formats and
protocols. While a spec is a document, and a promise not to sue, it
still requires code, or reverse-engineering in order to put this idea
to use on a computer. A spec attached to proprietary software is a
spec with a huge cost attached to it: the cost to build the free baseline implementation.
Someone who wants to improve the C# programming language
doesn't want to start with the spec, but with a compiler for that language. In fact, the spec could be generated from the compiler's
source code.
Free Software
97
Likewise, the best way to make sure that a new audio format
becomes available on every device is to make freely available the
code to read and write it. If the details are public, why not make
them useful to a computer?
4. Shared Source
Microsoft has also released some software under various shared
source licenses. One of the first products Microsoft released under
this license was C# and a portion of the .Net runtime. The language
spec was always free, and there was a decision made to release
some of the code as well in 2002. In addition to releasing the code,
Microsoft seeded research efforts by sponsoring 80 projects in universities around the world. However, there is little activity today,
and one reason is that the license is very restrictive:
You may not use or distribute this Software or any derivative
works in any form for commercial purposes. Examples of commercial purposes would be running business operations, licensing, leasing, or selling the Software, or distributing the
Software for use with commercial products.
Microsoft Shared Source CLI License
This is a license that shuns capitalism.
There is a free project managed by Novell called Mono which has

reverse-engineered the C# runtime and libraries from the public
spec. It has gained interest in a number of communities, and this
runtime, not Microsoft's, provides a way to run C# code on the Mac
and Linux. This free implementation could eventually become the de
facto one!
Just a Stab
Richard Stallman, who started the free software movement in
1985, might be right when he says that the freer intellectual property is, the better off society is.
However, I don't think this should automatically apply to everything, even software. Furthermore, it is a choice for every creator to
make. While it might make economic and moral sense for some
ideas to be given away, that doesn't mean ideas are no longer
owned. The GPL says that software is owned by its users.
Stallman reminds us that the concept of a free market is an idea
that has taken a long time for the general public to understand and
free software and the other intellectual property issues we grapple
with today will also take time for society to grasp.
98
Free Software
Computer pundit Tim O'Reilly makes the point that the GPL could
become irrelevant in the coming cloud computing world. Nowadays,
people focus on free code running on their own computer, but what
about if you are using GPL code which is doing work on your behalf,
but running on another processor? Currently, the GPL does not consider this scenario, but I think this is a loophole not within the spirit
of copyleft. Perhaps this will be the battle for GPL v4, some years
hence.
Patents & Copyright
99
PATENTS & COPYRIGHT

It has always been a strong goal of the JPEG committee that its
standards should be implementable in their baseline form without payment of royalty and license fees. The up and coming
JPEG 2000 standard has been prepared along these lines, and
agreement reached with over 20 large organizations holding
many patents in this area to allow use of their intellectual property in connection with the standard without payment of license
fees or royalties.
Statement by JPEG committee
If people had understood how patents would be granted when

most of today's ideas were invented, and had taken out patents,
the industry would be at a complete standstill today.
The solution is patenting as much as we can. A future startup
with no patents of its own will be forced to pay whatever price
the giants choose to impose. That price might be high. Established companies have an interest in excluding future competitors.
Bill Gates, internal memo, 1991
n February 2007, Microsoft lost a $1.5 billion judgment for

infringing on MP3 patents, even though Microsoft had licensed
MP3! The problem with MP3 is that multiple companies have
patent claims, which makes the situation a particular mess. We can
argue as to whether patents add value, but they indisputably add
costs.
The Software Freedom Law Center calculates that Microsoft has
paid $4 billion in lawsuits in the last three years to companies who
have won patent claims against the company, which works out to a
tax of $20 per copy of Windows. (The costs are actually higher
because some of the settlements are secret.)
As a Microsoft employee, if they filed a patent on technology you
devised, you received a three-inch marble keepsake cube and a few
hundred dollars. I remember interviewing for a new position within
Microsoft and feeling my hands sweat when I saw a stack of cubes
behind my interrogator. I received only two patents, late in my
career, so I always felt a tinge of jealousy when I saw someone else's
patents. On the other hand, I considered myself merely unlucky that
the features I was assigned happened to be undeserving of a patent.
My friend Alex Mogilevsky added background spell checking to
Microsoft Word '95 which draws red squiggly underlines below misspelled words. This is a feature we are all very familiar with now but
100
Patents & Copyright
which was the biggest enhancement of that release and some say
this was the most useful feature ever added to Word. In the end,
Alex received U. S. patent #5,787,451, but was this feature truly
worthy of a patent? These are the major elements of this patent:
Red underlines of misspelled words
Spell checking happens as you type, removing the need to
launch a dialog box as a separate step.
While adding this feature was a huge time-saving device, it isn't
something so unique that other word processors wouldn't have
eventually implemented it. Teachers have been circling misspelled
words with red pens since time immemorial, this is just the digital
version. Fortunately for the world, Microsoft has not enforced this
patent and squiggly underlines can now show up almost everywhere
you can type text.
For several years, British Telecom attempted to assert ownership
on the concept of the hyperlink in patent #4,873,662. Thankfully,
that patent was eventually invalidated on a technicality, but a lot of
money was spent on lawyers in the meanwhile.
One of Amazon's first patents was for 1-Click ordering. Once
Amazon has your payment and shipping information on file, you are
able to purchase a book with literally one click. However, isn't this
an obvious innovation for anyone building an e-commerce website?
Zero-click ordering would be an innovation worth patenting!
Amazon's patent didn't encourage innovation, it simply became a
stick their lawyers could use to beat up Barnes & Noble. We are told
that patents protect the little guy, but they actually create a complicated minefield that helps incumbents.
Patents & Copyright
101
One of the biggest areas of patent headaches for the computer

industry today deals with codecs (compression decompression),
little pieces of software that compress and decompress sound and
images. Patenting codecs is a bad idea because the fundamentals of
all codecs are the same:
There are an infinite number of ways of converting sound to and from bits,
but they are mathematically very similar. (The difference between codecs
has to do with merely their efficiency, and their cleverness in removing data
you cannot perceive.)
102
Patents & Copyright
There might be a new type of compression algorithm that is innovative, but the point of codecs is to enable the easy exchange of
video and audio bits. Patents, therefore, only serve as a hindrance to
this. The reason why digital audio and video is such a hassle today is
because of the mess of proprietary formats, patents and licensing
fees. These obstacles encourage the creation of even more formats,
which just makes the problem worse.
In the mid-90s, Apple, Microsoft, Real, and others were out there
hawking their proprietary audio and video formats, touting their
advantages over the others. We have not recovered from this. Microsoft employee Ben Waggoner wrote:
Microsoft (well before my time) went down the codec standard
route before with MPEG-4 part 2, which turns out to be a
profound disappointment across the industry it didn't offer
that much of a compression advantage over MPEG-2, and the
protracted license agreement discussions scared off a lot of
adoption. I was involved in many digital media projects that
wouldn't even touch MPEG-4 in the late '90s to early '00s
because there was going to be a 'content fee' that hadn't been
fully defined yet.
And even when they created standards like MPEG, certain companies would assert patent control over certain aspects. MPEG isn't a
codec so much as a system of codecs, a land mine of proprietary and
non-proprietary specifications that makes supporting MPEG very
difficult. The reason many websites do their video using the proprietary Flash control is because the various interests didn't come
together to produce a standard.
Many times in this industry, someone has invented a compression
mechanism, patented it, implemented the code for their own use,
but did not document it or give away code to encode and decode the
format. Then they asked everyone to use their new format. This
strategy is totally the wrong approach to making formats universally
usable by computers and devices.
What is important is that we pick a simple and efficient algorithm,
standardize it and then make the software to read and write it freely
available. That way, every device and every application will work
with every piece of sound or video. Today, there is nothing but chaos
and incompatibility.
The most popular audio format today is MP3. Here there is not
just one, but a number of different companies that have patent
claims that do not expire until 2015! The core logic of a codec is
Patents & Copyright
103
only a few thousand lines of software; this entire mess is over

details too small to disagree over, yet somehow we do.1 Patents and
standards serve diametrically opposite purposes.
Software is math
Software is math. In the 1930s, Alonzo Church created a mathematical system known as lambda () calculus, an early programming
language that used math as its foundation, and which could express
any program written today.
A patent on software is therefore a patent on math, something
that historically has not been patentable. Donald Knuth, one of
America's most preeminent computer scientists, wrote in a letter to
the U. S. Patent Office in 2003:
I am told that the courts are trying to make a distinction
between mathematical algorithms and non mathematical algorithms. To a computer scientist, this makes no sense, because
every algorithm is as mathematical as anything could be. An
algorithm is an abstract concept unrelated to the physical laws
of the universe.
Nor is it possible to distinguish between numerical and nonnumerical algorithms, as if numbers were somehow different
from other kinds of precise information. All data are numbers,
and all numbers are data.
Congress wisely decided long ago that mathematical things
cannot be patented. Surely nobody could apply mathematics if
it were necessary to pay a license fee whenever the theorem of
Pythagoras is employed. The basic algorithmic ideas that people are now rushing to patent are so fundamental, the result
threatens to be like what would happen if we allowed authors
to have patents on individual words and concepts.
I strongly believe that the recent trend to patenting algorithms
is of benefit only to a very small number of attorneys and inventors, while it is seriously harmful to the vast majority of people
who want to do useful things with computers.
Software doesn't look like math, but it is built up from just a few
primitive operations that have a mathematical foundation. Allowing
people to patent particular algorithms just means that a section of
our math is now owned by someone. If all of these patents become
owned by many different entities, then math could become unusable
by anyone. Then where would we be?
The free Vorbis decoder for embedded devices is less than 7,000 lines of code,
and much of that code is infrastructure logic to do, for example, all math using
integers because low-end processors often do not have floating point capabilities.
104
Patents & Copyright
Ironically, when defending itself in a patent infringement case

with AT&T, Microsoft argued that software cannot be patented:
Justice Scalia: You can't patent, you know, on-off, on-off code
in the abstract, can you?
Microsoft attorney, Ted Olson: That's correct, Justice Scalia.
[...] An idea or a principle, two plus two equals four, can't be
patented.
The Supreme Court ruled that unincorporated software, because

it is intangible information, cannot be typed a 'component' of an
invention. One might read this and conclude that the Supreme
Court, unknowingly, and at Microsoft's behest, outlawed software
patents!2
Justice Alito, in a concurring opinion states: To be sure, if these computers could

not run Windows without inserting and keeping a CD-ROM in the appropriate
drive, then the CD-ROMs might be components of the computer. But that is not
the case here. Actually, Linux Live CDs allow you to run the OS directly off the
CD-ROM. I don't see a distinction between using a CD-ROM every time, or using
it once and copying it to a faster medium and then discarding it. Reading the
decision made my head hurt.
Patents & Copyright
105
Software is big
Beyond software being math, software also differs from things
that were patented previously. In the biotechnology world, and even
the world of Edison, a patent typically covers one product:
Light bulb patent No. 223,898
Viagra Patent No. 6,469,012
Usually, a patent is one product.
Software, however, is different because it is enormous and

incorporates thousands of differing ideas. Software products today
are not patented in their entirety, only tiny portions of them. Richard
Stallman compares software to music: imagine if people could
patent chords, how would that affect our ability to create new
music?
Jerry Baker, Senior VP of Oracle wrote:
Our engineers and patent counsel have advised me that it may
be virtually impossible to develop a complicated software product today without infringing numerous broad existing patents.
106
Patents & Copyright
Microsoft has recently claimed that Linux violates 235 of its

patents spread throughout the software stack. This is an enormous
number; Verizon has threatened to put Vonage out of business for
violating just three obvious patents:
1.
2.
3.
Internet to phone-system connection technology

Internet phone features such as voice mail and callwaiting
Wireless to Internet phone calls
Software is a fast-moving industry

In the drug industry, it takes years of clinical trials to prove a
medicine. The Tufts Center For the Study of Drug Development
reported recently that the average time to get approvals for drugs
was 6 years. The current patent length of 17 years is an eternity for
the fast-moving field of computing. In his letter to the Patent Office,
software scientist Knuth also wrote:
Software patents will have the effect of freezing progress at
essentially its current level. If present trends continue, the only
recourse available to the majority of America's brilliant software developers will be to give up software or to emigrate.
Therefore, even if you believe in software patents, shrinking their

length of exclusive ownership to just a few years would be a compromise. Decreasing the duration of protection would also decrease
the number of spurious patents, which would make it easier to be in
compliance with all of those that are out there.
Copyright provides sufficient

protection
Patents are a powerful right because they give their owners
exclusive access to an idea. Proving that you invented an idea independently is not a defense. A much less exclusive right is what is
allowed in copyright law. Copyright law protects someone from
stealing words or code, but if you can prove you came up with it via
independent means, you have a sufficient defense.
In the old days of software, Word and WordPerfect kept adding
each other's features in order to convert users. They didn't need
exclusive access to an idea to be motivated to write software, and
patents would have decreased the level of competition.
Giving exclusive access to an idea can encourage people to rest
on their laurels, or even be a squatter on an idea, taking money from
Patents & Copyright
107
anyone who happens to run across it, but not using it for their own
purposes. Many times, patents are created merely as a defensive
measure against other companies. A company will patent things
with the hope it can trip up anyone who might come calling with
claims against them.
In a world filled with free software, it is the copyleft mechanism,
not the patent mechanism, that will provide protection for software.
Even proprietary software would not stop improving if software
patents disappeared, though their lawyers would scream like stuck
pigs.
Conclusion
In the early days of cars, there were patent lawsuits sent between
the lawyers, on horseback of course:
George Selden, the man who patented the car in 1895, didn't sell one until
14 years later. Was he just a squatter on this idea? The magic of the early
car was the internal combustion engine, which Selden did not invent, and
which required only the science of fire, something man exploited long
before he understood it.
108
Patents & Copyright
I believe research will show that 99% of software patents today

are natural extensions of an existing idea. Even for the few truly
unique ideas, it is probably better that the concept be widely available and used by many rather than guarded and stifled by a few.
If someone invents the software equivalent of an idea like E=MC2,
do we really want just one entity to own that idea? Would Einstein
want that? Anyone who supports software patents should hold up a
good one and explain how it is actually unique and yet an idea that
only one company should own.
If we outlawed software patents, the pace of progress in software
would increase, the squatters and their lawyers would disappear,
legal uncertainties surrounding software technology would
decrease, and there would still be many other motivations to write
software. In fact, many of the hassles in computing, like playing
MP3s and DVDs, exist because of patent issues. As of 2005, the U.S.
Patent and Trademark Office has granted 265,000 software patents,
a very large minefield.
There is a lawsuit known as In re Bilski that is now before the
Supreme Court (oral arguments on November 9, 2009) and which
has the possibility to outlaw software patents. The details of the lawsuit are tedious but the case has epic implications.
Patents & Copyright
109
Biotechnology Patents
There are more than 50 proteins possibly involved in cancer
that the company is not working on because the patent holders
either would not allow it or were demanding unreasonable royalties.
Peter Ringrose, Chief Scientific Officer, Bristol-Myers Squib
Pharmaceutical TV Ads
Side effects, warn ings, and disclaimers

Happy people walking around outdoors
Telling you what
the drug actually
does
Created by Zorak of graphjam.com
Chart of the content in pharmaceutical TV ads
Although this book is about software, it is worthwhile to look at

patents for biotechnology. Michael Crichton recently argued that
because companies have been able to patent genes, that are not
even invented by man, biotechnology research is threatened. Patenting drugs is one thing; patenting genes is something entirely different.
Against Intellectual Monopoly, by Michele Boldrin and David K.
Levine, is an interesting book that discusses many patent and copyright issues. It describes the case for removing patent protection for
pharmaceutical companies that produce relatively simple chemicals
requiring extensive research and testing. This is a brief summary of
their arguments against patent protection for drugs:
Before the pharmaceutical industry came along, the most
important chemical industry was for paints and dyes. The
authors' study of the history of those companies shows that
companies in countries with patent protection had slower
rates of innovation. One of the biggest reasons is because
110
Patents & Copyright

the chemical industry is a classic case of innovation chains
new compounds and processes built on the knowledge of
existing ones. Patents, therefore, get in the way of this.
Countries such as Switzerland and Italy, which for many
years didn't have patent protection on pharmaceuticals, still
had robust industries. Adding patent protection has actually
hurt investments in those countries' pharmaceutical
industries.
Pharmaceutical companies spend 13% of revenues on R&D
and spend twice as much on marketing and promotion. The
cost of R&D is not such a big deal as the drug companies say
it is.
Two-thirds of the money spent on pharmaceutical research in
the United States comes from the government. This public
money is getting turned into private patents.
Many drugs are inexpensive to produce, but drug companies
do not make them available to poor countries. If they did,
they could be re-imported, hurting their profits in the higherpriced Western markets. Furthermore, the fact that there are
such large profits to be made in Western markets has discouraged the development and production of drugs to treat
diseases plaguing the poor countries of Africa and Latin
America.
A study by the U. S. Centers for Disease Control and Prevention (CDC) found that of the top ten public health achievements of the 20th century in the United States, none was
motivated by a desire to patent.
The current patent system perverts pharmaceutical companies and makes them focus on copycat drugs. According to
an article from The New Republic:
Turn on your television and note which drugs are being marketed most aggressively. Ads for Celebrex may imply that it
will enable arthritics to jump rope, but the drug actually
relieves pain no better than basic ibuprofen; its principal
supposed benefit is causing fewer ulcers, but the FDA
recently rejected even that claim. Clarinex is a differently
packaged version of Claritin, which is of questionable efficacy in the first place and is sold over the counter abroad
for vastly less. Promoted as though it must be some sort of
elixir, the ubiquitous purple pill, Nexium, is essentially
AstraZeneca's old heartburn drug Prilosec with a minor
chemical twist that allowed the company to extend its
patent.
Patents & Copyright
111
Money spent on research and marketing of these me-too

drugs is money not being efficiently allocated. From Boldrin
and Levine:
Monopolies innovate as little as possible and only when
forced to; in general they would rather spend time seeking
rents via political protection, while trying to sell, at a high
price, their old refurbished products to powerless consumers, via massive doses of advertising.
The largest cost of making new drugs is clinical trials. In

some cases this can cost $800 million, and in others $6.5 million. Many of the clinical trials are for me-too drugs, which
are a waste. Furthermore, companies paying for clinical trials have a conflict of interest. Boldrin and Levine argue that
clinical trials should be paid for by competitive NIH grants.
The information about the effects of a drug is a public good
and therefore should be paid for by the public.
There are three stages of clinical trials, and so the authors
suggest drug companies pay for stage I, but that public
money be used for stages II and III. Taking this cost away
would allow the drug companies to focus more on fundamental R&D. The biotechnology firms still have a long way to go
in their understanding of chemistry and DNA.
Boldrin and Levine don't necessarily suggest that patents be eliminated; even shortening the period of protection would decrease the
adverse affects of the above issues. Furthermore, even if we assume
that lowering or removing patent protection will hurt investments,
the government could give R&D tax credits to offset the decrease.
Finally, they argue that too many drugs are available by prescription
only, which increases costs and lowers revenues for the pharmaceutical companies.
Congress, in mid-2007, was considering some patent reforms.
Some of the changes included changing the patent system to grant
them to the first one to file, not the first one to invent, and to limit
the amount of infringement damages. Both of these ideas are bad,
however, and are just nibbling around the edges. Boldrin and Levine
argue that the following reforms should be considered:
Proof that you independently invented something should be a
defense against a patent infringement.
Licensing fees for patents should be set close to R&D costs.
Have the government fund clinical trials.
Reduce the number of drugs requiring a prescription.
112
Patents & Copyright
Decrease the term of patents to four years.

Suspend drug patents in poor countries. To prevent drug reimportation, suspend free trade for medicines.
Boldrin and Levine argue that rather than allowing for the creation of patents, subsidizing a portion of R&D via tax credits is a
more efficient allocation of public resources.
Openness in Health Care

Throughout this book, I advocate cooperation in the future development of software. While doing research for this book, I came
across a report by the Committee for Economic Development (CED),
an independent, non-profit, non-partisan think tank that wrote a
report entitled: Harnessing Openness to Transform American
Health Care. According to the report:
$2 Trillion dollars, 16-17% of our GDP, is spent on health care
overuse, underuse, misuse, duplication, system failure, unnecessary repetition, poor communications and inefficiency.
The U.S. employer-based health care system is failing. Health
care costs are rising faster than wages, quality of care is low,
and access to coverage is deteriorating.
The report details ways in which better cooperation can improve

healthcare. For instance: The search for the human genome
sequence was a competitive race. Celera, a private-sector firm,
sought to be first to establish the sequence while, as was the norm,
keeping much of its data private, to be made available on commercial terms to other researchers.
In contrast, the publicly funded Human Genome Project (HGP)
followed an open model making its data publicly available and welcoming input from around the world. HGP pushed participating
researchers to disclose their findings as quickly as possible. While
Celera made important contributions to the sequencing, it was the
HGPs model of discovery that has transformed the research process
by reducing transaction costs and secrecy that may impede followon research.
The HGP researchers not only put raw sequencing data into the
public domain, but as the data were being produced, an opensource software program known as the distributed annotation
system (DAS) was set up to facilitate collaborative improvement
and annotation of the genome. This allowed any researcher to
choose the annotation they wanted to view and enabled the
ranking of annotations by the number of researchers that used
them, something akin to Googles methods for ranking search
results.
Patents & Copyright
113
This open model is now being used in a federally funded international effort to create a map of haplotypes (HapMap) which
describes variations in the human genome that tend to occur
together in neighborhoods or haplotypes. Data about the
genotype of the individual haplotypes is being released publicly
as soon as it is indentified. The openness of the HapMap effort
is reinforced by its use of a licensing system that is self-consciously modeled on the copyleft system of open-source software licensing and which prevents those who utilize the data
from attempting to close it to others via patents.
Utilizing the results of the HapMap process, a public-private
partnership, the SNP Consortium, is identifying panels of a few
hundred thousand single-nucleotide polymorphisms (SNPs) that
can be used to identify common variants in an individuals
entire 3-billion base-pair genome that might be associated with
a disease. As with the HapMap project, participants in the consortium have agreed to put the data they produce in to the public domain.
In the reasonably near future, according to Dr. Francis Collins,
leader of the National Human Genome Research Institute
(NHGRI) in the National Institutes of Health (NIH), the
HapMap should help make practical case-controlled studies
using SNPs to identify gene variants that contribute to diabetes, heart disease, Alzheimer disease, common cancers, mental illness, hypertension, asthma, and a host of other common
disorders. That future seems nearer than ever today with scientists finding correlations between diseases such as multiple
sclerosis and breast cancer and specific genetic variations.
To harness the power of openness to speed the development of

medicines and vaccines for less lucrative commercial markets, the
Bill & Melinda Gates Foundation now conditions its grants to
require researchers to share their results promptly so that others
can build on successes, avoid pitfalls, and eliminate redundancy. It
appears that Bill believes in openness for medical research, but not
for software yet.
114
Patents & Copyright
The CED reports concludes:

Openness is ultimately about an attitude that sees the opportunity for many to benefit from greater access to information, as
well as to contribute much to the benefit of us all. Greater
openness is likely to become increasingly important in more
and more areas driven by the relentless progress of information
and communications technology. We offer these recommendations with the hope that modest changes based on greater
access to information by more people, and more possibilities for
them to contribute based on their own expertise and energy,
can help improve health care in the United States and around
the world.
Currently, a lot of bioinformatics software is proprietary, following

the current path of PC software, but that is likewise slowing
progress.
The Scope of Copyright

The Congress shall have Power...To promote the Progress of
Science and useful Arts, by securing for limited Times to
Authors and Inventors the exclusive Right to their respective
Writings and Discoveries
U. S. Constitution, Section 8
The utility of patents is a big debatable question, but few argue

against copyright law. Copyright law is important even for the copyleft license because it allows for enforcement of the terms.
The two biggest issues surrounding copyright law are the duration of the copyright and your ability to do things outside the scope
of copyright, i.e. I'm legally using your copyrighted materials without needing to get your permission.
Length of Copyright
Our Founding Fathers had it right, once again, when they determined that authors and inventors should have exclusive rights to
their creations for limited times. Their thinking suggests a presumption that ideas will eventually flow into the public domain
because, at some point, increasing the years of protection isn't guaranteed to promote further progress; instead, it may serve as an
impediment.
When copyright was first created in the U. S., the term was 14
years, renewable once if the author was still alive. Today, the time
frame that has been chosen for U.S. copyright law is the life of the
author plus 70 years. This value was extended from the life of the
Patents & Copyright
115
author plus 50 years in 1998. (In Eldred v. Ashcroft, the U.S.

Supreme Court upheld the constitutionality of this law. The way to
fix copyright law is via Congress, not via the courts.)
It is hard to believe that people are motivated to make music so
that they can sell it after they are dead, so one could say that this
value is not limited times. Shortening the copyright law perhaps
decreases the motivation to create things, but it also allows more
people to have access to things they would never have otherwise.
We need a new compromise and having a copyright term of 10 years
should be the starting point for discussion. (In engineering, back of
the envelope estimates are a good basis for analysis. If 10 years is
too short, then next try 20, 40 and 80 years.)
Do you still buy 80s music? 90s music? How many of your movies,
music and books were less than ten years old when you purchased
them? I do not own any Beatles albums and I would never buy one
even if money were burning a whole in my pocket. The only way I
will hear Beatles music is if I can listen to it for free. I'll bet that
most things purchased at Amazon were produced in the last ten
years. Amazon could provide useful data to Congress on these
points; the writers of the Constitution didn't define copyright length
because they didn't have such data available to them.
When the term of copyright expires, it does not mean that no one
will sell the newly public domain item anymore. In fact, companies
might be willing to make CDs and DVDs of things in the public
domain for prices close to the marginal costs, yet still make a profit.
If you set the term of copyright beyond the point of drop off in
sales, you will not have materially hurt the artists, but you will have
created a richer world; the benefit of a free culture is that the next
generation of budding artists will have so much greater access to
art and they will become better artists for it.
116
Patents & Copyright
Fair Use
The fair use clause of copyright allows you to use copyrighted
materials for limited purposes without getting permission. This right
was recognized by the courts in common law and was incorporated
into U. S. Copyright law in 1976:
The fair use of a copyrighted work, including such use by
reproduction in copies or prerecords or by any other means
specified by that section, for purposes such as criticism,
comment, news reporting, teaching (including multiple copies
for classroom use), scholarship, or research, is not an
infringement of copyright. In determining whether the use
made of a work in any particular case is a fair use the factors to
be considered shall include
1. the purpose and character of the use, including whether
such use is of a commercial nature or is for nonprofit
educational purposes;
2. the nature of the copyrighted work;
3. the amount and substantiality of the portion used in
relation to the copyrighted work as a whole; and
4. the effect of the use upon the potential market for, or value
of, the copyrighted work.
There are questions unanswered here. Am I allowed to: make a

mix-CD and give it to ten of my friends? Let people download it via
the Internet? Sell it for shipping and handling plus $1?
The biggest problem with fair use is that it is biased against the
commercial domain. Fair use was not meant to take money out of art
which reuses other art, but it has had that unintended consequence.3
Creative Commons is a non-profit organization that has created several CC

licenses. These licenses allow creators to easily communicate which rights they
reserve, and which rights they waive for the benefit of users and other creators. A
number of the Creative Commons licenses make it easy to use their product, but
only in non-commercial enterprises. This anti-capitalist thinking goes against the
freedom to use an idea for any purpose and further takes money away from art.
Patents & Copyright
117
Digital Rights Management

(DRM)
Before the digital world, the enforcement of copyright law was
limited by the physical restrictions of the real world. Book authors
didn't worry much about people making unauthorized copies of their
work because the expense of making an unauthorized copy
approached the cost of purchasing a legal copy.
Enforcing arbitrary copyright provisions was not possible. You
could not charge someone more who planned to lend out a book
than someone who planned on keeping it for themselves, because
the cost of enforcement would be greater than the profits.
In a digital world, however, anything is possible when it comes to
creating rules for things. Stanford University law professor
Lawrence Lessig has written about what terms like copyright and
trespass, mean in a digital world in an interesting work entitled
Code 2.0. I will focus on just one aspect, digital enforcement of
copyright, or Digital Rights Management (DRM). In fact, with DRM,
arbitrary copyright provisions can now be created and enforced.
An unintended consequence, however, is that whatever rules are
created, whether onerous or not, you are provided no recourse
because the rules are encoded in software. A paper book cannot prevent me from accessing its contents, but a digital one can. Here is a
quote from a user of the Microsoft Zune music service:
The vast majority of the music I had purchased last year is completely gone. There's no refund, the music doesn't exist on the
service anymore, the files are just garbage now. Here's the
error: This item is no longer available at Zune Marketplace.
Because of this, you can no longer play it or sync it with your
Zune.
DRM has so far had mixed results. The first major use of DRM is
that which exists in DVD players. Most people don't realize that DVD
players have DRM because DVDs can be very easily exchanged,
unlike an iTunes song, whose hassles are well-documented.
118
Patents & Copyright
In fact the only time you can notice the DRM is when dealing with
region encoding:
DVD region encoding map
While this picture has a harmless, Benetton-esque, multicultural

feel to it, it also creates a hassle for anybody who crosses those
boundaries, and arguably decreases revenues. During layovers in
European airports, I have often browsed selections of books, music
and clothes, but I didn't purchase the European DVDs because they
will not play on my North American player. I have a friend who lives
in Latvia, and his choices in entertainment for his children are a tiny
fraction of what is available in the U. S. because of this mechanism.
The DVD region-encoding reinforces his country's third-world
status.
In general, the DRM in DVD players is not onerous because it
actually doesn't protect anything. The contents of the DVD are protected by DRM via encryption but the key to its contents is also
on the DVD!4 Typing in a password to play a DVD player would not
go over very well, so the password is put on the DVD. Once you can
read any DVD, you can read them all. It is impossible to lock an
object on its own.
Furthermore, the use of a DVD involves decoding it; the decoding
mechanism by itself cannot distinguish whether the user intends to
just watch the movie or post it on the Internet, so creating security
to prevent people from copying DVDs is impossible.
Moreover, if a DVD is illegally posted on the Internet, then the
security mechanism on all other copies of that DVD is now pointless
The DVD key is 40-bits which is very small, and was made small because of the
US export restrictions which existed in 1996. It has been said that computers
powerful enough to play a DVD are powerful enough to crack the password!
Patents & Copyright
119
because the information has already leaked out. In fact, the unencrypted version of the data is more useful because it doesn't have
the restrictions on it.
The key to security in any system is to have multiple layers of protection, a concept known as defense in depth. Inside a locked house,
you might have a safe to store important documents and jewelry. An
alarm system and cameras add additional levels of protection.
If you want to truly secure the contents of a DVD, you should put
it in a safe, encrypt the contents with a key stored away from the
DVD, secure access to the machine playing the DVD, make sure no
third-party applications are running on the computer playing the
DVD, etc.
One can obviously see that these mechanisms don't make sense
for the content we own. While I understand the interest in protecting copyrighted materials, we should not add complexity that serves
no purpose.
Because DVD decoding is protected by patents with per-copy
licensing fees, Microsoft decided not to include a means of playing
DVDs in Windows XP. (Microsoft likes to charge per-copy license
fees for its software but never likes to sign such license agreements
itself! DVD playback is one of the applications that hardware vendors must include with Windows to make it fully-functional.)
In fact, the DRM mechanisms create obstacles for the proprietary
software vendors, who try to legally jump through these non-existent security hoops, even more than for the free software guys, who
have just written code to do the job and been prosecuted.
If you sat down to write code to play your DVDs on your own
computer (not an easy task, I admit) you would be breaking the law!
Jon Lech Johansen, one of three people who first wrote free code to
play DVDs, suffered years of legal harassment.
The Digital Millennium Copyright Act (DMCA) of 1996 says that
writing or distributing DVD decoding software is illegal because it
circumvents a technological measure that effectively controls
access to a work. It is the DVD industry that has deigned itself
exclusive provider of technology to play your DVDs. Even if you
wrote this software just to play your own DVDs, you are breaking
the law, and Universal v. Reimerdes, the first lawsuit testing the
DMCA, has upheld this. According to Judge Lewis Kaplan:
In the final analysis, the dispute between these parties is simply
put if not necessarily simply resolved. Plaintiffs have invested
huge sums over the years in producing motion pictures in
reliance upon a legal framework that, through the law of
120
Patents & Copyright

copyright, has ensured that they will have the exclusive right to
copy and distribute those motion pictures for economic gain.
They contend that the advent of new technology should not
alter this long established structure. Defendants, on the other
hand, are adherents of a movement that believes that
information should be available without charge to anyone
clever enough to break into the computer systems or data
storage media in which it is located. Less radically, they have
raised a legitimate concern about the possible impact on
traditional fair use of access control measures in the digital era.
Each side is entitled to its views. In our society, however,
clashes of competing interests like this are resolved by
Congress. For now, at least, Congress has resolved this clash in
the DMCA and in plaintiffs' favor. Given the peculiar
characteristics of computer programs for circumventing
encryption and other access control measures, the DMCA as
applied to posting and linking here does not contravene the
First Amendment.
It appears that the defendants never argued the point that it

should be possible to write your own software to play your own
DVDs.
The Digital Media Consumer Rights Act, a bill proposed but not
passed by Congress, redefines such software to the status of barely
legal. You would be allowed to run unlicensed code to play DVDs
as long as you are doing it for legal purposes. This makes good
sense: playing a DVD isn't a problem, posting that movie on the
Internet is the problem!
DRM scenarios are also very limited in their scope. Your iPod's
DRM is incompatible with your cable box's DRM. As such, you will
never be able to watch your TV shows on your iPod because the
industry is unable to arrive at a DRM scheme on which everyone can
agree. (In fact, you can't even share videos between cable boxes in
your home today!)
There could certainly be uses of DRM that make sense. For example, you could rent a movie which would expire after a period of
time, using DRM as the enforcement of the expiration. The debate
should focus on rent versus own scenarios as wholly separate
classes of transactions.
Another compromise is to create a mechanism where media is
stamped with the textual string naming the owner, the digital equivalent of writing one's name on the inside cover of a book. This can
be done without encrypting the data.
In general, however, the focus shouldn't be on ways to restrict
information, but on ways to make it available. It seems like the
Patents & Copyright
121
recording industry spends more time coming up with new ways to

inconvenience customers than in coming up with new ways to
increase the quality of the user experience of digital media. When
you figure a global middle class of billions of people, worrying about
the pirates who live on the edge of society, or on the edge of subsistence, is a mistake.
People will steal if they want to steal. One of the reasons people
stole video games in the past was because it was not easy to try out
a game before you bought it. Now, with the Internet, you can download free demo games and make an informed decision what games
you want to buy in the future. Does providing free demos help or
hurt sales? I don't know the answer to this question, but it does
make for a better informed and discriminating customer; something
from which we all benefit in the long run.5
Music versus Drivers

Compared to the sociological need to protect music and movies,
hardware manufacturers in the computing industry are on weaker
footing when it comes to protecting the intellectual property of
the software drivers for their hardware. Drivers are simply the minimal glue that allows their hardware to work with the rest of a system.
Copyright laws were created to stimulate progress, but proprietary drivers serve as an impediment. Rather, they make people
hate their computer, dreading the day of work necessary to install a
fresh operating system. When you go to the store and purchase a
$500 video card, you would like that hardware to work with your
computer right out of the box. Given the time to design, manufacture and distribute hardware, software drivers should be available
first. There is no reason to have to download any drivers when setting up a new computer using Linux. (Linux often has drivers available before the hardware is released.)
While most drivers for Linux today are free, the graphics card
vendors are the most notable holdouts. They look at all the other
smaller drivers and argue that their complex software has more
worth. Of course, their current worth is a mixed bag because proprietary drivers not part of the free software tree are, by definition,
5
I downloaded a World War II flying simulator for my Xbox 360, but I couldn't
configure the controls like the controls on an RC model airplane. Don't the
programmers of that game recognize the crossover market? Relearning muscle
memory is hard and pointless.
122
Patents & Copyright
out-of-date and incompatible. Linux is working hard to make a 3-D

desktop, but it didn't run on an old computer because the proprietary driver didn't support the necessary features.6
Good driver support could be a competitive advantage for a hardware company, and Intel is aggressively moving into the 3-D video
card space with free drivers. However, it is best for the industry to
have the hardware guys fight it out on hardware features and cost,
not on software licensing agreements. (An important reminder here
is that much of the funding for the Linux kernel comes from hardware companies. Linux and the rest of the free software movement
is not being built only by volunteers.)
When I do enable their nascent 3-D features, the computer doesn't suspend and
resume properly anymore. And the cursor would sometimes become malformed.
Bugs like this would not survive long if the code were free. In fact, many teams
are re-engineering proprietary drivers from scratch just to be able to fix bugs.
Tools
123
TOOLS
You can tell a craftsman by his tools.
Socrates
The major cause of the software crisis is that the machines
have become several orders of magnitude more powerful! To
put it quite bluntly: as long as there were no machines, programming was no problem at all; when we had a few weak computers, programming became a mild problem, and now we have
gigantic computers, programming has become an equally
gigantic problem.
Edsger Dijkstra, 1972
Today's hardware is much more advanced than the software.
hile cooperation amongst our computer scientists is

important, the tools we use are as well, as they can either
facilitate progress, or impede it. The purpose of this chapter is to explore the biggest technical reason why people distrust
computer technology today. Many of the problems that frustrate
users, such as crashes, security violations, code bloat, and the slow
pace of progress, are directly related to the software used to create
our software.
124
Tools
In fact, the last hardware bug anyone remembers was the Intel
floating point division bug in 1994 nearly all the rest are software
bugs.1
The problem is this: the vast majority of today's code is written in
C, a programming language created in the early 1970s, or C++, created in the late 1970s. Computers always execute machine language, but programming these 1s and 0s are inconvenient in the
extreme, so programmers create high-level languages and compilers
that convert meaningful statements to machine code. We are asking
the same tools used to program the distant computer ancestors to
also program our iPods, laptops, and supercomputers, none of which
were even conceived of back then.
Imagine building a modern car with the tools of Henry Ford. A
tool ultimately defines your ability to approach a problem. While the
importance of cooperation in solving big problems is a major theme
of this book, the current set of tools are stifling progress. A complete unified set of libraries will change computing even more than
the worldwide adoption of the Linux kernel.
Metcalfe's law has unfortunately applied to C and C++; people
use them because others use them. These languages became so
dominant that we never moved on from our 1970s heritage. Programmers know many of the arguments for why newer languages
are better, but paradoxically, they just don't think it makes sense for
their codebase right now. They are so busy that they think they
don't have time to invest in things that will make their work more
productive!
Corporations today are struggling with the high costs of IT, and
many feel pressure to outsource to places like India to reduce the
costs of maintaining their applications. However, outsourcing doesn't decrease the manpower required to add a feature, it only
reduces the cost, so this is merely a band-aid fix.
In fairness to Intel, that bug happened on average once in every nine million division operations, and would return results that were off by less than 0.000061.
Testing a 64-bit processor's math capabilities involves 2128, or 1038 test cases! Intel
does continuously release errata lists about their processors. For example, initializing a computer is a very complicated process with plenty of room for ambiguity,
yet because user data isn't even in memory yet, there is no risk of a hardware bug
causing a newsworthy problem.
Tools
125
Brief History of Programming

Q. Where did the names C and C++ come from?
A. They were grades.
Jerry Leichter
If C gives you enough rope to hang yourself, then C++ gives
you enough rope to bind and gag your neighborhood, rig the
sails on a small ship, and still have enough rope to hang yourself from the yardarm.
Anonymous
For the first 60 years computers were programed, and input and
output was done using punch cards.2 Punch cards were invented for
the 1890 census by Herman Hollerith, founder of the firm which was
eventually to become IBM. Punch card machines operate on the
same principles as modern computers but have one millionth the
memory. Code reuse was impossible because one couldn't copy and
paste punch cards.
IBM's first computer, used for the 1890 census. Punch cards had been
around for 110 years before the world heard of pregnant chads.
This history ignores Fortran, Cobol, Algol, and other important languages, but this
book is PC-focused.
126
Tools
The first programing languages which ran on devices we would

recognize today as computers were known as assembly language,
and were first created in the 1950s. Programmers almost had to
comment explaining what each line of code did because it was
inscrutable otherwise:
.TITLE HELLO WORLD
.MCALL .TTYOUT,.EXIT
HELLO:: MOV
#MSG,R1 ;STARTING ADDRESS OF STRING
1$:
MOVB
(R1)+,R0 ;FETCH NEXT CHARACTER
BEQ
DONE
;IF ZERO, EXIT LOOP
.TTYOUT
;OTHERWISE PRINT IT
BR
1$
;REPEAT LOOP
DONE:
.EXIT
MSG:
.ASCIZ /HELLO, WORLD!/
.END
HELLO
Code to display Hello, World! in DEC's PDP assembly language.
I worked with an engineer in Office named Murray Sargent who

insisted that with practice, you could write in assembly language as
quickly as you could write in C. However, this is someone who wrote
a physics textbook, after creating a word processor to compose his
textbook in, so I never took that statement very seriously. (Working
at Microsoft was sometimes intimidating!)
As a college student, I took one assembly language class, which is
the perfect amount for a budding computer scientist. I learned
about the low-level concepts of computers: pointers, registers, the
stack versus the heap, etc.
In addition to being cryptic, the biggest problem with assembly
language instructions was that there were a lot of them, as each language contained details about the particular processor it was
designed to run on. As a result, in order to run your software on a
different computer, you were forced to alter nearly every line of
code.
Therefore, one of the biggest advancements in computing was
Bell Labs' creation of the programming language C in 1970. Compared to assembly language, code written in C is almost English:
#include <stdio.h>
int main(void)
{
printf("Hello, World!\n");
return 0;
}
Hello, World! in C. This is much easier to read than assembly language,

but still has weird things like #include.
Tools
127
C is more comprehensible, but more importantly, by simply swapping the compiler, it enables software written in C to run on every
computer in existence today with few changes required. In building
the standard replacement for assembly language, Bell Labs changed
computing.
Assembly language actually still exists in the nooks and crannies
of modern computers: deep in the code of the operating system kernel, inside game engines, and in code to program specialized graphics or math processors. However, most processor-specific knowledge
has moved out of the code a programmer writes and into the compiler, which is a huge step up.
While developers at Microsoft used assembly language in the very
early years, when I joined in 1993, nearly all code was written in C
or C++, and that is still true today.
Bill Gates
Microsoft, 1978
Bill Gates quit coding somewhere in this timeframe.
Paul Allen
Our most important codebases Office and OpenOffice, Internet

Explorer and Firefox, SQL Server and MySQL, IIS and Apache are
written in C or C++ today.
128
Tools
The world spent the first 1,000 man-years of software development primarily in assembly language, but most programs are written in C or C++. Man has spent 400,000 man-years in those two
languages, the same amount of time that the Egyptians spent building the Great Pyramid of Giza:
Man spent as much time programming in C and C++ as building the pyramids of Giza. Unfortunately, the software used to build our software looks
as old and cracked as those pyramids do today.
Today, the free software community and Microsoft are locked in a

closely fought battle: Firefox is comparable, but not dramatically
better than IE. This is because neither has yet adopted a productive
new programming language for such a large and important codebase. They are warring armies of patient cavemen.
In fact, what both C and C++ are missing is an innovation known
as garbage collection (GC), or automatic memory management, an
idea so subtle and clever that it would be worthy of the first Nobel
Prize in Computer Science.
Tools
129
Lisp and Garbage Collection

Please don't assume Lisp is only useful for Animation and
Graphics, AI, Bioinformatics, B2B and E-Commerce, Data Mining, EDA/Semiconductor applications, Expert Systems, Finance,
Intelligent Agents, Knowledge Management, Mechanical CAD,
Modeling and Simulation, Natural Language, Optimization,
Research, Risk Analysis, Scheduling, Telecom, and Web Authoring just because these are the only things they happened to list.
Kent Pitman, hacker
The future has already arrived; it's just not evenly distributed yet.
William Gibson, science-fiction writer
Memory, handed out in contiguous chunks called buffers or

arrays, is the scratch space for your processor. Your computer needs
to read bitmaps and e-mails from the disk or network into memory
in order to display or let you edit them. (Your operating system can
actually cheat and pretend to give out lots of memory, and write to
disk the infrequently used portions virtual memory. If you come
back to an application after not using it for several hours, you may
find your hard drive is busy for several seconds because it is loading
all of your data and code back into the real memory.)
Memory contains the current state of all the applications on your
computer, and is therefore an extremely precious resource. Garbage
collection is a system that transparently manages the computer's
memory and automatically reclaims the unused garbage from the
programmer when he is done using it. I will spend the rest of the
chapter explaining how this technology changes programming, so
sit tight if my explanation doesn't make sense yet.
130
Tools
John McCarthy created GC in 1959 when he created Lisp, a language invented ten years before C, but which never became
accepted into the mainstream:
(defun finder (obj vec start end)
(let ((range (- end start)))
(if (zerop range)
(if (eql obj (aref vec start))
obj
nil)
(let ((mid (+ start (round (/ range 2)))))
(let ((obj2 (aref vec mid)))
(if (< obj obj2)
(finder obj vec start (- mid 1))
(if (> obj obj2)
(finder obj vec (+ mid 1) end)
obj)))))))
The binary search function written in Lisp is a simple algorithm for quickly
finding values in a sorted list. It runs in Log (n) time because at each step, it
divides the size of the array in half, similar to how we look up words in a
dictionary. There are many faster and more complicated algorithms, and
search is a very interesting and large topic in computer science, but 99% of
the time, ye olde binary search is good enough.
2
C and C++ was based on BCPL and other languages before it,
none of which had garbage collection. Lisp is a language built by
mathematicians rather than operating systems hackers. Lisp pioneered GC, but also was clean and powerful, and had a number of
innovations that even C# and Java don't have today.3
Wikipedia's web page doesn't explain why Lisp never became
accepted for mainstream applications, but perhaps the biggest
answer is performance.4 So instead, people looked at other, more
primitive, but compiled languages. The most expensive mistake in
the history of computing is that the industry adopted the non-GC
language C, rather than Lisp.
3
Perhaps the next most important innovation of Lisp over C is functional programming. Functional programming is a focus on writing code which has no side
effects; the behavior of a function depends only on its parameters, and the only
result of a function is its return value. Nowadays, in object-oriented programming,
people tend to create classes with lots of mutable states, so that the behavior of a
function depends on so many things that it is very hard to write correct code,
prove it is correct, support multiple processors manipulating that object at the
same time, etc. Functional programming is a philosophy, and Lisp made it natural.
Most Lisp implementations ran 10x slower than C because it was interpreted
rather than compiled to machine code. It is possible to compile Lisp, but unfortunately, almost no one bothered. If someone complained about Lisp performance,
the standard answer was that they were considering putting the interpreter into
hardware, i.e. a Lisp computer. This never happened because it would have
sucked Lisp into the expensive silicon race.
Tools
131
While Lisp had many innovations, its most important was garbage
collection. Garbage collection requires significant infrastructure on
the part of the system and is a threshold test for an intelligent programming language.5
Because so few of the most important codebases today have
adopted GC, I must explain how it improves software so my geek
brethren start using it.
The six factors of software quality are: reliability, portability, efficiency, maintainability, functionality, and usability; I will discuss how
GC affects all of these factors. The most important factor is reliability, the sine qua non of software.
Some argue that static type checking (declaring all types in the code so the compiler can flag mismatch errors) is an alternative way of making software more
reliable, but while it can catch certain classes of bugs, it doesn't prevent memory
leaks or buffer overruns.
Likewise, there are smart pointers which can emulate some of the features of
garbage collection, but it is not a standard part of languages and doesn't provide
many of the benefits.
Apple's Objective C 2.0 has added support for GC in their language, but it is
optional, and therefore doesn't provide many of the benefits of a fully GC language, like enabling reflection or preventing buffer overruns.
132
Tools
Reliability
Therefore everyone who hears these words of mine and puts
them into practice is like a wise man who built his house on the
rock. The rain came down, the streams rose, and the winds
blew and beat against that house; yet it did not fall, because it
had its foundation on the rock. But everyone who hears these
words of mine and does not put them into practice is like a foolish man who built his house on sand. The rain came down, the
streams rose, and the winds blew and beat against that house,
and it fell with a great crash.
Matthew 7:24-27
If the software of Star Wars was no more reliable than today's...
Reliability, according to Wikipedia, is the ability of a device or

system to perform a required function under stated conditions for a
specified period of time. In software, the specified period of time is
forever, because unlike metal and other things containing atoms, it
doesn't wear out.
For example, reliability means that the computer does the same
things for the same inputs every time. If you create a function to add
two numbers, it should always work for all numbers. If your computer gets hacked and doesn't respond to your input, it isn't reliable.
Even performance is related to reliability: if your computer is wasting CPU cycles and is no longer responsive, it isn't reliable.
Tools
133
Today, few would say that computers are completely reliable.

Maybe your cable box crashes: Comcast: On Demand, Sometimes.
Maybe your laptop doesn't recognize the wireless network in one
particular Internet cafe. I have trained my family to know that whenever their computer is misbehaving, they should reboot the system
before calling me. This fixes any errors where the memory is in an
unspecified state, which is the most common computer problem.
Reliability is important because software is built up as a series of
layers. At one level is the code displaying a file open dialog box, and
far below is a device driver reading the file you selected as a series
of scattered data blocks on the hard drive. Every layer of software
needs to be reliable because it is either being used by other software or by a human. Both humans and software need a reliable
foundation.
134
Tools
A tiny bug in its software caused the crash of one of the European
Space Agency's Ariane 5 rockets, costing $370 million:6
The rocket's software was written in Ada, an old language, but with many of the
features of garbage collection. Code which converted a 64-bit integer to a 16-bit
integer received a number too big to fit into 16 bits, and so the conversion code
threw an exception. The code to handle this exception was disabled, and therefore
the computer crashed. When this computer crashed, it started sending confusing
diagnostic information to the flight control computer, causing it to fly in a crazy
way and break apart, triggering the internal self-destruct mechanisms.
Many blame management, but this was a simple design bug (you should be very
careful when throwing away data). This was compounded because they were
using a specialized embedded system with a non-mainstream programming language which allowed them the capability of disabling certain exceptions. This bug
could have been caught in testing, but they didn't use accurate trajectory information in the simulations. Perhaps clumsy tools made it hard to modify test cases,
and so they never got updated.
Tools
135
When your computer crashes, you can reboot it; when your rocket
crashes, there is nothing to reboot. The Mars Spirit and Opportunity
rovers had a file system bug which made the rovers unresponsive,
nearly ending the project before they even landed!7
While it isn't usually the case that a software bug will cause a
rocket to crash, it is typically the case that all of the software layers
depending on that buggy code will also fail. Software reliability is
even trickier than that because an error in one place can lead to failures far away this is known in engineering as cascading failures. If an application gets confused and writes invalid data to the
disk, other code which reads that info on startup will crash because
it wasn't expecting invalid data. Now, your application is crashing on
startup. In software, perhaps more than in any other type of intellectual property, a bug anywhere can cause problems everywhere,
which is why reliability is the ultimate challenge for the software
engineer.
Perfect reliability is extremely hard to achieve because software
has to deal with the complexities of the real world. Ultimately, a key
to reliable software is not to let complexity get out of hand. Languages cannot remove the complexity of the world we choose to
model inside a computer. However, they can remove many classes of
reliability issues. I'm going to talk about two of the most common
and expensive reliability challenges of computers: memory leaks and
buffer overruns, and how garbage collection prevents these from
happening.
Memory Leaks
Web banner ad for a tool to find memory leaks. There is a cottage industry
of tools to fix problems which exists only because the programming language is broken in the first place.
One of the most common forms of memory corruption code is

memory leaks a common source of computer unreliability and
frustration for users that worsens as our code becomes more complicated.8
7
8
The rover file system used RAM for every file. The rovers created a lot of system
logs on its trip from Earth to Mars, and so ran out of memory just as they arrived!
The problem is even worse because every big C/C++ application has its own
memory allocators. They grab the memory from the OS in large chunks and manage it themselves. Now if you are done with memory, you need to return to the
person who gave it to you.
136
Tools
Losing the address of your memory is like the sign outside a Chinese dry-cleaner: No tickie, no laundry. To prevent leaks, memory
should be kept track of carefully. Unfortunately, C and C++ do not
provide this basic feature, as you can allocate and lose track of
memory in two lines of code:
byte[] p = new byte[100]; // p points to 100 bytes of memory
p = NULL;
// p now points to NULL, reference
// to 100 bytes lost
New returns the location of the newly allocated memory, stored into variable p. If you overwrite that variable, the address of your memory is lost,
and you can't free it.
Writing code in C or C++, in which all unused memory is

returned to the operating system, is both complicated and tiresome:
Person* p = new Person("Linus", "Torvalds", 1969);
if (p == NULL)
//Out of Memory or other error
return;
Person* p2 = new Person("Richard", "Stallman", 1953);
if (p2 == NULL) //Out of Memory or other error
{
delete (p);
//Cleanup p because the p2 failed
return;
}
MarriageLicense* pml = new MarriageLicense(p, p2)
if (pml == NULL) //Out of Memory or other error
{
delete (p);
//Cleanup p and p2
delete (p2);
return;
}
Code in C or C++ to manually handle out-of-memory conditions and other

errors is itself bug-prone, adds complexity, and slows performance.
As programmers write increasingly complicated code, the work

required to clean it up when things go wrong becomes more difficult. This small example has only three failure cases and the code to
remedy these error conditions makes the code twice as complex as
it would otherwise be. Furthermore, code which only executes in
unlikely failure scenarios doesn't get executed very frequently, and
is therefore likely to have bugs in it. A rule in software is: If the
code isn't executed, it probably doesn't work. And if your error handling code doesn't work, your problems will accumulate faster.
Scale this dilemma into millions of lines of interdependent code
written by different people and the complexities compound beyond
Tools
137
our ability to fix them. To date, there is no non-trivial codebase written in C or C++ which is able to solve all of these error conditions,
and every codebase I saw at Microsoft had bugs which occurred
when the computer ran out of memory.9
MySQL, an otherwise highly reliable database, which powers popular websites of Yahoo! and the Associated Press, still has several
memory leaks (and buffer overruns.) Firefox's bug list contains several hundred, though most are obscure now.10
Let's take a look at why a memory leak can't happen when running on a GC language:
byte[] p = new byte[100]; // Variable "p" points to 100 bytes.
p = NULL;
// p now points to NULL.
// The system can deduce that no variables
// are referencing the memory, and therefore
// free it.
You don't have to call delete because the system can infer what memory is
in use.
After the second line of code executes, p no longer references

the 100 bytes. However, a GC system is smart, and it can take inventory of what memory is in use, and therefore discover that because
these 100 bytes are not being referenced, they can be freed. Likewise, if the Chinese laundry knew you had lost your ticket, they
would know to throw away your clothes.
It is this simple innovation that changes programming. Memory is
such a critical part of computers that we need to have the system,
not the programmer, keep track of it.11
9
Often near the end of a development cycle, after fixing our feature bugs, we
would focus on some of the out-of-memory bugs. While we never fixed them all,
we'd make it better and feel good about it.
It is true that when you run out of memory, it is hard to do anything for the user,
but not causing a crash or a further memory leak is the goal.
10 Here is a link to all active MySQL bugs containing leak:
http://tinyurl.com/2v95vu. Here is a link to all active Firefox bugs containing
memory leak: http://tinyurl.com/2tt5fw.
11 GC makes it easy for programmers to freely pass around objects that more than
one piece of code is using at the same time, and the memory will be cleaned up
only when every piece of code is finished with it. C and C++ do not enable this
and many other common scenarios.
To write code which allows two pieces of software to share memory and to return
it to the operating system only when both are finished is complicated and onerous. The simplest way to implement this feature in C/C++ is to do reference
counting: have a count of the number of users of a piece of memory. COM, and the
Linux and Windows kernels have reference counting. When the last user is finished, the count turns to zero and that last user is responsible for returning the
memory to the OS. Unfortunately, this feature requires complicated nonstandard
code (to handle multiple processors) and places additional burdens on the programmer because he now needs to keep track of what things are referencing each
138
Tools
It is also quite interesting that GC enables a bunch of infrastructure that Microsoft's OLE/COM component system tried to enable,
but COM did it in a very complicated way because it built on top of
C and C++, rather that adding the features directly into the language:
COM Feature Name
.Net & Java Feature Name
Reference Counting
Garbage Collection
BSTR
Unicode strings
Type Libraries
metadata + bytecodes / IL
IUnknown
Everything is an Object
IDispatch
Reflection
DCOM
Remoting and Web Services
COM contains a lot of the same infrastructure that GC systems have, which
suggests a deep similarity of some kind. Doing these features outside the
language, however, made writing COM more tedious, difficult, and errorprone. .Net completely supersedes COM, and in much simpler packaging,
so you will not hear Microsoft talk about COM again, but it will live on for
many years in nearly every Microsoft codebase, and many external ones.
Buffer Overruns
As bad as memory leaks are because they often cause crashes,
buffer overruns are worse because your machine can get hijacked!
Buffer overruns are the most common type of security bug, and a
major nemesis of the computer industry. Microsoft's code has fallen
prey to a number of buffer overruns; the Code Red virus, which
infected Microsoft's web server and caused rolling outages on the
Internet, is estimated to have cost the industry two billion dollars.
Free software is certainly not immune to this either; on a daily basis
my Ubuntu operating system downloads fixes to newly discovered
buffer overruns.12
Like with memory leaks, you can create and overrun a buffer with
just two lines of code:
int* p = new int[50]; //Allocate 50 entries, referenced 0-49
p[50] = 7;
//Write to 51st entry, off by 1 bug
C and C++ do not validate memory access, so a programmer can intentionally or unintentionally read or write to memory he shouldn't have access to.
other. Handing out references, which should be simple, is now error-prone. Even
so, reference counting is insufficient: if two objects point to each other, but no one
points to them, they will keep each other alive.
12 One of the security fixes awaiting me right now is: SECURITY UPDATE: arbitrary
code execution via heap overflow, from CVE-2007-3106.
Tools
139
If hackers find a piece of code which doesn't validate a buffer

length, they can send to that computer, not the usual data of an email or picture, but a carefully crafted block of evil code, causing
the system to start executing the hacker's software. A buffer overrun exploit is like a virus in the biological world: imagine cells which
would unquestioningly incorporate any DNA they happened upon.
In addition to knowing what memory is in use, a GC system knows
the size of every buffer, and so can validate all reads and writes to
memory. The computer, not the error-prone programmer, makes
sure memory doesn't ever leak or become corrupt. Garbage collection is a necessary, if insufficient, precondition of the better code
that we seek.
While GC is necessary for reliability, it provides many other
advantages.13
13 The Linux kernel is an example of reliable C code. Many might use this as proof
that it is possible to write reliable code without garbage collection. However,
there are several reasons why this lesson may not be valid:
The kernel's primary job is to provide glue code to make the hardware work. Seventy-five percent of the kernel's code is hardware specific, and much of the code
is tiny, simple components. Much of the remaining code implements decades-old
operating system concepts like threads. This makes the feature requirements and
design relatively stable over time. All of the things we consider smart software,
like grammar checkers and speech and handwriting recognition, involve writing
code which has no clear design and would never be part of a kernel.
The kernel doesn't have the need to inter-operate with as much other software
like applications do. Software written by many disparate and interdependent
teams makes GC more important.
The Linux kernel has a big team per line of code compared to other codebases.
This gives them the luxury of using inefficient tools.
The Linux kernel does have a few memory leaks even as I write this. (Bugs 5029,
6335, 8332, 8580)
The kernel has a number of specialized requirements which make it harder to do
garbage collection, but it would benefit from it, and it wouldn't surprise me if
most of the kernel's code which is not written in an assembly language is written
in a GC language one day.
For now, we should focus on porting application code and leave the Linux kernel
as the very last piece of code written in C. Once the world has collaborated on a
new software stack, written in our new programming language, we can knock on
Linus's door with our now mature tools and figure out what his requirements are.
140
Tools
Portability
Fifteen years ago, Intel, with HP's help, decided to create a new
processor architecture which would enable them to incorporate all
they had learned in creating their x86 processors, first introduced in
the 1970s. This chip, known as Itanium or IA-64, is a 64-bit chip
which removed all of the previous compounded ugliness, leaving a
simpler and therefore potentially faster design in its place. Microsoft was already working on porting its C tools and code to IA-64
when I joined in 1993, though the chip wasn't released until 2001:
Beautiful new chip, but even though Intel included x86 compatibility hardware, it was still incompatible with existing C and C++.14 Code written in a
GC language is already portable to chips that have yet to be created.
In spite of the years of R&D and Intel's enormous resources, the

chip is hardly used today, even on servers. The reason this chip hasn't yet taken off is literally a billion-dollar question, causing Sun
founder Scott McNealy to dub it the Itanic.15
14 Building emulation in hardware was not a good idea because you cannot mix 32bit and 64-bit code because you cannot hand a 64-bit pointer to 32-bit code. You
would have to boot the computer into a 32-bit mode where it never created a 64bit memory address in order to let all code to run, and such a feature would
defeat the purpose of building a 64-bit OS.
In 2006, Intel removed their x86 hardware interpreter, which ran the code 10
times slower than native, and replaced it with software that would dynamically recompile the x86 binary to Itanium instructions. The code still couldn't interoperate with true 64-bit code, but it could run at full speed, and it lowered the cost of
building an Itanium. It is ironic that Intel had put into its hardware precisely what
it was trying to get rid of!
15 The two most common theories for the failure of IA-64 are: performance benchmarks which weren't significantly better than Intel's current processors, and
Tools
141
The biggest obstacle Intel faced was the fact that our pyramid of
C and C++ code running on PCs today is compiled for the x86 processor. Such programs won't run without at least re-compiling the
source for another processor, and it might even require changes to
the software because it is easy to write non-portable C/C++. Consequently, the adoption of new hardware is significantly limited by the
fact that you have to find every piece of code out there, and recompile it. What this ends up meaning is that while a lot of your stuff
works, some of it doesn't. Itanium Linux had no Flash player or
Adobe Reader till very recently, two significant stumbling blocks for
desktop deployments, and even one obstacle can be too many.16
GC solves portability issues because programs written in languages such as Java, C#, Python, etc. are no longer compiled for
any specific processor. By comparison a C/C++ executable program
is just a blob of processor-specific code containing no information
about what functions and other metadata are inside it. Its contents
are completely opaque to the system, and the processor just starts
blindly executing it. To a GC system, a blob of binary code is insufficient.
Like stores in the real world, GC systems in principle need to
close to take inventory. However, unlike stores, they cannot just kick
all the existing customers out, or wait for them to leave. So it does
the equivalent of locking the doors, not letting any customers in or
out (halts execution of code), and tabulating what is on the shelves
and in the shopping carts (what memory is in use). Once it has an
accurate account and it can then re-open the doors and let existing
customers leave, and new ones enter (program execution
resumes.)17
incremental mechanisms to add 64-bit computing to x86 without starting from

scratch and creating something 100% incompatible.
16 If that code were free software, it would have been ported already by a geek just
trying to get his hardware to work. However, if the code were written in a modern
GC programming language, the issue of porting our pyramid of code wouldn't
exist.
17 A more precise analogy is that it doesn't let any customers put new things into
their cart until the inventory is done. It also needs to tag shopping carts as having
been scanned so that it doesn't take inventory of a cart twice.
142
Tools
When GC pauses code, it needs to know what function the processor is currently executing, and even where in that function it is:
void ExampleFunction()
{
int x = SquareNum(3); //If execution stops here, no memory
//allocated yet.
object o = new object(); //This allocates memory into 'o'.
DoStuffWithObject(o); //If execution stops here, 'o' is in use.
int y = SquareNum(4); //If execution stops here, 'o' is no
//longer in use, and can be cleaned up.
}
GC programs in principle need to know what your processor is doing at

each moment in time to precisely inventory memory.
The garbage collector needs to be able to know what objects are

live in the system at every moment in time, but this depends on
exactly what line of code the processor is executing when the inventory process is taking place. A C or C++ executable isn't required to
have this sort of information available, but GC requires a rich understanding of the code in order to do its job.18 Therefore, when a program is distributed for a GC language, it is delivered either in
source form, or converted to a bytecode, a more efficient format to
parse, but with very little loss of information compared to the original source.19
.assembly helloworld {}
.method public static void MyMain() cil managed
{
.entrypoint
ldstr "Hello, World!"
call void [mscorlib]System.Console::WriteLine(string)
ret
}
Hello, World! in .Net's bytecode. This is similar to the original C#, though
more verbose.
In (all of the common) GC systems, the programmer ships source

code or bytecode, not a machine-specific binary blob. If all code
18 The Boehm GC doesn't require metadata, so it has to guess what is a pointer by

walking the stack and heap and looking for bit patterns that look like pointers!
Therefore, because Boehm isn't true GC, it doesn't enable things like reflection.
Also, Boehm can't compact memory.
19 That GC basically requires you to ship your source code seems like a hint from the
Author of computer science that free software is a good thing.
Tools
143
written for the Macintosh was written in a GC programming language, it would have been zero work for Apple and third-parties to
switch to the Intel processor once the GC runtime was ported!20
In fact, an application written in a GC programming language is
automatically portable to chips that haven't even been created yet.
We impede our freedom to create new processors when software is
not written in portable languages.
Portability is one of the holy grails of computing, and while GC
code doesn't completely solve cross-operating system portability, it
does solve the situation of running the same code on different processors itself an enormous step.21
With the source code or bytecode, the GC system has all the information it needs to figure out exactly what is going on when it stops
execution. In fact, it also has a lot of information that enables other
cool features like reflection, which allows code to query information
about an object at runtime. These features create a more dynamic
system.
We've discussed two advantages of GC: greater reliability and
portability. The next topic is code performance, which is the biggest
worry when using modern tools. I have had many discussions with
smart geeks who insisted that languages such as C# simply weren't
suitable for their fast code.
Efficiency
It doesn't matter how fast your code runs if it doesn't give the correct result, but processing power is still an important resource. In
fact, code efficiency is even less important than memory usage
because if you waste memory, your computer will eventually crash,
but if you waste processor cycles, your computer will just be sluggish and annoying. Microsoft has always focused on performance,
and has often promoted it as a competitive advantage.22
If you walk up to a programmer on the street and ask them what
they think of Java, one of your answers will be: slow. At one time,
20 Today, Mac users have to worry about whether a program is a PowerPC binary or
an Intel binary. There is even a rumor that one day there will be four binaries, two
for the 32-bit and 64-bit versions of both processors!
21 If we all use the same OS, then this OS cross-platform problem disappears :-)
22 GC code can give better performance because it has the exact hardware in front
of it. C compilers are forced to generate generic code which will run on all models
of a processor. New processors aren't just faster than previous versions, they add
new instructions and other capabilities, which often go unused.
144
Tools
Netscape started working on a web browser written in Java, but

they abandoned the effort after one year because of performance
worries, and before the Java runtimes got fast.
Languages with garbage collection manage more details for you
compared to unmanaged code, and this by definition adds a performance cost to your application. The need to pause a program
while taking inventory of memory is one of the most common reasons cited for not using automatic memory management.23 Add to
this a bias: programmers, from their first class in computer science,
are taught that analysis of algorithms revolves primarily around an
analysis of code performance.
Both pieces of code below are less than 20 lines and look fairly
similar, but the one on the left is significantly faster:
static void quicksort(int[] a, int l,
int r)
{
int i, j;
char x, y;
i = l; j = r;
x = a[(l+r)/2];
static void bubblesort(int[] items)

{
int i;
int j;
int temp;
int x = items.Length;
for( i = (x - 1); i >= 0; i-- )
{
for( j = 1; j <= i; j++ )
{
if( items[j-1] > a[j] )
{
temp = items[j-1];
items [j-1] = items[j];
items [j] = temp;
}
}
}
do {
while((a[i] < x) && (i < r)) i++;
while((x < q[j]) && (j > l)) j--;
if(i <= j) {
y = a[i];
a[i] = a[j];
a[j] = y;
i++; j--;
}
} while(i <= j);
if(l < j) quicksort(a, l, j);

if(i < r) quicksort(a, i, r);
Algorithm analysis teaches you that the code on the left should be about
50,000 times faster than the code on the right at sorting one million numbers. The speed of code, not the speed of the language is what matters.24
23 You only need to in principle pause a program while doing GC. The good news is
that in many types of applications, from word processors to web servers, a brief
pause is acceptable. This number is proportional to the amount of memory in use
by the application, and therefore obeys the only pay for what you use rule of
engineering.
However, in certain situations, pausing is not acceptable, such as video games or
code interacting with hardware. There are many solutions such as implementing
GC via reference counting, or making it incremental and scheduling it proactively
during idle moments, etc.
24 To sort an array of n numbers, Quicksort will do it in a time proportional to
n * log2(n) and Bubblesort will do it in n2. If you plug in one million for n, it means
Tools
145
Performance is rightfully important because the difference

between fast code and slow code can mean the difference between
milliseconds and hours. However, I spent years analyzing performance bottlenecks in various codebases at Microsoft, and every performance problem was caused by inefficient algorithms or bad
design, not the speed of the language. Fast code written in C would
be fast in a GC programming language, and slow code written in a
GC programming language would also be slow in C.25 The first lesson in the Bible of Computer Science, Donald Knuth's compendium of software algorithms, is that the speed of code is determined
by the algorithms it incorporates.26
In the early days of C, you could ask an assembly language programmer what he thought of C, and he'd have said that it was too
slow. No one says that today, however, because of C's 30 years of
maturation, the dramatic progress of hardware, and because its significant other advantages overshadowed its early lackluster performance issues. My first group at Microsoft built FoxPro, which was
for many years the fastest PC database, and they quit writing code
in assembly language the day they re-wrote a bit of it in C and found
it faster!
Java and C# have some performance issues today, but the primary reason is that most of the community is still expending energy
continuing to optimize C/C++ tools, and not even beginning to
explore the many ways to speed up GC systems that are unavailable
to the older languages. Even so, let's assume that modern tools slow
down code by 20%.27 Taking a one-time 20% hit is worth it because
our current software is ineffectively, or wastefully, using today's processing power. My pedestrian laptop is a now-standard dual-core
that the fast code will be about 50,000 times faster than the slow code.
I wrote some code to compare Mono's Array.Sort() to bubble sort, and found that
the fast algorithm was 10,316 times faster than the slow one, which is the same
order of magnitude as 50,000.
25 Firefox will get faster in certain ways when it is re-written in a modern programming language. One of the performance boosts will come when it throws away its
Javascript interpreter. If Firefox ran Javascript at compiled native speed, it would
be a significant performance boost.
Another example: Google's web spiders hit nearly all of my web pages nearly
every day, even though I add new entries only on a monthly basis, and my articles
do not change after they have been posted. The best way to make their servers
use less processing power is not to do micro-optimizations, like making it 10%
faster at parsing HTML, but to make it smarter about the big things, like not
fetching and parsing unchanged web pages in the first place.
26 Knuth's Art of Computer Programming is about as hard to read as the King James
Bible because it uses a made-up, primitive programming language.
27 That many focus on this 20% means they are focusing on a mere constant and
ignoring the order of magnitude analysis which should be the primary focus.
146
Tools
computer, and it requires only one of the cores running at 2% usage

when I type text as fast as I can into this book; even though I am
working on a long document, it is redrawing the screen, autocorrecting, spell-checking, updating document statistics, etc. Sometimes CPU usage noticeably peaks in OpenOffice, but that's usually
because it is redrawing the screen over and over, or doing some
other useless task.
Today's software is primitive, i.e. it cannot do a processor-consuming task of intelligent analysis of my writing rather, the computer is mostly just sitting there. It knows to correct teh, but not
much more. The computer industry should focus less on not wasting
a few CPU cycles and more on dreaming up new ways of using them
to make our lives easier and better. Our stupid software still has a
long way to go.
Tools
147
Maintainability
The major incentive to productivity and efficiency are social
and moral rather than financial.
Peter Drucker
During the years we worked on Viaweb I read a lot of job
descriptions. A new competitor seemed to emerge out of the
woodwork every month or so. The first thing I would do, after
checking to see if they had a live online demo, was look at their
job listings. After a couple years of this I could tell which companies to worry about and which not to. The more of an IT flavor the job descriptions had, the less dangerous the company
was. The safest kind were the ones that wanted Oracle experience. You never had to worry about those. You were also safe if
they said they wanted C++ or Java developers. If they wanted
Perl or Python programmers, that would be a bit frightening
that's starting to sound like a company where the technical
side, at least, is run by real hackers.
We were always very secretive about our competitive advantage of Lisp. Robert Morris says that I needn't be because even
if our competitors had known, they wouldn't have understood
why: If they were that smart they'd already be programming in
Lisp.
Paul Graham, Hackers and Painters
I worked on a team where one of our developers spent weeks

integrating an HTTP stack written in C into our C codebase. In Java
or C#, this basic functionality is already there, but even if it weren't,
it would take mere hours to integrate such a library.28
For weeks after I started programming in C#, I would write code
with a grin on my face, and my fellow co-workers also learning C#
understood why. With a modern language, one finds that by having
the system manage many of the details, more time is spent thinking
about higher-level software issues.
Sun Microsystems researchers, in a paper called FreeTTS A
Performance Case Study, analyzed this relationship between the
productivity and performance of two speech synthesis engines
one written in Java, named FreeTTS, and one in C, named Flite:
When we started our study of the performance characteristics
of a speech synthesis engine programmed in the Java programming language, our expectations were that it would hopefully
be able to run nearly as fast as the native-C counterpart.
Through using some straightforward optimizations and relying
28 It took so much time because C libraries often use their own string classes, synchronization primitives, error handling schemes, etc. Furthermore, modern languages do a better job at many little details like backward compatibility. If you
reorder the elements of a data structure written in C, you need to recompile all
code which uses it.
148
Tools
on the aggressive optimizations performed by the Java HotSpot
compiler, we were pleased to find that FreeTTS runs two to
four times faster than its native-C counterpart, Flite.
Clearly, it would be possible for us to roll some of these optimizations back into Flite with the likely result of improving
Flite's performance to levels similar to FreeTTS. The lack of
Java platform features such as garbage collection and high-performance collection utilities, however, makes performing these
optimizations in Flite much more time consuming from a programming point of view.
The programmer's time is finite and is thus the limiting factor in

software engineering. One could say that the only software quality
that matters is maintainability, as that gives you the time to focus on
every other aspect. We need garbage collection for reliability, but
the payoff is increased maintainability, which will pay for the transition costs.
Many people in the software industry no longer believe that programming languages affect productivity. This misconception exists
primarily because C++ didn't end up being significantly more productive than C. The debate is mostly between C and C++. Quoth
Linus:
Quite frankly, even if the choice of C were to do nothing but
keep the C++ programmers out, that in itself would be a huge
reason to use C.
C++ leads to really really bad design choices. Developers
invariably start using the 'nice' library features of the language
like STL and Boost and other total and utter crap, that may
'help' you program, but they cause infinite amounts of pain
when they don't work and inefficient abstracted programming
models.
At the same time, few argue that assembly language had close to
the productivity of C, so this contradiction is unresolved in the
minds of many. Many computer geeks like to argue about C versus
C++, but compared to a modern and elegant language like C#, this
is like choosing between Britney and Paris.
While C++ added object orientation features to C, it had a fatal
flaw: it was a superset of C. In fact, the early compilers just converted their code to C, which was a great way to bootstrap the new
language to the many places where C was already used, but by beging a superset it tied the designers to the baggage of C. While C++
added new object-oriented features, it also added significant complexity.29 I used C++ for many years, and I liked some of the
29 Another way to analyze the maintainability of code is to analyze the maintainability of the compilers for that code. One finds that C++ compilers are big, ugly and
Tools
149
improvements over C, but the language is mind-numbingly complicated and generally provides many more ways to screw up than in C.
That C++ didn't start with a clean slate is the second biggest mistake in the history of computing.
If you could double developer productivity at the cost of half of
your current performance, would you take it? There isn't any universal agreement on the answer to this question among computer engineers today. However, Moore's law states that Intel's computers take
merely 18 months to become twice as fast, and in ten years, computers will be another 100 times faster than they are today. A 20% drop
in performance to enable garbage collection would take Intel hardware progress 4 months to counter-balance, and we would pay it
only once. Anders Hejlsberg, architect of C#, has said it is the best
use of Moore's law to come around in years.
Reliability issues also add a variability to the engineering cost of a
project because developers spend an unpredictable amount of time
fixing bugs. In adopting GC, developers would pay a fixed performance cost in exchange for decreased engineering costs and variability.
Functionality and Usability

Functionality and usability are soft features lower priority
amongst the other software quality factors. If your computer corrupts your files while saving them to disk, you don't care how rich or
usable the application was up to that point.
Many consumer devices today are often not very usable because
the people building those devices write 100% custom code, often in
assembly language. It takes a lot of effort merely to get the software
working, and once it is, the device manufacturers become fearful of
changing anything lest they break it.
One of my groups at Microsoft, called Spot, ran C# interpreted on
a $50 watch. The code was efficient because we focused on the
high-level performance issues that you never seem to get to when
you write in C or C++. In fact, the processor ran at just a few percent on average. Our code was so fast that it hardly ever ran, which
was great for battery life! The watch failed in the marketplace, but it
slow because the languages are big and complicated. If the tools for a language
are complicated, it is also likely the code written in them will have maintainability
issues. It is perfectly possible to build a language which looks nothing like C, but
whose compiler emits C. My favorite C++ gotcha is the need to make destructors
virtual. One of the best description of C++'s flaws is at http://yosefk.com/c+
+fqa/defective.html.
150
Tools
was reliable, rich, and easy to use. We were able to tweak the UI
based on feedback right up until the end simply because we could
without breaking things. Modern tools can go everywhere, even to
devices constrained by cost and size, and their greater use will
make consumer devices easier to use.
Conclusion
We can only see a short distance ahead, but we can see plenty
there that needs to be done.
Alan Turing, father of modern computer science
Their policy on the problem of maintenance was but a game
they seemed to be playing with a piece of rubber that could be
stretched a little, then a little more.
Ayn Rand, Atlas Shrugged
Phew thanks for making it through all that! I had to explain GC

because the fact that it is missing from so much code is the biggest
reason why programs crash and people don't trust computers today.
When software, the most fundamental building block of mankind's
21st century intellectual activity, protect themselves with copyleft,
and use a standard GC programming language, mankind will enter a
shining age. Perhaps we only need free software, but I don't think
we can ever get there with C/C++, which are missing an important
piece of technology invented 50 years ago.
Software, like math, is a pure world. Any aliens who have
invented software will have created the same Quicksort algorithm
we have. One of my mentors at Microsoft, Eliyezer Kohen, designed
some of the best, and most complicated, code at Microsoft. He
showed that code should be written as carefully as a mathematician
constructs a proof. The compiler, the operating system, and everything else is simply the external machinery created to run your perfect software. His thoroughness left a strong impression on me, but
better tools make it easier to write and share code, no matter what
language you use. In fact, with better languages and runtimes, you
can make your code perfect faster.
If Microsoft were to move its important code to C# and the free
software world were to fail to move their code away from C/C++,
Microsoft could leave them in the dust. In general, the task of converting the free software codebases to better tools is much easier
than porting Microsoft's code because the free codebases are ten
years younger, and much smaller and simpler. In fact, Microsoft's
codebases are so big and old that I'm not sure they could port them,
Tools
151
and if they did, they'd likely break backward compatibility so badly

that users wouldn't necessarily switch to the new version. If they did
provide perfect compatibility, they'd build something almost as complicated as what they have today, and they wouldn't be able to
divorce themselves from 20 years of legacy.
Switching to modern tools could be a huge competitive advantage
for the free software community. The tremendous inefficiency of the
free software community is its greatest weakness, and tools play an
enormous part of that, and is a big part of how millions of free software programmers can continue to lose to Microsoft's thousands.
Upgrading to modern tools is a task that few of the important
teams in free software are even considering. Every free software
group needs to create a five-year plan and start today. We must do
this for reliability reasons, but we will receive many other benefits.
What language should they choose? That is the subject of the next
chapter.
152
The Java Mess
THE JAVA MESS

I think everybody hates Java as a desktop thing. I see Java mentioned a lot lately, but all of the mentions within the last year
have been of Java as a server language, not as a desktop language. If you go back a year and a half, everybody was talking
about Java on the desktop. They aren't anymore. It's dead. And
once you're dead on the desktop, my personal opinion is you're
dead. If servers are everything you have, just forget it. Why do
you think Sun, HP everybody is nervous about Microsoft?
It's not because they make great servers. It's because they control the desktop. Once you control the desktop, you control the
servers.
It's no longer something that will revolutionize the industry. It
could have revolutionized the industry if it was on the desktop,
but I don't see that happening anymore. I hope I'm wrong.
Really. I just don't think I am.
Linus Torvalds, 1998
A company should build a process that systematically looks at
every product, every service, every process, every policy, every
market with the question, If we weren't doing this already,
knowing what we now know, would we start it, would we go
into it? If not, how quickly can we get out?
Peter Drucker
class JavaProgram
{
public static void main(string args[])
{
int counter = 1; //Add up all the integers from 1 thru 10
int sum = 0;
//Store value in variable sum
while (counter <= 10)
{
sum = sum + counter;
counter = counter + 1;
}
System.out.println("Numbers 1-10 add to: " + sum);
}
}
Java is relatively elegant, and should have replaced C and C++. The highlighted portions show the very few places that would need changing to port
to C#.
n 1995, Sun Microsystems created a next-generation programming language called Java, something that could have been the
most significant part of their legacy, more important than their
Sparc processor, Solaris (their flavor of Unix), or anything else. Java
The Java Mess
153
had the potential to change the software industry by becoming the

next generation C/C++, but now I think it is a dying language. (Java
is similar in name to Javascript, the programming language of web
browsers, and both are similar to C, but the languages and runtimes
are incompatible and different.)
At first glance, Java looks quite like C, and very much like
Microsoft's C#. Sun's Java adopted much of the syntax and semantics of C but added object orientation and many other language
innovations, including garbage collection, and they made no effort
to be 100% backwards compatible like C++ did.
While Java has many important technical advancements, it has
achieved only a small fraction of the universal status it should have.
Java should have replaced C and C++, languages desperate to be
taken out back and shot! However, Java did not, and one of the biggest reasons why software is in shambles today is because Sun
repeatedly screwed the pooch with Java.
In mid-2006, after Scott McNealy's departure as CEO, Sun began
making Java free, attempting to solve some of the issues mentioned
in this chapter, but Java grew up as proprietary and this has left
indelible scars on it. I will discuss why making it free won't change
things at the end of the chapter.
For those not in the software business, it is hard to describe the
ire many programmers have felt towards Sun. This frustration is
rooted in the special part tools play in the life of a software developer. A programmer is a tinkerer; he tinkers with his own tools as
much as he concentrates on the job at hand. Software tools are continuously refashioned. If I need to process columns of data in a text
file I receive from the Internet, it is almost as easy to fashion a tool
to accomplish the task as it is to do it by hand. (A spreadsheet would
be a useful tool, but there are other possible ones. A web server isn't
going to load a spreadsheet to extract and process simple data from
a text file.) A suitable tool might not exist, or might not be exactly
suitable to the particular task until it is tinkered with and made into
the proper tool for the job.
154
The Java Mess
Sun locked up the code

Brilliance is typically the act of an individual, but incredible stupidity can usually be traced to an organization.
Jon Bentley
Sun's first mistake was that they failed to do what Bell Labs did
with C and C++: create freely available compilers and runtimes for
people to experiment with and extend. Instead, Sun locked up the
Java codebase, letting few see it, and letting even fewer improve it,
so that today, there exists only a small community of people, outside
of Sun itself, improving Java.
For example, because Sun locked up their code, no one was able
to port it to other processors. As I mentioned in the Linux chapter,
Debian calls itself The Universal Operating System because it contains 18,000 software components that run on 15 different processor architectures:
Intel x86 / IA-32
AMD64
Motorola 68k
Sun SPARC
Alpha
Motorola/IBM PowerPC
PowerPC 64-bit
ARM
MIPS CPUs
HP PA-RISC
IA-64
S/390
SuperH
Big-Endian ARM
Renesas's 32-bit RISC
Debian supports 15 processors, but Sun's Java web page lists just four.
A wide variety of platforms allow the programmer to decide

which processor is the best choice for his product. By contrast, on
Sun's website, you can only download four architectures: x86, PowerPC, AMD64 and Sparc.
The fact that Java doesn't offer broad support on hardware platforms is a significant consideration. Why invest in a programming
language that might not be usable with future hardware? C, and
other free languages, provide tremendous hardware flexibility in
their widespread adoption.
Flexibility and adaptability of programming tools is especially
important for embedded devices because, unlike the PC environ-
The Java Mess
155
ment where processors and operating systems have similar functionality, the software must be customized for embedding on low-end
hardware, something that Sun did not enable.
For its first five years, when everyone was seriously considering
using it, Java ran ten times slower than C because it was interpreted
rather than compiled, exactly like Lisp. This would have been fixed a
lot faster if Sun had involved and encouraged the existing free compiler development community.
Even in 2008, Java programs on my computer don't look like they
should. For example, when a program asks to display a file chooser
dialog box to the user, below are the results for a Java and a native
application:
Java file chooser on Linux

Standard Linux file chooser
How does Java not look native in 2008? Let us count the ways.
Java had many significant limitations for many years because all
progress was held up by Sun.
156
The Java Mess
Sun obsessed over specs

A spec is close to useless. I have never seen a spec that was
both big enough to be useful and accurate.
And I have seen lots of total crap work that was based on specs.
It's the single worst way to write software, because it by definition means that the software was written to match theory, not
reality.
So there's two MAJOR reasons to avoid specs:
1. They're dangerously wrong. Reality is different, and
anybody who thinks specs matter over reality should
get out of kernel programming NOW. When reality and
specs clash, the spec has zero meaning. Zilch. Nada.
None.
It's like real science: if you have a theory that doesn't
match experiments, it doesn't matter how much you like
that theory. It's wrong. You can use it as an approximation, but you MUST keep in mind that it's an approximation.
2. Specs have an inevitable tendency to try to introduce
abstractions levels and wording and documentation
policies that make sense for a written spec. Trying to
implement actual code off the spec leads to the code
looking and working like CRAP.
The classic example of this is the OSI network model protocols.
Classic spec-design, which had absolutely zero relevance for
the real world. We still talk about the seven layers model,
because it's a convenient model for discussion, but that has
absolutely zero to do with any real-life software engineering. In
other words, it's a way to talk about things, not to implement
them.
And that's important. Specs are a basis for talking about things.
But they are not a basis for implementing software.
So please don't bother talking about specs. Real standards
grow up despite specs, not thanks to them.
Linus Torvalds
Imagine designing a car by a committee on a computer, and then

sending the design directly to the assembly line without building a
car first and letting anyone test-drive it.
Sun didn't let anyone read or tinker with their code but they
wanted to be seen as a community-oriented effort, so, they created
the Java Community Process where human-readable specifications
describing Java were produced. Unfortunately, they focused on the
specs and not on releasing software for people to try out so that the
feedback came years later, or not at all. The biggest feedback that
The Java Mess
157
Sun would have received, but did not, was how insanely complicated
their Java specs were. As an example, here is what a menu item
looks like on the screen:
and here are the 456 functions a Java MenuItem class implements:1
action, actionPropertyChanged, add, addActionListener, addAncestorListener, addChangeListener, addComponentListener,
addContainerListener, addFocusListener, addHierarchyBoundsListener, addHierarchyListener, addImpl, addInputMethodListener,
addItemListener, addKeyListener, addMenuDragMouseListener, addMenuKeyListener, addMouseListener, addMouseMotionListener,
addMouseWheelListener, addNotify, addPropertyChangeListener, addVetoableChangeListener, applyComponentOrientation,
areFocusTraversalKeysSet, bounds, checkHorizontalKey, checkImage, checkVerticalKey, clone, coalesceEvents, computeVisibleRect,
configurePropertiesFromAction, contains, countComponents, createActionListener, createActionPropertyChangeListener,
createChangeListener, createImage, createItemListener, createToolTip, createVolatileImage, deliverEvent, disable,
disableEvents, dispatchEvent, doClick, doLayout, enable, enableEvents, enableInputMethods, equals, finalize, findComponentAt,
fireActionPerformed, fireItemStateChanged, fireMenuDragMouseDragged, fireMenuDragMouseEntered, fireMenuDragMouseExited,
fireMenuKeyPressed, fireMenuKeyReleased, fireMenuKeyTyped, firePropertyChange, fireStateChanged, fireVetoableChange,
getAccelerator, getAccessibleContext, getAction, getActionCommand, getActionForKeyStroke, getActionListeners, getActionMap,
getAlignmentX, getAlignmentY, getAncestorListeners, getAutoscrolls, getBackground, getBaseline, getBaselineResizeBehavior,
getBorder, getBounds, getChangeListeners, getClass, getClientProperty, getColorModel, getComponent, getComponentAt,
getComponentCount, getComponentGraphics, getComponentListeners, getComponentOrientation, getComponentPopupMenu, getComponents,
getComponentZOrder, getConditionForKeyStroke, getContainerListeners, getCursor, getDebugGraphicsOptions, getDefaultLocale,
getDisabledIcon, getDisabledSelectedIcon, getDisplayedMnemonicIndex, getDropTarget, getFocusCycleRootAncestor,
getFocusListeners, getFocusTraversalKeys, getFocusTraversalKeysEnabled, getFocusTraversalPolicy, getFont, getFontMetrics,
getForeground, getGraphics, getGraphicsConfiguration, getHeight, getHideActionText, getHierarchyBoundsListeners,
getHierarchyListeners, getHorizontalAlignment, getHorizontalTextPosition, getIcon, getIconTextGap, getIgnoreRepaint,
getInheritsPopupMenu, getInputContext, getInputMap, getInputMethodListeners, getInputMethodRequests, getInputVerifier,
getInsets, getItemListeners, getKeyListeners, getLabel, getLayout, getListeners, getLocale, getLocation, getLocationOnScreen,
getMargin, getMaximumSize, getMenuDragMouseListeners, getMenuKeyListeners, getMinimumSize, getMnemonic, getModel,
getMouseListeners, getMouseMotionListeners, getMousePosition, getMouseWheelListeners, getMultiClickThreshhold, getName,
getNextFocusableComponent, getParent, getPeer, getPopupLocation, getPreferredSize, getPressedIcon, getPropertyChangeListeners,
getRegisteredKeyStrokes, getRolloverIcon, getRolloverSelectedIcon, getRootPane, getSelectedIcon, getSelectedObjects, getSize,
getSubElements, getText, getToolkit, getToolTipLocation, getToolTipText, getTopLevelAncestor, getTransferHandler, getTreeLock,
getUI, getUIClassID, getVerifyInputWhenFocusTarget, getVerticalAlignment, getVerticalTextPosition, getVetoableChangeListeners,
getVisibleRect, getWidth, getX, getY, gotFocus, grabFocus, handleEvent, hasFocus, hashCode, hide,
imageUpdate,fireMenuDragMouseReleased isBorderPainted, init, insets, inside, invalidate, isAncestorOf, isArmed,
isBackgroundSet, isContentAreaFilled, isCursorSet, isDisplayable, isDoubleBuffered, isEnabled, isFocusable, isFocusCycleRoot,
isFocusOwner, isFocusPainted, isFocusTraversable, isFocusTraversalPolicyProvider, isFocusTraversalPolicySet, isFontSet,
isForegroundSet, isLightweight, isLightweightComponent, isManagingFocus, isMaximumSizeSet, isMinimumSizeSet, isOpaque,
isOptimizedDrawingEnabled, isPaintingForPrint, isPaintingTile, isPreferredSizeSet, isRequestFocusEnabled, isRolloverEnabled,
isSelected, isShowing, isValid, isValidateRoot, isVisible, keyDown, keyUp, layout, list, locate, location, lostFocus,
menuSelectionChanged, minimumSize, mouseDown, mouseDrag, mouseEnter, mouseExit, mouseMove, mouseUp, move, nextFocus, notify,
notifyAll, paint, paintAll, paintBorder, paintChildren, paintComponent, paintComponents, paintImmediately, paramString,
postEvent, preferredSize, prepareImage, print, printAll, printBorder, printChildren, printComponent, printComponents,
processComponentEvent, processComponentKeyEvent, processContainerEvent, processEvent, processFocusEvent,
processHierarchyBoundsEvent, processHierarchyEvent, processInputMethodEvent, processKeyBinding, processKeyEvent,
processMenuDragMouseEvent, processMenuKeyEvent, processMouseEvent, processMouseMotionEvent, processMouseWheelEvent,
putClientProperty, registerKeyboardAction, remove, removeActionListener, removeAll, removeAncestorListener,
removeChangeListener, removeComponentListener, removeContainerListener, removeFocusListener, removeHierarchyBoundsListener,
removeHierarchyListener, removeInputMethodListener, removeItemListener, removeKeyListener, removeMenuDragMouseListener,
removeMenuKeyListener, removeMouseListener, removeMouseMotionListener, removeMouseWheelListener, removeNotify,
removePropertyChangeListener, removeVetoableChangeListener, repaint, requestDefaultFocus, requestFocus, requestFocusInWindow,
resetKeyboardActions, reshape, resize, revalidate, scrollRectToVisible, setAccelerator, setAction, setActionCommand,
setActionMap, setAlignmentX, setAlignmentY, setArmed, setAutoscrolls, setBackground, setBorder, setBorderPainted, setBounds,
setComponentOrientation, setComponentPopupMenu, setComponentZOrder, setContentAreaFilled, setCursor, setDebugGraphicsOptions,
setDefaultLocale, setDisabledIcon, setDisabledSelectedIcon, setDisplayedMnemonicIndex, setDoubleBuffered, setDropTarget,
setEnabled, setFocusable, setFocusCycleRoot, setFocusPainted, setFocusTraversalKeys, setFocusTraversalKeysEnabled,
setFocusTraversalPolicy, setFocusTraversalPolicyProvider, setFont, setForeground, setHideActionText, setHorizontalAlignment,
setHorizontalTextPosition, setIcon, setIconTextGap, setIgnoreRepaint, setInheritsPopupMenu, setInputMap, setInputVerifier,
setLabel, setLayout, setLocale, setLocation, setMargin, setMaximumSize, setMinimumSize, setMnemonic, setModel,
setMultiClickThreshhold, setName, setNextFocusableComponent, setOpaque, setPreferredSize, setPressedIcon,
setRequestFocusEnabled, setRolloverEnabled, setRolloverIcon, setRolloverSelectedIcon, setSelected, setSelectedIcon, setSize,
setText, setToolTipText, setTransferHandler, setUI, setVerifyInputWhenFocusTarget, setVerticalAlignment,
setVerticalTextPosition, setVisible, show, size, toString, transferFocus, transferFocusBackward, transferFocusDownCycle,
transferFocusUpCycle, unregisterKeyboardAction, update, updateUI, validate, validateTree, wait
The list of functions the Swing Java MenuItem class implements, with
bizarro names like isFocusTraversalPolicyProvider and addVetoableChangeListener. Imagine if you needed to become familiar with 456
things to use your oven.
Such bloated code is hard to understand and even harder to make

reliable. And this is just the Swing class library's opinion of a menu
item, one of three popular implementations of Java widgets, and
supposedly an efficient one!
Other widget libraries, like the Gtk#'s MenuItem have hundreds functions as well,
but many do not and Java is the worst. The WxWidgets MenuItem class has 50
members, and it uses native widgets on Mac, Windows and Linux. Apple's NSMenuItem has 55.
158
The Java Mess
Java is plagued by too much complexity as you can see, and

releasing specs instead of code is a big cause of this problem. In
fact, many of the Java specs aren't widely used because they do not
meet developers' needs, or they became heavily amended after they
finally got into customers' hands, which was often years after the
specs were originally written.2
Java was not growing carefully and organically, as Linux has
done, by following the ship early, ship often methodology. A lot of
people who hate Java do so because of its pervasive extraneous complexity.
Sun locked up the design

When it created Java, Sun didn't want to release the code because
they were worried people would take their code, tweak it slightly to
better meet their needs, and create incompatibilities that would hurt
the brand of Java confusing a 99% friend with an enemy. This
paranoia occurred because Sun made a rookie mistake in not understanding that a language needs to be extensible.
In the C and C++ world, there is even a convention that any nonstandard keywords would be prefaced with two underscores, like
this: __inline. In some cases, a feature started out as a vendor-specific addition, was adopted by others, and eventually became added
to the language in an official way. The creators of C and C++ didn't
have the hubris to believe they had built a perfect language.
Even more importantly, the primary way to extend a computer
language is by creating new functions and data types, which was
also not allowed in any of the domains specified by Sun.
Sun got it backwards people will only make minimal changes
necessary to Java to make it work for them. That is what happened
in C and C++. Sun didn't have faith in its customers and didn't
understand the realities of how large bodies of systems-level code
are created.
Two good examples of basically unused class libraries are:

1. JXTA, Java's peer-to-peer protocol that standardizes many aspects of P2P but is
not used because it solves the wrong problems, and,
2. J2ME, Java's underpowered class library for mobile devices which has primitive
widgets, is unable to access a phone's address book, calendar, camera, GPS, or
web browser, and has no ability to make phone calls. J2ME's goal should be rich
enough to build an entire phone user interface in Java, but it only has the ability
to make toy applications, and even then, programmers often complain that the differences between the implementations make it a write once, test everywhere
scenario.
The Java Mess
159
Sun fragmented Java

Because of fear, Sun decided to lock up their code to prevent it
from being extended by anyone else. This short-sighted idea, but
even worse, the law of unintended consequences came up and bit
them in the tush. Because Sun made only the specs but not the code
available, many third parties started producing their own Java runtimes so they could control their own destiny. Wikipedia documents
34 different third-party Java runtimes, and their list doesn't include
any of the embedded Java implementations, which are harder to
count.
Everyone who wanted to use but improve Java would have to start
from scratch. This created much bigger areas of incompatibility
because building a Java implementation from scratch requires a lot
of work, and they could even unknowingly cut corners.
The biggest problem with so many runtimes, besides the wasted
efforts, is that the imperfect specs became the reference rather than
the code because there wasn't an agreed upon codebase containing
the official Java design.
Backward compatibility is the greatest cause of complexity, and
therefore unreliability, in computers today. As described in the Linux
chapter, one of the Linux kernel's major innovations is that it avoids
backward compatibility issues by encouraging all code into one tree.
This makes it much easier to make changes to the design.
With an official source tree, Java could become whatever we need
it to be, avoiding the mess of different codebases, and serving as a
reference when the paper specs are ambiguous. It sounds chaotic,
but things quickly converge, and on something with minimal excess
complexity, because it can be removed.
While the Java specs were not good because feedback came years
later, they became etched in stone because once 30 runtimes have a
particular piece of functionality, it becomes very hard to change
them all. Even if Sun let you alter their design, it might take years to
make any changes. In contrast, if the Linux kernel is missing a feature, you can send some code to Linus and have it in the official
codebase immediately.
160
The Java Mess
Sun sued Microsoft

Another Sun mistake was getting into a war with Microsoft over
Java. Microsoft's tools division was an active participant in Java for
its first several years, and PC Magazine wrote in 1997 that:
Microsoft's Java environment was the fastest and most compatible
on our tests.
When Microsoft improved Java in certain ways that by definition
were incompatible, Sun sued them. Microsoft's changes to Java
were minor, the features justifiable, and the incompatibilities almost
non-existent.3 A public lawsuit, the exchange of $20 million in damages, and much bitterness ensued.
Eventually, Microsoft quit supporting Java and went off to create
C# instead. C# might be considered Java 2.0. While it is very similar
to Java, it had five years of refinement and applied Microsoft's many
years of experience as a dominant force in the tools business. C#
created unnecessary chaos and confusion in the industry and
allowed Microsoft to be 100% incompatible with Java rather than
just .01% incompatible. Microsoft would not have created C# if Sun
hadn't sued them.
Java as GPL from Day 0

If Sun had created Java with a GPL license, the landscape of the
software industry would look completely different today, and computers would be more reliable and smarter. Java would be more reliable, simpler, richer, and faster, and a community, like C and C++,
that Microsoft tools developers would have been a part of.
Many more programmers would have contributed to Java and
built something much better. These mistakes stunted Java's growth,
but even worse, it caused many software programmers just to con-
One of the things Microsoft changed was to simplify and improve the performance
of how a Java developer called in to native operating system functionality. Java has
always allowed developers to write non-portable, operating system-specific code
when the operating system provided a feature that a developer wanted access to
but that Java did not support. By definition this code was not cross-platform, so
one would imagine that Microsoft-specific syntax to access Windows-specific features would not have been a big deal. In fact, the way a Java programmer would
call the native functionality on Windows had the same syntax between Microsoft's
RNI and Sun's JNI. It was only the way the native method was declared that was
different. Microsoft did make other changes, such as adding a keyword, delegate, but if you search the web you can find many people who wished Sun had
added that feature and code samples to enable that feature in Java.
The Java Mess
161
tinue to plod along in their C or C++ codebases. It gave rise to C#

and to other free languages like Javascript, PHP, Python, and Ruby.
These were created to fill niches that Java didn't meet.
The success of C was in part due to its widespread support, which
made it easier to share code than did its predecessor on the PC
assembly language. It is hard to share code across languages, and
the proliferation of new ones has gotten worse since Java was created, when it should have gone in the opposite direction.4 Even
within Sun, there is a group of people trying to build another programming language called Fortress, which will be suitable for scientific computing. They too believe Java is not sufficient as it is.5
Sharing code across languages has historically been very difficult. The search
engine Lucene is a recent example of something which started in Java but has
been forked into versions in PHP and C#.
There are many reasons for it being so difficult to share code between different
languages. They have different ways of deploying the code, different naming conventions, and different low-level details. Interoperating layers usually adds significant performance costs because everything from the strings on up the pyramid
have to be marshaled, which usually involves making a copy of the data. If you
can't agree on what a string is, then you will have difficulty sharing code. (Going
from C# to C/C++ is fast because a copy of the data doesn't need to be created
a C Unicode string is a subset of a C# string.)
Fortress's syntax is arbitrarily different from Java in ways not at all related to high
performance computing (HPC). A programming language is by definition extensible by creating new functions and classes, and it is also possible to alter the runtime, without changing the language, to optimize it for specialized hardware, etc.
The big challenge in HPC is to robustly divide up the work amongst a farm of
servers, pass messages and maintain and monitor the system new software, not
new language features.
162
The Java Mess
Pouring Java down the drain

The Lord said, If as one people speaking one language they
have begun to do this, then nothing they plan to do will be
impossible for them.
Genesis 11:66
Scott McNealy, co-founder and CEO of Sun Microsystems from 1984 2006,
and an impediment to free software even though his company made most
of their money on hardware.
Sun's first major decision after co-founder Scott McNealy's departure in mid-2006 was a promise to make much of their software,
including Java, free. This announcement was big news and some
6
My bible says God scattered the Babylonians because of their pride. I can't
know what they thought, but I do believe that learning about ourselves, using
technology to save and improve lives, and attempting to understand the true
nature of the world, only gives us more appreciation for what He has created. I
believe the creator of liberty and science wants us to understand these concepts,
as long as we don't lose track of our relationship with the creator somehow in the
process.
The Java Mess
163
think it marks the potential for a new day in the life of Java. However, this effort is starting more than a decade after Java's creation,
arguably too late to fix most of the problems.
Even with Sun's new GPL license, there is a lot of Java code out
there that is not free and cannot be made so by Sun. Apple, for
example, took Sun's PC implementation and worked extensively to
port it to Macintosh. This code isn't free, and it doesn't seem likely
that Apple will ever make it such. There is a company making Java
run on mainframes, and their enhancements are proprietary as well.
Even within the code Sun has recently made free, there are pieces
Sun licensed from third parties, and they are having to re-write it
because they can't get permission to change the license.
Even a free Java on its current trajectory will not absorb other
communities. C# is considered better than Java, so those programmers are not making a switch. There are millions of web pages written in PHP, and I don't see them rushing to switch to Sun's Java
either. Sun's code has been locked up for so long, there are few people in the outside world able to contribute. Why should the community fix a mess that was created only because Sun didn't work with
the community in the first place?
C# and .Net is the dominant platform on Windows, PHP is the
dominant language on the Web, and Python and Mono are the most
popular GC runtimes on the Linux desktop. So while Java is taught
in universities, used for custom applications in enterprises, and has
some success on mobile phones, these niches aren't enough to to
ensure its long-term viability. Java has accumulated so much baggage, only some of which I have discussed here, that I think the software community should abandon it. This would also take a big step
in lessening the problem of too many programming languages. In
earlier drafts of my book, I proposed Sun create a next generation
programming language, but I now believe there already are suitable
codebases: Mono and Python.
164
The Java Mess
Let's Start Today

All economic activity is by definition high risk. And defending
yesterday that is, not innovating is far more risky than
making tomorrow.
Peter Drucker
When there's a will to fail, obstacles can be found.
John McCarthy
Sometimes the real hurdle to renewal is not a lack of options,
but a lack of flexibility in resource allocation. All too often,
legacy projects get richly funded year after year while new initiatives go begging. This, more than anything, is why companies regularly forfeit the future -- they over invest in what is
at the expense of what could be.
New projects are deemed untested, risky, or a diversion of
resources. Thus while senior execs may happily fund a billiondollar acquisition, someone a few levels down who attempts to
borrow a half-dozen talented individuals for a new project, or
carve a few thousand dollars out of a legacy budget, is likely to
find the task on par with a dental extraction.
The resource allocation model is typically biased against new
ideas, since it demands a level of certainty about volumes,
costs, timelines, and profits that simply can't be satisfied when
an ideal is truly novel. While it's easy to predict the returns on a
project that is a linear extension of an existing business, the
payback on an unconventional idea will be harder to calculate.
Managers running established businesses seldom have to
defend the strategic risk they take when they pour good money
into a slowly decaying business model, or overfund an activity
that is already producing diminishing returns.
How do you accelerate the redeployment of resources from
legacy programs to future-focused initiatives?
Gary Hamel, The Future of Management
When software, the most important building block of mankind's

intellectual property, uses a standard GC programming language
and protects itself with copyleft, mankind will enter a shining age. I
think better runtimes are critical to the success of Linux on the
desktop.
Unfortunately, fixing this problem is a big task because inertia is
such a powerful force. The good news is that the difficulty of porting
to a different language is not the same in every case. Porting from
one modern language to another is ten times easier than moving
from C to C#. Right now, much of our code and infrastructure lives
in the dark ages, and this first port is the hardest, but it will provide
big benefits.
The Java Mess
165
During my tenure at Microsoft, I witnessed numerous re-writes.

Unfortunately, most of them failed. Those that failed had at least one
thing in common: they usually were re-writing C code in C, or C++,
and because they did not use significantly better tools as part of
their plan, they never ended up saving any time in the end. It doesn't make sense to re-build something from scratch unless you plan
to make it a lot better. But if the teams out there move from C or
C++ to a much more productive language, they will save time in the
end and it is much more likely to succeed assuming they start.
The key is to make a plan and break down the problem. For example, the OpenOffice word processor could use the existing layout
engine written in C and replace the UI with one written in a GC language. This first step would give the team expertise to help them
make the subsequent steps. They also might find that a bold vision
will bring in new contributors, especially if it is easier and more
enjoyable to work on the new codebase. Whatever choice makes
sense for each team, the important teams like Firefox, OpenOffice,
and Gnome should agree on a language because that will bring
along the rest of the Linux desktop.
Today, the most popular GC runtimes on the Linux desktop today
are Mono and Python. Both are built by a worldwide community, but
each has certain advantages and disadvantages.
One of Mono's biggest advantages is that it is compiled. While
performance is not the most important factor of software quality,
any language which wants to become the standard for Linux needs a
compiler. Even if only a small fraction of any piece of software is
performance-critical, the language doesn't know what that part is,
and so it should run all code fast. Another advantage of Mono is that
it comes with a complete end-to-end solution for creating programs:
a graphical editor and debugger, code browser with auto-complete,
etc. With Python, you must piece together these components yourself, and there isn't a standard one with critical mass.
Mono's biggest disadvantage stems from the fact that it is an
implementation of .Net which is a standard created by Microsoft.
The C# language and the core libraries are not created with much
input from the worldwide community and the enhancements are
improved only on Microsoft's schedule. Microsoft has hired some of
the best language designers on the planet, like Anders Hejlsberg,
but he is not as smart as the combined knowledge of the rest of the
166
The Java Mess
languages designers on the planet. This has practical effects: C#

added dynamic language features in the version 4.0 of the language, while Python has had these for many years.7
Another disadvantage of Mono is that there is an enormous
amount of FUD (fear, uncertainty and doubt) that surrounds it.
Microsoft has historically been hostile towards free software, calling
it a cancer and sometimes threatening to sue over claimed patent
infringements, and while this should create fear all over the free
software stack, it appears to be directed mostly at Mono!
I believe Linux programmers are being irrational because the
.Net specification is publicly available, and Microsoft has repeatedly
promised not to sue for implementations of it. Programmers should
also realize that Mono extends the reach of .Net to other processors
and operating systems, which benefits Microsoft's customers.
One other concern about Mono is that it doesn't use a copyleft license. Excerpt
from the FAQ on the Mono website:
When a developer contributes code to the C# compiler or the Mono runtime
engine, we require that the author grants Novell the right to relicense his/her
contribution under other licensing terms. This allows Novell to re-distribute the
Mono source code to parties that might not want to use the GPL or LGPL versions
of the code. Particularly: embedded system vendors would obtain grants to the
Mono runtime engine and modify it for their own purposes without having to
release those changes back.
The Java Mess
167
Python's advantage is that it is created with input from programmers all over the world, and has a larger set of libraries than any
other GC language:
A computational fluid dynamic (CFD) visualization of a combustion chamber. The Python community has created a wide variety of libraries for everything from gaming to scientific computing.
Python's biggest disadvantage is performance: it is not compiled

and doesn't support multiple threads. These features are a requirement for any serious runtime and so one might wonder whether
Python is still at a prototype phase of development. Even fans of
Python very loudly complain that while it might be suitable for little
scripts and application plugins, it is too slow for writing large applications.
These features are a big amount of work that are not even on
Python's to-do list. And as the language has never been compiled
there are even some questions as to whether it can be made fast.8 In
addition, Because the Python runtime is slow, the developers are
forced to write a lot of support code in C, which further slows its
development progress. Is is a shame that Python isn't fully using a
8
In some performance tests, the IronPython implementation on .Net runs slower

than the standard Python implementation. So even though IronPython has access
to a compiler, it apparently isn't able to take advantage of it enough to be faster.
168
The Java Mess
tool which would have made their job easier their own language.
There are efforts such as the PyPy project which has built a Python
compiler, in Python, which outputs C. Unfortunately, this piece of
elegance is not yet the mainline codebase and is not considered for
such.9 And as mentioned earlier, the language doesn't have a standard graphical programming and debugging environment. So
Python today has impediments for both casual programmers looking
for an easy way to get into programming, and professionals who
care about building high-performance applications.
There are other interesting languages and runtimes out there, but
I believe the Linux desktop community should focus on these two.
Some have proposed merging the Python language with the Mono
runtime. Mono already supports other languages in addition to C#,
some that even look like Python.
There many good programming languages, in fact there are too
many, but I also think they are mostly good enough. One could further tweak the letter of the English language to make it easier for
your eyes to distinguish the letters, but it isn't necessary because
what we have is good enough. Likewise, it is much more important
to build a complete set of libraries for all aspects of computing, a
Wikipedia of free code, than to worry that further language innovation is the gating factor towards future progress in software.
One of the biggest missing features is that it doesn't support Python extension
modules.
The OS Battle
169
THE OS BATTLE
Free software works well in a complex environment. Maybe
nobody at all understands the big picture, but evolution doesnt
require global understanding, it just requires small local
improvements and an open market (survival of the fittest).
Linus Torvalds
I've been a big proponent of Microsoft Windows Vista over the
past few months, even going so far as loading it onto most of
my computers and spending hours tweaking and optimizing it.
So why, nine months after launch, am I so frustrated? The litany
of what doesn't work and what still frustrates me stretches on
endlessly.
Take sleep mode, for example. Vista promised a new low-power
sleep mode that would save energy yet enable nearly instantaneous resume. Poppycock. The brand-new dual-core system I
built a few months ago totters off to sleep but never returns. I
have to cold-start it to bring it back. This after replacing virtually every driver inside.
Take my media center PC, for example. It's supposed to serve
up photos, videos, and music. Instead, it often simply drops off
the network for absolutely no reason.
I could go on and on about the lack of drivers, the bizarre wakeup rituals, the strange and nonreproducible system quirks, and
more. But I won't bore you with the details.
Jim Louderback, Editor in Chief of PC Magazine.
indows revenue of $16 billion is larger than the GDP of

many countries, and it is this dominance that provides
technical and financial competitive advantages that fuel
Microsoft's success everywhere else. The battle for the desktop
computing OS is an epic struggle seldom seen in the history of business, and the success of Linux on the desktop will reinvigorate the
PC as a tool much more powerful than a word processor and a web
browser.
To a Linux distribution, the kernel is just the software that makes
all the other software run, and is one of the thousands of components they integrate. The Linux kernel by itself will not defeat Windows it requires an entire distribution. Therefore, it is worth
analyzing the state of the OS market.
There are many producers of Linux distributions, but in the PC
world the four most important teams are: Red Hat, Novell, Debian,
and a four year-old upstart, Ubuntu. There are hundreds of other
Linux distributors, but most of them are merely using the big four's
170
The OS Battle
efforts and tweaking them further for more specialized markets.

Each of these four will be discussed over the next few pages, but it
is worth mentioning a name that isn't on the list, IBM.
IBM
Prodigy, a powerful ad describing the power of free software, created by

IBM in 2003. Given IBM's tepid support for free software in 2008, their
management should watch this ad a few more times.
IBM was the first, and is still the biggest computer company, and
has built many operating systems over the decades. However, it has
yet to exhibit an interest in producing a Linux distribution for PCs.
In fact, you cannot get Linux pre-installed on any of their computers
today, even though IBM/Lenovo laptops are extremely popular in the
Linux community. It is possible they gave up hope after their
antitrust lawsuit in the 1970s, the distraction of mainframes and
minicomputers, and the expensive and humiliating defeat of OS/2 by
Windows in the early 1990s.
IBM has touted its support of free software and Linux for eight
years, but has done very little to even ensure its hardware runs
smoothly on Linux. I met an Intel employee whose job was to write
device drivers for IBM hardware because IBM wasn't working on
The OS Battle
171
the problem. On my IBM laptop, the system generally works, but a

number of the devices do not have Linux support.1 IBM could have
easily done the work, or just put a little pressure on its suppliers to
get them to write the drivers. If you can design hardware, you can
easily write the corresponding drivers!
In general, various people have labored to make IBM's hardware
work for the important user scenarios. But if I un-dock my DVDplayer on my IBM laptop, Linux will hang. This sort of dynamic functionality isn't needed, as one can just swap out parts when the computer is off. This code is tricky, the work is boring, and it is not
necessary, thus the last bits like this have been left out. It is the sort
of code that will not become functional until IBM understands free
software can achieve world domination if only they'd lend a hand!
In addition to the above issues, IBM has made chunks of their
code free, but still left a fair amount proprietary. For example, random products like Notes, Rational and Websphere have long been
sold to enterprises. I wouldn't run those old dogs on my computer at
any price, which is why IBM might feel that if they've got suckers
who are willing to continue to pay for it, then they should just continue to take their money!
Nevertheless, there still must surely be a divide inside the company on the utility of free software. IBM has purchased the rights to
some code and made them free, but often the codebases are
ungainly (Eclipse developer tool) or irrelevant (Cloudscape database). The complacency of IBM and others toward Linux and free
software contributes significantly to the continuing dominance of
proprietary software.
These are the most notable devices that don't work: the fingerprint reader, the
broadband modem and the accelerometer which can signal to the hard drive to
park itself if the computer senses it is being dropped.
There are also missing utilities, like the ability to enable an external monitor, the
ability to recondition the laptop battery, or set the maximum charge to only 95%
of capacity to lengthen the lifetime of the battery. All of these exist on Windows.
Then there is the gray area of support: my old laptop contained an ATI graphics
card that generally worked, but because it the driver was proprietary, didn't
support 3-D, and was buggy.
172
The OS Battle
Red Hat
The first Linux distribution I used was Red Hat, the largest commercial Linux producer. One might conclude, therefore, that a big
part of the reason Linux hasn't taken off with desktops is because
Red Hat didn't really focus on building a user-friendly experience.
Red Hat built a platform usable by Google but not by our moms.
Instead, they focused on developers, servers, grid computing, the
web, etc. For many years, certain basic user scenarios, like setting
up a shared printer, have been cumbersome on Linux.
Red Hat's chief technology officer, Brian Stevens, was recently
asked, When is Red Hat getting into the desktop space? His
response:
To us, the desktop metaphor is dead. It's a dinosaur. Today's
users aren't sitting at home, sitting at a desk in isolation anymore. They are collaborative. They are sharing. They work and
play online. We don't believe that re-creating the Windows paradigm with just pure open source models does anything to
advance the productivity or the lives of the users.
Stevens is confusing the desktop as a personal and powerful tool

with mass appeal, with the desktop as a user interface metaphor! If
they don't get what the desktop is, it should be no surprise that the
Linux desktop hasn't taken off after 10 years of work. Furthermore,
there is nothing wrong with the current desktop metaphors of windows, icons, and menus. These concepts do allow people to work
and play online, and share documents over the Internet. We might
The OS Battle
173
come up with user interfaces like those out of science-fiction

movies, but that isn't necessary, or sufficient for that matter, to
enable the scenarios that Stevens describes.
Red Hat produces two versions of a Linux operating system: a
free, community-oriented product known as Fedora, and a commercial version known as Red Hat Enterprise Linux (RHEL), which is
typically purchased by corporations as part of a support contract.
The wisdom of this dual-market strategy is debatable because it
separates the user base and the brand. Red Hat's enterprise product
generates the most revenue, so this is where the company's focus
lies. Unfortunately, few normal users run the enterprise product
because it is rather expensive2 and it ships every 18 months, which
is a long time given the fast-moving state of free software.
Red Hat's concentration on its for-profit enterprise offerings has
distracted it from creating a healthy community of interested geeks
who could have contributed free improvements and additions to
both versions.
It is an open question as to how much revenue Red Hat would
make if they chose not to charge for their enterprise version, but
charged for support. Red Hat has been charging for their enterprise
version for many years, and effectively bundling support for free,
but in fact people typically buy RHEL because they want support,
not because the software is significantly better than the community
version. In fact, Red Hat could decide to sell support for both versions. These changes might even generate more revenue for them!
With only two million RHEL customers today, the potential for
growth is huge.
$349 - $1300 per year, depending on whether you get two-day web response or
24/7 phone response.
174
The OS Battle
Novell
Those who have heard of Novell remember it for their proprietary,

server-oriented operating system known as NetWare. The product
never gained critical mass and soon lost out to IBM and Microsoft.3
However, Novell acquired German Linux distributor SuSE in 2003
for $210 million, and the company has since refocused itself on
Linux and the free software movement.
While Red Hat has concentrated on data centers, Novell is targeting business desktops, which need directory services, management
tools, and Microsoft compatibility. Novell's experience with NetWare
and its other products gives it the institutional knowledge necessary
to understanding how to meet the needs of enterprises through free
software.
This merger appears to be working fairly well and Novell is a
well-respected company in the free software community. SuSE is
one of the most popular Linux distributions, but like Red Hat, they
have separate community and enterprise versions, and communities,
that is potentially stunting their growth.
The kernel also had fatal flaws for server scenarios: it didn't support preemptive
multitasking and ran application code in the kernel, therefore hurting reliability.
The OS Battle
175
Debian
Large organizations cannot be versatile. A large organization is
effective through its mass rather than through its agility. Fleas
can jump many times their own height, but not an elephant.
Peter Drucker
Welcome banner for the 2007 Debconf. Debian is one of the largest engineering teams you've never heard of.
176
The OS Battle
Debian, the big dog of community-based Linux distributions, is

used on servers, desktops and embedded devices, and is respected
for its power and reliability. Like Wikipedia, it is built by a community rather than a corporation, and was created by a college sophomore in 1993. The name Debian is a conjunction of his girlfriend
Debra, and Ian, the founder's first name.
Ian Murdock, a student at Purdue, wrote in his Debian Manifesto that writing an operating system is a job which is too boring
for any corporation to do well. Ian was dissatisfied with the quality
of the products produced by commercial companies building and
selling Linux distributions:
Operating systems are neither easy nor glamorous to construct
and require a great deal of ongoing effort from the creator to
keep the distribution bug-free and up-to-date: to ensure that
the system is easy for others to install, is installable and usable
under a wide variety of hardware configurations, contains software that others will find useful, and is updated when the components themselves are improved.
This argument is the reverse of the typical argument made about

free versus proprietary software. Perhaps some software, like an
operating system, is so large and tedious to maintain that no one
corporation can do a good job. Whether Ian is right or not, it is true
that the free software community can build a good product: Debian
is a completely free software distribution which has harnessed an
army of 1,000 developers to build a rich and reliable system. There
are challenges facing Debian, but none of these challenges relates
to the quality of the engineering
At last count, these engineers have incorporated 283 million lines
of code. By one metric, the total cost to write the software is $14 billion! Debian sets a high standard for inclusion into their platform,
and they have been very influential in making sure that the free software community stays reliable and free. Nobody knows when free
software will take over, but much of the code is already written and
available in Debian.
Windows and Apple's Macintosh don't get better with more users
because those users can't contribute anything technical back. Any
conversation between a customer and a Microsoft employee is just
an unprofitable distraction from writing and selling new software.
To a lesser extent, this is even true with Red Hat and Novell's Linux
because most of their development, even of the free versions, is
done mostly by full-time employees as their efforts to harness their
community of users has long been an afterthought.
The OS Battle
177
Team size is a very important metric, in fact one of the most critical ones. All other things like productivity being equal, the team that
has the most engineers will win. The biggest reason why Internet
Explorer beat Netscape is that Microsoft created a bigger team.4 A
larger team can do more, including absorb new people faster and
build up institutional expertise in more areas. The lesson of Metcalfe's law is that the first to achieve critical mass wins: Google,
YouTube, Wikipedia.
In 2007, I went to the annual Debian Conference and was very
impressed with the strength of the team; many of the attendees
were of a similar caliber to my former co-workers at Microsoft, even
though they weren't screened via a day-long intensive interview
process.
In fact, many of Debian's components are packaged, updated, and
maintained by people who are using that component for their personal or professional use. Hewlett-Packard and other companies
contribute to Debian to make it work better for their customers, and
all have a voice at the table. This perspective which transcends companies and geographies, can lead to a healthy state of affairs.
Debian has governance structures, although their leaders play a
very small role in guiding the team in any particular direction. A feature is added simply because someone decides it is a good idea.
Details are hashed out in e-mail discussions, blogs and conferences.
The person doing the work makes the final decision, which is why
free software has been called a Do-ocracy, but the Internet allows
him to get questions answered and leverage the expertise of others.
Debian's collective expertise in understanding all of the software on
its DVD is its greatest asset.
A Debian developer's primary job is to update the many free software components to the latest version and then find and fix interaction bugs between components in the system. Debian does write its
own code, but that is a small yet important part of what is actually
on a Debian CD.
Debian is a spiritual leader of the free software community and
has a very ambitious goal embodied in their motto: to be The Universal Operating System. While they have yet to achieve this objective, they have come very close. Debian contains 18,200 software
applications that run on 15 hardware platforms and support hundreds of languages.
4
Microsoft's nearly unlimited supply of smart and experienced developers, whose

expertise in text processing, windowing, control widgets and dealing with the
challenges of building large codebases meant Netscape didn't stand a chance.
178
The OS Battle
Debian's biggest mistake so far is that its releases have never

truly targeted a desktop operating environment. Their product has
shipped every two to three years, which is an eternity in the fastmoving free software world. Long shipment cycles encourage procrastination by contributors, thereby lengthening the release cycle
even more.
With software distribution, the ship date is as much a choice of
the community as it is a technical limitation. As I wrote earlier,
Debian writes little code and ships only stabilized versions of its
components. That is why they can ship on any schedule they choose.
The latest code could be pushed out every day if you had thorough
testing. In fact most Debian developers run the Debian Unstable
release which is precisely that.
While Debian has a 2-3 year release cycle, many other free software organizations have moved to more frequent releases, and
there haven't been a lot of complaints about decreasing quality. In
fact, the faster you release, the faster you can incorporate the latest
and greatest code, which is better in every way. Furthermore, frequent releases increase the excitement and discourage procrastination so that bugs get fixed faster.
Debian had the potential to be the Wikipedia of Linux distributions, but it hasn't come close to achieving this status. It was one of
the most popular Linux distributions in the early years, but it has
now fallen far behind the leaders to number seven.
So what happened? Mark Shuttleworth, a former Debian developer whoc that became a dot-com billionaire, decided he wanted his
own Linux-based operating system. And so, with $10 million and
what is arguably an unethical interpretation of the spirit of free software, created a brand-new Linux distribution system.
Enter Ubuntu which sat on a platform that was 100% Debian's.
Now, with changes and improvements, Ubuntu is 99.9% Debian, and
is taking over the Linux distribution market primarily by leveraging
the strengths of Debian.
The OS Battle
Ubuntu
Mark Shuttleworth, dot-com billionaire, space tourist, and creator of

Ubuntu Linux
179
180
The OS Battle
The GPL (General Public License) allows anyone to access

Debian's code and take it in a new direction. That is what each of
the 130 derivatives of Debian have done. However, most of that
work has so far only involved relatively small changes, mostly subsetting Debian, with little fragmentation of the Debian community or
brand. All that changed when Ubuntu was created.
What began in 2004 has become the most popular and fastestgrowing Linux distribution system. While its name is bizarre,
Ubuntu is all the buzz in the Linux community. Free software luminary Eric Raymond made a very public statement when he switched
from Red Hat to Ubuntu in February 2007. Dell announced in May
2007 that they were going to offer Ubuntu on several Dell computers, very big news indeed. The latest public numbers are that as of
November 2006, Ubuntu had eight million users and was doubling
every eight months.
Ubuntu comes from the African language Bantu and is a humanist
ideology:
A person with ubuntu is open and available to others, affirming
of others, does not feel threatened that others are able and
good, for he or she has a proper self-assurance that comes from
knowing that he or she belongs in a greater whole and is diminished when others are humiliated or diminished, when others
are tortured or oppressed.
Desmond Tutu
In short, Ubuntu means: humanity towards others, or I am

because we are. Lovely. Shuttleworth, a South African himself, with
$10 million a minuscule amount by the standards of the wellfunded IT industry and a copy of the Debian code, hired ten of its
best volunteers to work full-time.
Ubuntu took Debian's solid base of packages and added a few
innovations: they ship every six months, much more frequently than
Debian's two to three-year release cycle, and they focus on ease of
use and better support for proprietary drivers.
The OS Battle
181
Perhaps for the first time, a Linux for human beings was created,
Ubuntu's motto.
Today, Linux (and computing in general) is mostly for male human beings.
Periodically, Ubuntu designates stable releases for which they will

provide paid support. This is a different model from that employed
by Red Hat and Novell, which provide support only in their Enterprise versions that are separate products with separate life cycles.
With Ubuntu, the code is free, and there is all sorts of community
support on forums and blogs, in short, a search engine is your
friend. English majors can follow the steps to fix a problem, even if
they don't fully understand them! In fact, it is often faster to type in
a query to a search engine than to call someone who will usually ask
many silly questions (Is your printer on?)
Shuttleworth's success demonstrates that it is possible to foster a
community of volunteers around a team of hard-core, full-time, paid
developers. Ubuntu's full-timers spend their hours tracking down
the hard kernel bugs and maintaining the larger, complicated packages, which often have more arcane gruntwork associated with
182
The OS Battle
them. The volunteer community today focuses on maintaining the

bug list, creating artwork, translating the user interface, and adding
applications that Ubuntu is happy to provide on their servers but not
necessarily fully support. With a large user base, if everyone just
chips in a little, a lot of progress can be achieved.
Many Ubuntu users are passionate about the potential of free
software and want to help, but they are unable to contribute back in
any technical way. This is one of the reasons why donations of
money should be used for free software, a topic I will discuss more
in the next chapter.
Perhaps the biggest reason why money isn't donated to free software is that hardly anyone is asking. On Ubuntu's website, they
encourage you to volunteer your time, but they do not ask for
money.
Just $1 per computer per year would cover Mark Shuttleworth's
costs to run Ubuntu. Computers cost between $500 and $2,000, so
five dollars for the five years you might keep your new computer is
not very much to ask!
Should Ubuntu Have Been

Created?
I am both a Debian and an Ubuntu developer, and Im sometimes amazed that Ubuntu discusses technical choices that
were discussed (and solved) a few weeks earlier in Debian.
Lucas Nussbaum
What many people don't understand about Linux development
is that it's truly a team effort:
Red Hat develops the kernel,
Novell develops the applications,
Debian does the packaging,
and Ubuntu takes the credit!
Joke found on a bathroom stall at LinuxWorld Boston 2005
The next chapter discusses challenges facing the free software

community, but Debian/Ubuntu is a specific one to discuss here
because it is a case study on the software vehicle that today is the
most likely replacement for Windows and the Mac.
While the Linux community has benefited greatly from Ubuntu's
investments and focus on the deficiencies of Debian, it is not clear
why Shuttleworth needed to fork Ubuntu to improve Debian in the
first place. Hiring volunteers to work full-time is a good way to
The OS Battle
183
speed up progress, but they could have done their work inside of
Debian if Mark had told them to. In addition, there is an argument
to be made that both Ubuntu and Debian are hurt by the split.
It is widely accepted in the free software community that Ubuntu
and Debian have a special relationship. Ubuntu's website says that
Debian is the rock that Ubuntu is built upon. Given that Debian is
installed on millions of machines, has been around for 15 years, and
has 1,000 developers, this analogy is apt.
While everyone agrees that Ubuntu's hurting Debian is bad for
the free software movement, no one knows the extent of damage to
Debian. There are no accepted and published metrics that geeks can
use to help analyze the problem like the number of Debian users
who have switched to Ubuntu.5 The Debian developer community is
growing linearly, while Ubuntu and other free software efforts are
growing exponentially.6 There certainly must have been a slowdown
at Debian around the time of Ubuntu's creation in early 2004.
We know the changes in Ubuntu could have have been achieved
by Debian because Shuttleworth hired ten of the best Debian volunteer developers, and they started work in a 100% Debian codebase.
Debian is very highly respected within the Linux community, and the
pre-Ubuntu consensus was that it was just missing a little polish and
dynamism. This could have been easily fixed, especially with the
shot in the arm of a few volunteers transitioning to full-time developers. Other computer companies have done their work directly in
the Debian codebase, and Shuttleworth has never given justification
as to why he couldn't adopt a similar strategy.7
Geoffrey Moore's recent book Dealing with Darwin talks about
companies getting eaten up by context; irrelevant things not
core or important to the business. With Ubuntu, Shuttleworth cre5
6
7
Other good ones are the rate of growth of Debian users versus other, non-Ubuntu,
distros. Also useful is the number of Debian developer-hours per person per week.
Based on a conversation with former Debian leader Anthony Towns, who said that
the number of developers joining Debian has been constant over the last few
years.
One of the biggest challenges would be for Debian to have two release cycles, one
every six months, and one when it is ready, which is Debian's current modus
operandi. This is non-trivial, but doable.
Former Debian leader Martin Michlmayr argues in his PhD thesis that Debian
should switch to time-based releases. Debian believes they have, but it is an 18month release cycle, and they still allow themselves to slip. I think a yearly
release perhaps on Debian's birthday would be a good thing.
Wider use of Debian Testing would be another possibility. Debian Testing contains
the latest tested versions of all the applications all the time. New versions of the
applications are pushed to Testing after sitting in the Unstable branch for a few
days without any major bugs being found. The package manager even supports a
way to install older versions of packages, so you are never stuck.
184
The OS Battle
ated yet another bug-tracking system, source control software,

wikis, forums, and in general invested in a lot of infrastructure
already in place at Debian. None of that work made Linux any more
ready for human beings. Perhaps one-half of Shuttleworth's early
team focused on this non-core work. Even assuming the Debian
infrastructure needed work to make it ready for human beings, it is
cheaper and better to improve an existing system than to build a
new one just to add a few features.
New Linux users are joining the Ubuntu team and contributing to
the Ubuntu codebase and community because that is the one set up
for them. By analogy, it's as if someone took Wikipedia, made a few
small changes, and this became the website people used rather than
Wikipedia itself. If the changes were so good, why were they not
made a part of Wikipedia, leveraging Wikipedia's expertise and processes? Putting the changes directly into Wikipedia would also be
better because it would improve Wikipedia, which is what everyone
is already using. Forking a codebase is primarily social engineering,
with widespread impacts, and goes against the spirit of cooperation
inherent in science and free software. In the Linux kernel, the good
ideas are incorporated and the bad ones weeded out, but this is all
done within the context of one codebase and team. Anyone who
wants to improve on Linus's work can just e-mail him some code
changes, there is no reason to create a separate, competing kernel.
There is evidence of the inefficiency in the separate organizations. Today, Debian developers and Ubuntu developers are working
mostly on separate tracks. One of the best practices I learned at
Microsoft was to give a developer full responsibility for a feature.
The person adding the footnote code to Word would be responsible
for the changes to the user interface, the file format, and the layout
engine. This meant collaborating with other developers, but it also
made just one person the feature expert, enabling him to make all
decisions with a holistic view. Most importantly, it is efficient for this
one person to fix any feature bug.
The hard part about doing something is learning how to do it.
Making a gourmet dinner, landing the Space Shuttle, or performing
open heart surgery is a few hours' work if you know what you're
doing, but it is time-consuming and nerve-racking if you do not.
When a Ubuntu developer adds a feature, he designs and implements a feature in Ubuntu, but not in the Debian codebase. Now, the
only person who understands the change is the Ubuntu developer
The OS Battle
185
who made it. Ubuntu publishes its source code on a website, but if a
Debian developer grabs it, and runs into problems, he is not an
expert in this code yet because it was the Ubuntu developer who
first made the change. Therefore, he will need to spend time getting
up to speed.
The time to get up to speed is comparable to the time to do the
work in the first place. In fact, the Debian developer who integrated
a huge set of X.Org patches from Ubuntu told me that they were
just a starting point, unsurprisingly providing little more help than
if he had done the work from scratch. I believe that if Shuttleworth
understood this concept, he would not have created a separate
Ubuntu.
If a different codebase had never been created and all the Debian
and Ubuntu developers were working in the same one, they would
automatically work more efficiently. They wouldn't need to redo, and
therefore re-learn, what someone else has just done. This would
enforce a division of labor and would increase the pace of progress.
Having separate teams is inefficient, but it also hurts Ubuntu's
quality. Whenever a Debian developer is re-learning about a software change first made in Ubuntu, he isn't using that time to move
forward on new things.
Furthermore, Ubuntu isn't the beneficiary of Debian's greater
expertise, which means their code is buggier than it could and
should be. Debian and Ubuntu's buglist is one of the best metrics
today for the set of obstacles preventing world domination. Ubuntu's
user base has grown dramatically, but their small and young team
has shown no ability to keep up with the new issues that have come
piling in along with the new users. In May 2006, Ubuntu had 10,000
active bugs, and in February 2008, Ubuntu had 40,000.
I discuss more about the challenge of bugs in the next chapter,
but the fact that Ubuntu has so many bugs means that there are
unsatisfied users, and this is stunting Ubuntu's growth. Debian's
much larger and more experienced team could provide great assistance, but because the team's release cycles and bug list are separate, there is no unified effort being made to resolve this challenge.
Because the Ubuntu team is smaller than the Debian team, they
argue that they are too busy to take ownership of their work inside
Debian. This idea is flawed because if someone else is redoing your
work, then you aren't actually accomplishing anything. Ya can't
change the laws of physics, Captain Kirk! If you are not accom-
186
The OS Battle
plishing anything, it doesn't matter how busy you think you are.
Smaller organizations should actually be more sensitive to wasted
work because they have fewer employees.
Shuttleworth claims that Ubuntu and Debian are going after different markets, but he can give no examples of features Ubuntu
wants that Debian doesn't want. If you consider the areas in which
Ubuntu has already made engineering investments: simple menus,
3-D graphics, faster startup, and educational software, it should be
obvious that Debian wants all of these features as well.8
Many of the features that Ubuntu has added, like better suspend
and resume for laptops, Debian is no longer motivated to add
because almost any Debian user who wanted this feature is now
using Ubuntu. Even if Debian does the work, they might not find the
bugs because it doesn't have that many users testing out the feature. Debian is being consigned to servers and embedded, which has
always been their strength, however, these are areas now being targeted by Ubuntu.
In a recent blog post, Shuttleworth wrote that he admires
Debian's goal of building a universal operating system, but he also
said in the same post that he believes its objectives are unrealistic.
Mark should trust his idealistic side and realize that because software is infinitely malleable, all of his software innovations can be
put directly into Debian. There are strongly unified teams building
Wikipedia and the Linux kernel, and their success stories can be
applied here.
There is understandably a fair amount of bitterness around,
which itself decreases the morale and productivity of the community. Debian has spent over a decade doing foundational work, but
Ubuntu has made just a few improvements and grabbed all the
excitement and new volunteers. I believe Debian has been terminally damaged by the split.
A separate user community is inefficient, but this is dwarfed by
the inefficiency of the separate developer community. The greatest
long-term threat to Debian is that they stop accumulating institutional knowledge. The best way to prevent this is to encourage
Ubuntu users to join the Debian community as well. Debian is filled
8
Some argue that supporting as many processor platforms as Debian does is more
work than supporting the 3 that Ubuntu supports, but there is very little architecture-specific code in Debian most of it lives in the Linux kernel and the C compiler. Additionally, Debian has platform maintainers, who are constantly watching
if anything breaks. Like with many things, Debian already has the infrastructure
and is already doing the work.
The OS Battle
187
with many experienced programmers and is a great place to receive

mentoring. And because changes are automatically propagated over
to Ubuntu, work in Debian helps Ubuntu.9
I was honored to be a speaker and discuss some of these issues at
the annual Debian developer conference in June 2007 in Scotland. A
great thing about the free software community is how you can meet
and talk with everyone. However, because Ubuntu is one of the most
successful Linux distros and has been around for several years, the
status quo is accepted. People who only look at the success of
Ubuntu are ignoring the opportunity cost of doing things in a better
way.
If Ubuntu and Debian were to combine their resources, it would
eliminate animosity and the end result would be more innovative
and reliable. This organization would be in a very good position to
replace Windows. The first free software distribution with a community of 10,000 developers wins.
One Linux Distro?

A free society is about voluntary communities cooperating
through the division of labor.
John Stossel
One of the big questions in the free software community is

whether there should be just one Linux distribution tasked with taking on Windows. In theory, Windows provided a standard platform
for developers. Windows '95 had a look and feel little different from
Windows NT, and the differences were minor from a developer's perspective, and were mostly bugs.10
One distribution is an efficiency gain over many distributions, but
the cost isn't as high with Linux. Most free software component's
greatest dependency is on the kernel, compiler and other low-level
tools, and these are the same across Linux distributions.11
There is a gaming team that recently decided to do all their work in Debian and
just let the changes flow downstream. If all patches flowed in both directions, as
everybody claims to want, and Debian and Ubuntu shipped on the same day, how
would someone decide which distro to install?
10 The Unicode support in the US release of Windows '95 was minimal and buggy.
Every non-trivial Windows application queries the version of the operating system
to do different behavior.
11 If a distribution finds a bug in the Linux kernel, it will put a bug into the bug database. The fix usually goes into the next release of the software, but a distro can
backport into their current version. The difference between the kernel in different
distros is the version and the set of backports, which are usually not noticeable to
applications.
188
The OS Battle
The job of installing a software component and working through

any compatibility issues is the responsibility of the Linux distribution. Because the cost for each additional distro isn't paid for by the
component developer, it doesn't slow progress. In fact, the cost of
today's many Linux distributions is relatively small in the grand
scheme of this million-man movement.
The best way to increase efficiency is to increase the amount of
code sharing between the distros. A good example of such an effort
is FreeDesktop.org:
The OS Battle
189
FreeDesktop.org Projects
Avahi is a multicast dns network service discovery
library
cairo is a vector graphics library with cross-device
output support.
CJK-Unifonts open source CJK unicode truetype fonts
with additional support for Minnan and Hakka
languages.
Clipart is an open source clipart repository.
D-Bus is a message bus system.
Desktop VFS is a Virtual File System aimed at message
loop (gui) applications.
desktop-file-utils contains command line utilities for
working with desktop entries and .menu files
DRI is a framework for allowing direct access to
graphics hardware in a safe and efficient manner.
Enchant is a new cross-platform abstract layer to
spellchecking.
Enlightenment is a desktop environment and
application toolkit suite with lots of pretty pixels.
Eventuality is an "application automation meets cron"
type DBUS based framework for creating a means to
schedule arbitrary "actions" performed by conforming
apps.
Fontconfig is a library for configuring and customizing
font access.
GNU FriBidi is a library implementing the Unicode
Bidirectional Algorithm and Arabic Joining/Shaping.
Galago is a desktop-neutral presence system.
glitz is an OpenGL 2D graphics library and a backend
for gl output in cairo.
GStreamer is a streaming media framework.
GTK-Qt Theme Engine is a project to unify the GTK and
Qt theming engines.
HAL is a specification and an implementation of a
hardware abstraction layer.
HarfBuzz is the common OpenType Layout engine
shared by Pango, Qt, and possibly others.
Hieroglyph is a PostScript rendering library.
icon-slicer is a utility for generating icon themes and
libXcursor cursor themes.
icon-theme contains the standard and also references
the default icon theme called hicolor.
IMBUS is a common tier-1 architecture of IM
frameworks for connecting input method engine
containers and client application libraries.
immodule for Qt is a modular, extensible input method
subsystem for Qt.
IPCF is an inter-personal communication framework.
LDTP - Linux Desktop Testing Project
libburn is an open source library suite for reading,
mastering and writing optical discs.
libmimetype is a simple implementation accessing the
shared-mime-database included in PCManFM, a
lightweight graphical file manager featuring speed,
low resource usage, and tabbed-browsing. This small
GPL'd lib can be used for mime-type handling as a
lightweight replacement of xdgmime.
liboil is a library that makes it easier to develop and
maintain code written for MMX/SSE/Altivec extensions.
libopenraw is an open source library for Camera RAW

file decoding and processing.
libspectre is a small library for rendering Postscript
documents.
Mesa The Mesa 3D Graphics Library, an
implementation of OpenGL.
OpenSync is a project to create a standardized
synchronization framework.
Oyranos is a cross platform colour management
system.
poppler is a PDF rendering library, forked from xpdf.
Portland provides a set of high level desktopintegration APIs
SCIM Smart Common Input Method platform, is a
platform to develop input method services.
Scratchbox2 is a cross-compilation tool.
shared-mime-info is a package containing a large
number of common MIME types, created by converting
the existing KDE and GNOME databases to the new
format and merging them together, and software for
updating the database based on the share-mime-info
specification.
startup-notification is a sample implementation of
startup notification (telling the desktop environment
when an app is done starting up).
Tracker is a highly memory efficient file indexer and
metadata harvester.
uim is a library to support input many languages.
UTF-8 is a project to document and evangelize the use
of UTF-8 encodings for open source projects.
unicode-translation aims to translate Unicode
character names and other data into many languages.
vaAPI provides a decode only video acceleration API
for all video formats. Currently in proposal stage.
waimea aims to be a standards compliant window
manager for the X Window System making use of the
cairo graphics library for all rendering.
xdg-utils is a set of command line utilities to simplify
integration with a Free Desktop. It has simple
functions for creating menus, opening files, setting
mime types, and so on. It is part of the Portland
project.
xdg-user-dirs is a tool to handle well known directories
in the users homedir
Xft is a library for client-side rendering of fonts.
xfullscreen is a useful module for applications or
window managers supporting fullscreen modes.
Xgl is an X server architecture layered on top of
OpenGL.
Xorg is the XOrg Foundation's Public Implementation
of the X Window System.
xkeyboard-config is a central project for keyboard
configuration.
xprint is the X11 printing system.
xsettings is a reference implementation.
X Testing provides information on various software for
testing X Servers and Clients.
X Window information is a window information utility
for developers of applications, toolkits, and window
managers.
When Linux standardizes font rasterizers and such, unification

automatically occurs.
If we moved towards a truly universal PC operating system, what
would happen to the long tail of Linux distributions? Eventually, in a
biological system, the biggest and best gain momentum and squeeze
out the others. However, we are still in the early years of Linux, and
so the natural selection process hasn't yet happened. Today every
190
The OS Battle
Linux distribution team is growing when you have a worldwide

market of only 3%, it is a target-rich environment. In addition, you
can have subsets and other derivatives.
Apple
After Woz hooked his haywire rig up to the living-room TV, he
turned it on, and there on the screen I saw a crude Breakout
game in full color! Now I was really amazed. This was much
better than the crude color graphics from the Cromemco Dazzler. ... How do you like that? said Jobs, smiling. We're going
to dump the Apple I and only work on the Apple II. Steve, I
said, if you do that you will never sell another computer. You
promised BASIC for the Apple I, and most dealers haven't sold
the boards they bought from you. If you come out with an
improved Model II they will be stuck. Put it on the back burner
until you deliver on your promises.
Stan Velt, former Editor-in-chief, Computer Shopper
Apple's iPod and iPhone may be sexy and profitable, but these
small devices are specialized in function, so there isn't a lot to say
about them. An iPod is busy when playing music, whereas when
your computer plays music, it uses less than 1% of its computing
power, which is not even noticeable.
For most of Apple's existence, they never really got the idea of the
relationship between market share and a developer community. For
example, Macs have historically not been allowed in enterprises
because no one added the necessary features and applications
because it never got the requisite market share to make anyone
want to bother.
Microsoft understood the virtuous cycle between users and developers, and knew that making it easy to build applications would
make Microsoft's ecosystems successful. Bill Gates brags that
Microsoft has ten times as many partners as Apple, and tools like
Visual Basic and FrontPage were important reasons why.
This internal focus that has limited the Mac's potential marketshare is now playing itself out with their new devices. Symptoms of
this mindset are noticeable in the most basic scenarios: you cannot
drag and drop music on and off an iPod, as you can with a digital
camera. Even if you could copy over your files, unless it is in one of
the few formats Apple can be bothered to support, you would still
not be able to play it.
While Moore's law will push these new devices further up the
computing value chain, it isn't clear Steve Jobs understands the
value of having a developer community extend his platforms
The OS Battle
191
because he doesn't see past the potential complexity it creates. In a

New York Times interview, Jobs said he would not allow third-party
technology contributions to his new iPhone because he doesn't want
the telephone network brought down by rogue applications.12
Apple later reversed-course and now there are now third-party
apps available for the iPhone, and Apple even runs ads touting a
capability they initially shunned, but Apple's mindset of ignoring the
outside world is still embedded into its culture. When the iPhone
software development kit (SDK) was launched, third-party iPhone
developers were: legally banned from sharing programming tips,
discussing code or asking questions of one another in forums or
over e-mail. Note that an SDK only contains the public information
necessary for a developer to write an app, so why they would be so
restrictive is inexplicable.
After lots of complaints by developers who were unable to get
their code working, and yet were not able to talk to each other,
Apple in October 2008 reversed course here as well, so maybe they
are slowly learning that not all good ideas originate in Cupertino.
One of the reasons Microsoft has won so many battles is that it
knows the key to success is to just get early market share and let
Metcalfe's law take over, with the ensuing profits. Steve Jobs is presumably satisfied with Macintosh's 3.5% worldwide marketshare
and the state of the iPod today. It is this attitude that ultimately will
ensure their failure, I believe. Apple's motto for a time was Think
Different but a more apt one might be Think Small.
IPod and Digital Music

In stumbling upon the business of building fashionable music
players, Apple has touched upon one of the important questions of
the digital age: How do we acquire and archive our music?13 If we
12 It should not be possible for software to take down a network. There are also
many ways to sandbox applications to limit their ability to do things.
13 Another important question that is not close to being answered: How does my
music device plug into my car? A device should not expose itself as a set of files to
your car radio because that requires your car to understand how to decode all the
various formats. Even if the car does understand MP3, it won't understand audio
bookmarks. My car radio wants to fast-forward through hour-long podcasts a couple of seconds at a time, whereas the RockBox OS on my iPod gets increasingly
faster and is configurable in this regard. What we likely need is a streaming protocol: the iPod will implement functions like fast-forward, next song, show directory, search, etc.
A question: since you can understand voice that's 50% faster, should that playback setting be available in the car radio UI, or be a setting of audio files? Suppose you wanted to override that behavior, would you be able to do that through
the UI of the radio?
192
The OS Battle
do it right, we can store our music in one format for decades, even
forever.
A digital format is something not tied to a hardware medium in
the way that the VHS format was tied to VHS tapes; everything digital can be copied to a hard drive or a USB key. (Iraqi terrorist Zarqawi's death was a big setback for al-Qaeda because we recovered a
USB key containing his terrorist documents and music. One day, the
only thing we will ever need to waterboard terrorists for is their
passwords.)
We have been able to create basically one format for digital cameras JPEG, which is free and efficient. We should have been able
to do this for audio as well because the underlying math is similar!
Apple made the digital audio format problem worse by endorsing
one that only they use, AAC. And, by adding DRM, they only allow
you to play your music on their one device and in their one application. Apple has added hassles and created doubt about whether you
will ever control, and therefore truly own, your music.
Steve Jobs is ecstatic that iTunes has sold 2.5 billion songs in five
years, but when you consider that the music business is a $40-billion-dollar per-year industry, and Apple has no serious competitors
in digital music, that number is modest.
I've met many people who have told me that they won't buy an
iPod again because of these and other issues. In fact, I run an alternative OS on my iPod, called Rockbox.14 It supports more audio formats, lets me copy files back and forth like you can with a digital
camera, it even sounds better because of a feature known as crossfeed. (When you listen to music on room speakers, each ear can
hear music, slightly delayed, from both the left and right channel.
However, this does not happen with headphones. Rockbox uses
clever algorithms to simulate the sound from the opposite channel,
slightly delayed, to make headphones sound more natural.)
Rockbox also comes with more software and other advantages. It
is perfectly logical that the free software community can do a better
job than Apple because Apple likely had 20 software developers
writing code for the iPod device, whereas Rockbox has had more
than 300 contributors, and itself reuses a lot of other free code.15
14 iPodLinux doesn't quite work yet. The installer repeatedly crashed, after installation I couldn't find a way to play music, I had to reboot into Apple's iPod OS to
copy music over, etc. The momentum is with Rockbox.
15 My biggest complaint with Rockbox is that they haven't standardized the back
button.
The OS Battle
193
Even worse for iPod's future, I suspect that Apple scavenged some
of the best people from their iPod team when staffing up iPhone
Microsoft would have done so.
Apple is now reversing course and providing more music DRMfree, but they are still making it difficult to put music onto your
device, not letting other devices play the songs in your music library,
and are now preventing the installation of third-party software on
their new hardware. It is only because Apple has such little overall
marketshare that they can get away with this sort of behavior.
Mac OS X Kernel
Apple giving about as much attention to the iPod as the Mac is like Ford
focusing their R&D on bling because they suddenly started making half of
their profits on chrome rims.
Compared to the task of decoding compressed music, the word

processor I used to compose this book is 1,000 times more complicated. So, what about Apple's computers? The first relevant fact is
that Apple is making half of its profits in the (currently) profitable
portable device market which has caused Apple to lose interest in
the computer as the most powerful tool ever created.
Apple's computers have an elegant hardware design, but the biggest difference between a Mac and a PC is that Apple's computers
run a completely different pyramid of software, that is mostly proprietary like Microsoft's.
From its introduction in 1984 till 2001, the Macintosh ran a kernel built by Apple for its first Macintosh. It was written in Motorola's
assembly language, and didn't have memory protection, preemptive
194
The OS Battle
multitasking, and many other features of a kernel you would read

about in a college textbook. In the late 1990s, this caught up with
Apple as Macintoshes were considered unreliable.
Apple eventually threw away their original kernel in creating Mac
OS-X 10.0, but it took them a while to get going on it as they had
four CEOs in the 1990s.
Mac OS 9, the last release of the original Apple kernel and the official version until 2001.
Mac OS-X, Macintosh's first release based on a kernel that is free software.
The OS Battle
195
Apple's second kernel wasn't built from scratch, but is based

heavily on Berkeley Software Distribution (BSD) Unix code.16 This
code is a lot like Linux, but with a smaller development community
and a non-copyleft license agreement. That Apple is depending on a
smaller free kernel community, and yet doing just fine, does say
something about free software's ability to deliver quality products.
Apple's new kernel is certainly much better than the one they tossed
away after 20 years of investments!
Portion of the Unix family tree. The biggest reason why Unix and Linux hasn't beaten Windows yet is that the workstation vendors didn't work together
on the software, and so kept re-implementing each other's features.
16 Apple's kernel also contains pieces of the Mach kernel, but their OS contains
much more code from BSD. In fact, I can see little to no benefit from using any
Mach code as opposed to its equivalent from BSD. Some have described Apple's
combined kernel as a software Frankenstein.
196
The OS Battle
Unfortunately, in choosing this software, Apple gave life support

to a group who should have folded their code and resources into
Linux. The free software community can withstand such inefficiency
because it is an army of millions, but from a global perspective, this
choice slowed progress and hurt Apple's customers.17
Apple didn't publicly say why it chose BSD over Linux, but at the
time it was widely surmised that one of the biggest reasons was that
Apple feared the copyleft aspect of Linux's GPL license. Apple didn't
understand that it is perfectly legal to ship proprietary code on top
of Linux.
However, Apple should have asked itself why it chose a free kernel as the foundation of its future software, but not other free code.
Apple has not embraced free software, even though it has come to
depend on it, and Apple's software total less than 10% of revenues!
Apple's hardware team should be happy to run Linux or Windows on
its machines. With Macintosh at 3.5% marketshare, PCs are 30
times more an opportunity than a threat.
Supporting other OSes is no big feat of engineering; for a period
of time, Linus did his Linux development on a Mac after discarding all of Apple's software. If Apple did decide to go aggressively
after the PC OS market, it would need to make just a few hardware
changes: their keyboard is not quite suitable, and they have stubbornly resisted the move to two-button mice. Apple already produces a multitude of keyboards for different writing scripts, so they
already understand the necessary work.
17 When Apple started shipping computers with four and more processors, BSD
Unix's performance was worse than Linux's, and in 2003, BSD added scalability
optimizations that Linux had added four years earlier. Linux kernel guru Greg
Kroah-Hartman has said that Linux runs faster inside a virtual machine than the
BSD-based Mac OS hosting it.
The OS Battle
197
Software
Nobody has ever had more contempt for customers than Steve
Jobs.
Eben Moglen
In a better iWorld, Apple would have used Linux instead of BSD.
With less of an internal focus, Apple could even have done a few
things to make their OS more compatible with Windows, to increase
sales and to better tempt Windows users into switching. There is
plenty of free software out there to enable interoperation with Windows technologies. Out of the box, Windows Media is treated by the
Mac like a text file; there is code out there to fix this, but Apple
doesn't provide it. It's as if supporting Windows Media is a concession that weakens Apple. We aren't even talking about having musiccreating software support that format simply the ability to play
these files!
Their iChat program doesn't support MSN messenger, though it
does support AOL and Yahoo. Given the unnecessary hurdles Apple
has created for Windows users, it is not surprising that their market
share remains so low. There is even a free Win32 implementation
known as WINE that would allow Windows software to run on the
Mac; yet another way to storm Microsoft's beachfront that Apple
hasn't adopted.
198
The OS Battle
Apple creates proprietary software because they use it to bundle

with their hardware, and to lock you into their platform. However,
the threat of free software is already approaching Apple. The Linux
community is currently focused on building the Web and Office,
the minimum necessary to build a platform that competes with Windows. Once these are covered and more people start using Linux,
then Linux contributors can focus on building video-editing software, and other features that Apple has leveraged to carve out its
niche. In fact, supporting the few relevant Apple standards like
QuickTime is a much simpler problem than supporting Microsoft's
many complicated and popular standards.18
When I visit coffee shops, I increasingly notice students and computer geeks purchasing Macs. Students have limited budgets and so
should gravitate towards free software. If Apple doesn't support free
software, their position in the educational market is threatened.
Many computer geeks buy a Mac because of its Unix foundation.
For example, in the terminal window of both the Mac and Linux, if
you type ps -a, you see the list of processes. (Windows doesn't
support the rich Unix command-line world except via an add-in that
is little used by Microsoft or third-party developers.) Apple has good
Unix compatibility only because their programmers never removed
it. It was never a goal of the Mac OS-X to appeal to geeks Apple
just got lucky. Smartly, Apple now mentions Unix compatibility in
their marketing material.
As a long-time Windows user, and a Linux convert, I tried out the
Mac OS X for a couple of days. Here are some of my impressions:
A Mac OS has more code than ever before, and a lot of it is

based on free code, but it doesn't have a repository with
thousands of applications like Linux. There are several thirdparty efforts to provide this feature, but none are blessed or
supported by Apple. The Mac comes free with iPhoto, but
they really want me to buy Aperture for $159, which they tell
me just added 100 new features! Apple ships a new OS every
year, but you don't get free upgrades it costs $140 to
upgrade from OS X 10.4 to 10.5.
Many of the Mac's UI details like how to maximize windows,
and shortcut keys, are dis-similar from Windows. Linux, by
contrast, feels much more natural to a Windows user.
18 WINE is one of the biggest pieces of Windows compatibility code out there, and it
alone is much bigger than all the Apple compatibility code one would need to convert someone from a Mac to Linux.
The OS Battle
199
Another example: when you double-click on a picture, it loads

the iPreview application that stays around even after the window displaying the picture is closed. How about just creating
a window, putting the picture in that window, and having it
all disappear when I close the window? I tried to change the
shortcuts to match the Windows keystrokes, but their software contains bugs because it didn't change it in all applications.
The Mac feels like a lot of disparate pieces bolted together.
The desktop widgets code has its own UI, and it doesn't integrate well into the OS desktop. The Spaces is a clone of an
old Unix feature and doesn't implement it as well as Linux
does. (For example, there is no easily discoverable way to
move applications between spaces.)
As mentioned above, the Mac doesn't support as many of the
Microsoft standards as Linux does. One of the most obvious
is WMA, but it also doesn't ship with any software that reads
and writes DOC files, even though there is OpenOffice.org
and other free software out there.
It is less customizable. I cannot find a way to have the computer not go to sleep when the laptop screen is closed. The
mouse speed seems too slow and you can only adjust the
amount of acceleration, not the sensitivity. You neither resize
the system menu bar, nor add applets like you can with Linux's Gnome.
Having used all three, it is my view that Linux is better than the
Macintosh, which is better than Windows Vista. I hear that internally, Microsoft is very afraid of the Macintosh, but Linux presents a
greater long-term threat.
Apple has been good for Linux because it has forced software
vendors to think about writing cross-platform code. By using free
software like the BSD Kernel, Apple is more of a part of the free
software community than before, although not by much. Apple's
web browser Safari is based on the engine WebKit, which Apple
derived from the free KHTML, but they worked for a year on the
fork before attempting to integrate their changes back, and KHTML
developers said that their relationship with Apple was a bitter failure.
Apple does use Apache, and oversees the printing system CUPS,
but they have proprietary office-type applications, and a variety of
200
The OS Battle
multimedia and other tools. They are neither using a fraction of the
free software they could, nor releasing much of their code as free
software.
A Free Macintosh OS?
Can Apple's software stand alone?
In general, other free software should be looked at as an opportunity for Apple. But what about making all of Apple's software free? If
Apple's software were made free, Apple could work more closely
with the global free software community and create a better product
for their customers. However, the downside is that this software
could get ported to Windows and Linux, and create less reasons to
purchase a Mac.
However, if a Linux desktop takes off, Apple's OS will suffer the
same fate as Windows and force Apple into being a hardware-only
company like Dell and the rest, all the value they have built up will
be gone. Once people are satisfied using free software, they don't
usually go back to proprietary software. For example, free code to
convert from proprietary formats has much more demand than code
to convert to them.
If Apple's free software is used on other OSes and hardware, they
risk becoming a company competing primarily on hardware features, but they have that risk already, and at least would be in control of their own destiny. Apple could be leading the free software
movement! In addition, there are many ways to monetize users of
The OS Battle
201
Apple's free software who are running it on other hardware platforms. It seems unlikely that Apple's free software would be widely
used, but not their hardware.
Windows Vista
Screenshot of Windows Vista. It is five years of evolution beyond Windows

XP, but has many of the same limitations. Whether it will still be prone to
viruses and malware is one of the biggest questions.
Microsoft's Windows Vista was released in December, 2006. The

first major release of Windows since Windows XP in 2001, it is an
evolutionary step forward in many ways. Every component was
improved with new features; a glassy 3-D user interface, integrated
search, improved kernel and a new version of Internet Explorer.
A new IBM Thinkpad computer came with Vista Premium and I
spent several hours with it. Here are my impressions:
When I turned on the computer for the first time, it spent five
minutes checking its performance before I could do anything;
it's rather ironic that this task is slowing down my computer.
Internet Explorer provided two toolbar textboxes to search
the web, both using MSN's search. When I create a new tab
using the shortcut key after I launch IE, it would say it could
202
The OS Battle
not connect, even though the computer had long-ago established an Internet connection. Apparently, the typing part of
IE was initialized before the networking part.
Windows Update wanted to install an Office 2003 service
pack even though I could find no Office 2003 installed. Office
2007 was installed as demoware, even though I chose not to
purchase it.
There was a tool to resize the hard drive to make room for
Linux, but it insisted on reserving 44 gigs of free space for
Vista!
The disk defragmentation tool said my hard drive contained
3055 fragmented files and 13,265 excess fragments and
that the drive was heavily fragmented. I guess when Vista
checked the computer's performance, it didn't notice the
filesystem fragmentation issues. The defrag tool was nagware and kept encouraging me to purchase a better version.
(In Linux, the file systems have much less fragmentation
because the code is smarter.19)
I found seven copies of system files like i8042prt.sys.
I wanted to get Windows Media to play my OGG audio files. I
told Windows Media I want to install plugins and it took me
to a Microsoft-maintained web page with links to 3rd party
plugins. I find one and install it. Once installed, I cannot see
the plugin in the media player's list of installed plugins.
I downloaded an OGG file in Internet Explorer, and Windows
Media finds and plays the file, but I can't see where it was
put. I use the search feature to look for it, but search doesn't
find it. I try to refresh the index, and it warns me that performance will be slow because it is a background task, which is
true because it doesn't even attempt to rebuild it.
I checked on the Windows Media settings for ripping music.
The default is to rip into 128 kbps WMA, which has never
been good enough for archival purposes. It doesn't allow me
to rip into anything higher than 192 kbps. I then discover
that there are four types of Windows Media, including an
option to use variable bit rate. (VBR encoding uses more or
less bits depending on how much the music needs at that
moment to maintain a certain level of quality.) The default is
19 My ext3 user partition is 67% full and has 4% non-contiguous inodes. My system
partition is 52% full and only 1.6%. You would expect the system partition, where
not much writing happens, to not be very fragmented. In Windows, it was the system files that were fragmented because it had no user files on the computer yet.
The OS Battle
203
inexplicably not to use that clever feature. Windows Media

allows me to rip into MP3, but I can only set four levels of
quality, and I can't change any other MP3 encoding options
like I can on Linux. (Microsoft got sued for billions of dollars
for minimal MP3 support.)
There are now three status icons for wireless Internet, one
more than in Windows XP. Let no one say that Vista is not an
upgrade: Now with more 802.11 status icons!
After Vista Service Pack 1 is released, I go to Windows
Update to try to download it. It isn't there, and shows me no
new updates. I log-off, and it tells me it is installing a bunch
of updates ones that it just told me it didn't have.
I have pages of notes like this from just a few hours with the OS.
A friend bought Vista on a new HP computer, and while he could see
the Internet, he couldn't really transfer data because the Norton
anti-virus nagware that came with the computer was incompatible
with the Vista firewall. This former MS programmer needed to call
HP to get help because the symptom of sort-of Internet access didn't lend itself to any obvious solutions. Removing the nagware
solved the problem, but that process took 40 minutes building a
high-performance uninstallation program on Windows is apparently
nontrivial, even for a company with their act together enough to
build a virus scanner.
While Vista is the next generation of Windows XP, it still has the
same fundamental limitations. It will have driver reliability issues
because of its backward compatibility constraints and complicated
code. It will become obsolete because new hardware is released
continuously, but Microsoft puts products into a maintenance mode
after they ship, and places most engineering resources onto the
future source of revenue.20
20 Microsoft isn't able to just give you a new kernel with updated hardware support
without giving you a whole new OS. Driver writers building code for the next OS
don't want to support the old OS in addition, especially if that next version of Windows is just around the corner.
Microsoft has to threaten and plead with hardware vendors to get them to do all
the work that MS creates for them. In order to get the 'Designed for Windows
2008 Logo' you need to do this. I heard a Microsoft Windows Server evangelist
say that their 32-bit operating system was going to be retired after Vista, which
was why everyone should just go ahead and start work on the 64-bit drivers right
now this coming from a company that shipped remnants of 16-bit DOS and Windows 1.0 in Windows Me. A 64-bit Windows simply requires that all hardware vendors recompile their drivers, but because there is no unified tree, this is a hard
problem involving a lot of coaxing.
204
The OS Battle
Beyond the reliability and interoperability issues, Vista's biggest

limitation is that it doesn't contain any major functionality beyond
what we all think of as Windows today. If you want an audio creation
tool or educational software, Windows is not the answer.
Like the Linux kernel itself, a Linux distribution has all of its code
in one big tree. This means that upon installation of a new version,
all the applications work out of the box because they have been built
and tested together. A distribution can fix any bug in any of its code
because it has total access to it all. By contrast, Russ Cooper, a
senior information security analyst at Cybertrust, has said that:
I'm going to Vista...when my VPN supplier tells me that they
have drivers that work, and when my anti-virus vendor tells me
that they have non-beta versions that work.
Microsoft has to spend much more on backward compatibility

than Linux does, and the huge cost has imperfect results because
Microsoft doesn't have the ability to fix code in third-party applications. In a free software world, bugs can be fixed anywhere, and
therefore in the proper place.
In addition to the larger variety of applications in Linux, you find
more choices all throughout the free software stack. The user interface is more customizable than Windows or the Mac, and there are
communities around even the tiniest parts of a Linux operating system, like themes: every Linux desktop can be personalized to a
much greater degree. I can even download a theme to make my
computer have a glassy user interface like Vista, or a user interface
like the Mac, just two of many choices.
A new version of Windows used to mean a lot because new hardware and software was made a part of the platform. However,
because of the dearth and death of third-party Windows developers,
and because most developers are building applications on the Internet, a new version of Windows matters a lot less these days.
Furthermore, because deployments of Windows take years, developers are left in a quandary: If I write code that depends on a new
Windows feature, what happens if my code is running on an old version of Windows? Free software propagates with less friction,
therefore developers can more easily take advantage of new features, and assume their existence. Someone remarked to me that
Vista is Different but not different. They've tweaked the icons and
moved things around, but the same functionality is there. I remember the OS wars of the past, and it seems like we are in for another
one. Grab popcorn!
Challenges for Free Software
205
CHALLENGES FOR
FREE SOFTWARE
Bill doesn't really want to review your spec, a colleague told
me. He just wants to make sure you've got it under control.
His standard MO is to ask harder and harder questions until
you admit that you don't know, and then he can yell at you for
being unprepared. Nobody was really sure what happens if you
answer the hardest question he can come up with, because it's
never happened before.
Watching nonprogrammers trying to run software companies is
like watching someone who doesn't know how to surf trying to
surf. Even if he has great advisers standing on the shore telling
him what to do, he still falls off the board again and again. The
cult of the MBA likes to believe that you can run organizations
that do things that you don't understand. But often, you can't.
Joel Spolsky
The mode by which the inevitable is reached is effort.
Felix Frankfurter, US Supreme Court Justice
ree software has been around since 1985, and yet has only
1% marketshare on the desktop today. Free software has
tremendous potential, but the community needs to execute
better to win. In fact, until free software succeeds on the desktop,
its last and biggest challenge, many will continue to question
whether it is even viable.
As a side note, while Microsoft doesnt have a great reputation
right now, especially because of Vista and malware, there is a lot of
clever code and brilliant engineers inside the company. I was fortunate to learn from a bunch of great people in groups that fostered
cultures of very high quality software.
When Microsoft got serious about Internet Explorer, Netscapes
rag tag team of kids just out of college didnt have a chance. MS
took some of their top engineers in text engines, networking, forms,
internationalization, performance, and object models, and put
together a large, world-class team. Microsofts institutional knowledge of so many areas of software could be applied to any effort.
People forget all that, but this is why the reviewers consistently
would say that Windows was better than OS/2, Microsoft Word was
easier to use WordPerfect, IE was faster than Netscape, Excel was
richer than Lotus 1-2-3 3, ad infinitum. PC Magazine wrote in 1997:
206
Microsofts Java environment was the fastest and most compatible

on our tests. Microsoft had built a better runtime than Sun, the creator of the language!
Many think Microsoft won all these battles merely because they
somehow magically became a monopoly, but the atriums in the
buildings at Microsoft have shelves overflowing with awards from
independent reviewers. Microsoft succeeded in all these areas
because they used their financial success with DOS to hire an enormous army of smart programmers working in all areas of software,
which could be applied to new areas, and passed on to new employees like me. And every time I switched groups, Id take the knowledge and best practices I had learned in my previous groups.
I am not an apologist for Microsoft, and I think they are in deep
trouble over the long-term, but the idea that Microsoft came to
utterly dominate the computer industry and lay waste to countless
companies was not mere chance. In fact, a big part of Microsoft's
success was the strategic or technical incompetence of its competitors.
More Free Software

If you look at the magazines about the use of the GNU/Linux
system, they're filled with ads for non-free software that you
could run on top of the system. Those ads have a common message. They say: Non-free software is good for you. They call
these things value-added packages, which makes a statement
about their values. I call them freedom-subtracted packages.
Because if you have installed a free operating system, then now
you are living in the free world. You enjoy the benefits of liberty
that we worked for so many years to give you.
Richard Stallman
One of the best ways to cure the problems of free software is simply to create more of it. If you talk to someone about their Linux
experience, they might complain that they can't play DVDs out of
the box, that some proprietary drivers are missing, or that iTunes or
other proprietary software does not natively run on Linux.
Getting vendors to provide their proprietary software on Linux
might sound like a great help, but building a completely free software stack should be the principal focus. Richard Stallman wrote:
Adding non-free software to the GNU/Linux system may
increase the popularity, if by popularity we mean the number of
people using some of GNU/Linux in combination with non-free
software. But at the same time, it implicitly encourages the
207
community to accept non-free software as a good thing, and forget the goal of freedom. It is no use driving faster if you cant
stay on the road.
Some GNU/Linux operating system distributions add proprietary packages to the basic free system, and they invite users
to consider this an advantage, rather than a step backwards
from freedom.
Richard Stallman looks at free software as a struggle requiring

constant vigilance, and he is correct. Linux is just like the real world
in that respect: the parts that don't have freedom are the worst
parts. Ultimately, it is good news if the solution to your problem is to
do more of what you're already doing as this means you are heading
in the right direction, and success is possible.
In addition, proprietary software is generally an older, decrepit
distraction. Microsoft's profitable products are simultaneously burdened by decrepit code, as is much of the industry's successful proprietary software. Adobe Photoshop is one of the most popular tools
used by graphics designers today, and one of Adobe's employees
wrote of Photoshop:
In One Piece At a Time, Johnny Cash tells the story of building a Cadillac from 20 years' worth of evolving, mismatched
parts. I've gotta say, I know the feeling. Photoshop has been
accreting power & users for the better part of two decades, and
looking at some parts of Photoshop is like counting the rings in
a tree: you can gauge when certain features arrived by the
dimensions & style of the dialog. No one wants to work with
or work on some shambling, bloated monster of a program.
I think proprietary software, if it becomes popular and long-lived,

is destined to become a mess because it does everything by itself
rather than leveraging free software components. Furthermore, it
doesn't receive the constant tending that a garden the size of a city
would require. If Adobe were to put a team of 100 developers on
Photoshop to clean things up, it would no longer be the profitable
product it is today, even with its $1,700 price tag.
Cash Donations
It is every mans obligation to put back into the world at least
the equivalent of what he takes out of it.
Albert Einstein
Drupal is a piece of free elegance to help manage the content of a

web site. Like many software projects, it has a link where you can
donate money via PayPal. While Drupal had 45,000 users in 2006, it
received only $1,000 in donations, for an average donation of 2 pen-
208
nies per person. Free software, as Eric Raymond says, is about

scratching an itch fixing a problem that you care about. But what
if you have the itch and some money but not the ability to scratch it?
The fact that much free software can flourish without money is
amazing, but free software should have financial donations as a sustaining element, in addition to time.
Bounties, a collection of a large amount of money set aside
towards a specific purpose, is another great idea that the free software community has explored but not widely adopted. If the eight
million users of Ubuntu gave their one dollar towards bounties, and
the bounties were set at $5,000 each, this would create enough
money for 1,600 projects. A major effort of Linux distributors could
be to gather proposals, take donations from interested users, and
manage the development. Doing this would motivate free software
users to donate money because it would be applied directly towards
projects users care about.
It is still early in the free software movement. The Linux kernel
has thousands of contributors, but many other teams are not as well
funded yet. An OpenOffice developer told me that this is the breakdown of developers from various companies:
Sun: 30
Novell: 15
RedHat: 1 to 2
Ubuntu: ~1
There are also RedFlag & IBM with large numbers of supposed
contributors, but sadly with a large imaginary component ;-)
The people at Sun who are responsible for OpenOffice do not realize its importance to put such few developers on it. In spite of the
fact that Sun has over 30,000 employees, Microsoft has twice as
many programmers working on Internet Explorer as Sun has working on all of OpenOffice! I wouldn't work on something as big and
complicated as OpenOffice without getting paid, which is why I
donated money.
We are used to paying hundreds of dollars for proprietary software applications, and so we should be willing to invest smaller
amounts for free software. I went on a guided tour of Mt. Saint
Helens, and at the end the Forest Service Guide said that if you
liked the tour, you should donate $5, but if you didn't, you should
donate $20 so they could improve the tour. That should be the
attitude of the free software community: if you want more, contribute more. The nice thing about this model is that people can donate
what they can afford. If you're a student or a business in India, the
209
amount you can donate back for use of a free database is much less
than if you are Amazon corporation and you have a farm of servers
running the same software.
Devices
The inside of a computer is dumb as hell but it goes like mad.
Richard Feynman, 1984 (Moore's law says this statement is
65,536 times more true today.)
There are enormous markets for free software below the PC for
which Microsoft has not built a dominant ecosystem, nor one with
strong ties to the desktop. These markets should be much easier for
free software developers to infiltrate and dominate:
Survey of Embedded/Realtime OS Usage

(Source: market strategy firm Venture Data Corp.)
7%
12%
37%
12%
Other
No formal OS
Windows
Linux
VxWorks
32%
The fragmented state of embedded operating systems. This survey also
showed that planned use of Linux is expected to increase 280% for the
next wave of projects. However, that would still only give it 33% of the
market.
The proprietary OS Symbian has 70% of the high-end cellphone

market but is not free software, and technically inferior to Linux,
and therefore should be easy to replace. The good news is that ABI
Research forecasts that by 2012, more than 127 million devices will
210
be enabled with Linux, up from 8.1 million in 2007. Many of the proprietary embedded OSes enjoy their success primarily because of
inertia.
Moore's law doesn't just apply to our computers, it will also
ensure that small new devices will proliferate like mice in springtime. And those mice will need software to make them interesting.
One device I'm looking for is a robotic mouse for my cat:
Fierce soldier of cuteness on kittenwar.com.
He's regal and mellow and while he happily enjoys lots of quality
time on da cowsche and refrigerator, his hunting instincts are prolific.1 I've purchased many toys for him to discover how he learns
and plays, but while each one has provided a few minutes to a few
hours of entertainment, none of the toys provides him the thrill and
challenge of stalking and catching live prey.
One of Davis's favorite toys is a ping-pong ball. However, while it
bounces around and can quickly disappear, it isn't alive and doesn't
1
As a bachelor who has spent too much time with computers, I've relied on Davis
to teach me lessons about how to better interact with humanoids. When I come
home at night, tired and ready to relax, my cat is just waking from his day-long
slumber and is ready to play. If I ignore him, he will persistently meow at me, letting me know he is bored and wants to be entertained; living inside an apartment
simply doesn't provide enough stimulation for him.
211
change direction under its own power. I can watch him calculate a
path to pounce on a toy, realize it isn't a challenge for his prowess,
sneer, and lose interest.
I'd like a robotic mouse that is quiet, agile and fast. Maybe when
he catches it, the mouse could open a hatch and release some food.
It would be nice to come home from work and be able to sit on the
couch and commiserate with my cat about how exhausted we both
are from earning our meal.
Better toys are in our future, and so are safer ones. My sister has
a beautiful Husky, an interesting animal because, like my sister, it
has no sense of awareness of its location. Huskies are bred for
endurance and strength and can head in any given direction for
great distances but are never able to retrace their steps. I'd like to
be able to give him a collar with GPS and a cellphone-like transmitter that could report back on his location. Anyone who has lost a pet
understands the frustration. A similar thing could also be used for
young children. A GPS-device would allow you to sleep soundly at
night knowing you can always find your vulnerable family members.
These are just two of the many innovative devices I can imagine,
and these devices require little new hardware technology beyond
what we already have at our fingertips. Building such devices would
cost hundreds of dollars today, but the relentless march of Moore's
law suggests that the price will keep halving every 18 months; this
combined with free software will make the future very interesting!
Reverse Engineering
Imagine trying to fix a car without having any of the documentation produced by the car manufacturer. Much technology in the
computing world today is undocumented, which means that a lot of
what free software developers have to do is reverse-engineering, a
tedious and time-consuming task. I don't believe the Publisher format is publicly documented, so someone would need to create lots of
little files and look through the on-disk binary to figure out the particulars of how formatting is stored on disk. It is sort of amazing that
millions of people create files in formats that are not publicly
described, yet this is the state of the industry today. Unfortunately,
this means the user is enslaved to his software vendor, and arguably
doesn't even truly own their documents.
If everyone used free standards, let alone free code, software
progress would be faster. Today, if a standard is created, the community almost never gets any code to go with it. It might be best to
212
have code be the official reference, rather than a written specification, but today we often have neither. If the computing world ever
breaks out of this vicious cycle, maybe we could focus our collective
efforts on the truly hard challenges.
PC Hardware
We love Linux, and we're doing our best to support the Linux
community. we see the Linux desktop as a customer-driven
activity. If customers want it, well, Dell will give it to them.
Michael Dell
Michael Dell's quote demonstrates that he doesn't consider the

situation where people aren't running Linux on his hardware
because it doesn't work! I have a Dell Vostro laptop that doesn't
have a Linux driver for wireless Internet. If I can't take the computer to a coffee shop and surf the web, it is useless. I thought about
putting a bullet through the laptop and mailing it back. Even laptops
by Dell that ship with Linux still contain proprietary drivers, drivers
that aren't in the kernel, and so forth so it seems clear that Dell
management doesn't much care about Linux yet. Dell also inexplicably has 30 models of laptops, so it might just be a case of corporate
cranial-rectal inversion.
Overall, PC hardware support in Linux is generally in good shape,
and has improved a lot in the years since I first started using it. I
believe it mostly requires that we continue to press on, working
through the chicken and egg issue where hardware vendors are
reluctant to support Linux until it has more users, but users won't
run Linux if it doesn't fully support their hardware.
I singled out IBM in the OS chapter, but many other hardware
companies are under-investing in free software even though public
statements by their VIPs suggest that they believe in it. For example, Intel claims to be a strong support of Linux, but is doing only a
decent job in its support of Linux drivers.2 An Intel engineer told me
at a Linux conference that their Linux efforts are just 1% of the
manpower that their Windows efforts receive. Doubling their Linux
development team would cost less than .1% of their total R&D. Intel
My Intel wireless card resets and frequently re-associates or doesn't always

resume properly. (These problems appear to now be fixed in 2009, but it took
years.) In general, Intel's video drivers are considered slow, buggy, and behind:
Intel added Linux driver support for TV-out years after the hardware was
released! There are recent news reports that Intel's GMA 500 drivers are a
bloody mess.
213
is under-investing in Linux not because they can't afford to increase

costs by .1%, but because they're suffering from Stockholm syndrome!3
If Linux could recognize all of your hardware, all of your free software would run; if there is a bug anywhere in your hardware or
device drivers, then it is quite possible that no software will run.
Therefore, step one of World Domination by Linux is World Installation. The software incompatibilities will be better solved as soon as
the hardware compatibilities become better solved. Therefore, it is
the kernel that is currently holding up the PC revolution.4
I've installed Linux on a number of computers and found several
problems: the fingerprint reader and other optional hardware often
doesn't work, the computer doesn't always come out of sleep, sometimes the modem drivers aren't available, etc. None of my particular
problems were a barrier, but others aren't so lucky. I recently tried
to upgrade my dad's eMachines computer from Ubuntu 6.06 to 8.10
but the install failed because of APIC incompatibilities.5 There are
many long discussion threads on the Internet about Linux hardware
incompatibilities.
Linus has overseen the design of an incredible platform, with support for more hardware than any other OS ever, and his kernel is the
best piece of large free code ever written, but it needs further work.
One of the biggest challenges I see for the kernel development community is a lack of respect for their buglist, the most important metric describing the state of the kernel.
Wikipedia: Stockholm syndrome is a psychological response sometimes seen in

an abducted hostage, in which the hostage shows signs of loyalty to the hostagetaker, regardless of the danger in which they have been placed. Loyalty to a more
powerful abuser in spite of the danger that this loyalty puts the victim in is
common among victims of domestic abuse, battered partners and child abuse. In
many instances the victims choose to remain loyal to their abuser, and choose not
to leave him or her, even when they are offered a safe placement in foster homes.
The X Windows (video) subsystem is likely the next biggest obstacle. Video problems are especially tricky because it is hard to fix your computer when the display
is not working! The external monitor stuff has been difficult for years and a UI for
setting these options has only recently been added to Linux. Even the free drivers
have some bugs, but the proprietary ones are years behind.
I even passed in NOAPIC NOLAPIC flags to the install program, but the install
still did not work. I re-installed 6.06 and it worked fine, so this appears to be a
kernel regression.
214
Fix the F'ing Hardware Bugs!

Never, never, never, never give up.
If you are going through hell, keep going.
I am easily satisfied with the very best.
Winston Churchill
At Microsoft, we had it beaten into our heads to fix bugs: a bug

meant an unhappy customer, and a bug that affected just 1% of
users meant that there were millions of unhappy customers!6 Software that doesn't work is not worth anything.
As I write this, the Linux kernel has 1,400 active bugs, and the
median age of those bugs is ten months a number three times
longer than its release schedule. In general, a bug, whenever it is
found, should get a fix put in the next release. The Linux kernel is
shipping frequently, but many bugs are taking three to four releases
to get fixed.
Bugs get stale if they sit around for too long. If I file a bug against
some device driver, and the owner of the code gets back to me one
year later, I may not have all the necessary pieces set up to reproduce the bug again. Therefore, the bug must be resolved as notreproduced and can get fixed only when the next unhappy customer
finds it. Furthermore, the fact that a buglist already appears voluminous discourages people from adding to it.
One might believe that the Linux programmers are fixing bugs as
fast as they can, but in fact they aren't. The Linux kernel development team has thousands of contributors and thousands of bugs, so
fixing their bugs is a very manageable task if they make it a priority.
The good news is that it is easier to fix bugs than to write code
because writing code involves design, which in turn requires difficult choices to be made. (e.g. How do I add this feature?) In the bugfixing phase of a software product cycle, most of the hard decisions
have already been made, so it is mostly a matter of making small
tweaks the architectural equivalent of changing a blueprint to
move a sink a few inches.
You might not believe that, but I have the early bad reviews to prove it!
Microsoft's software is considered unreliable today not because the developers
don't care, but because they are burdened by old code, and have a development
organization far too small for the vast scope of technology they release. They have
perhaps 10,000 developers in total, whereas the Linux kernel alone has 3,000
developers.
215
Fixing these bugs might be tedious, especially when the developer doesn't have access to the hardware he is trying to debug, but
everything about hardware is tedious so they may as well just do it
now. PC hardware contains 10,000 devices, along with a tremendous
amount of unnecessary complexity and undocumented designs, but
the messiness of the PC has existed from the moment Linus first created his kernel. Linus, however, has the means to make all PC hardware work on Linux: lean on hardware vendors to support free
software, and crack the whip on his kernel developers!
1,400 bugs for something as big and actively evolving as the
Linux kernel is not a disaster at all. But if Linux kernel developers
focused on the buglist for just a few months, the bug total could be
brought down to less than 50. Linux's goal should be to fix 90% of
new bugs in the next release, and 99% within two releases. This
accomplishment would represent yet another breakthrough for a
project of this size and complexity. Linux is a superior kernel to Windows today, but it needs a bit more work on the mess that is today's
PC hardware.7 The whip Linus must crack on his kernel developers
is actually nothing more than a feather boa.
Metrics
One of the great mistakes is to judge policies and programs by
their intentions rather than their results.
Milton Friedman
What's measured improves.
Peter Drucker
Microsoft conducts yearly polls that give them statistically significant data on the state of the company. If they want to know whether
developers were happier than they were five years before, they
can get hard numbers on this. By gathering the right data and
watching it over time, management can fix any organizational
trends heading in the wrong direction.
This data was collected and managed by the Human Resources
Department, but every other Microsoft team had metrics as well,
and by far the most important was the bug count. The Microsoft
buglist almost isn't even considered a metric because it is the driving force in the development process.
7
In fact, if hardware companies did not finalize their hardware until their Linux
driver was written, hardware itself would become more reliable. Most hardware
today is designed before the software is written. When the drivers finally get written, it can expose hardware bugs. Free drivers allow for better hardware and simpler software.
216
A lesson of success is the importance of metrics in driving

change. Rudy Giuliani explained in his book Leadership that turning
around New York City required a lot of hard work, but the key was
to gather and publicize data about problems and track progress on
them.
People might argue that data is a bureaucratic waste of time and
rightly point out that by itself it doesn't change anything. What they
are missing is that it gives people information which serves as motivation. When kernel developers see that the median age of Linux's
bugs is ten months, they will realize that they should do better. This
will motivate them to spend a bit more time working on the older
bugs.
Every software team needs to find metrics that give good analysis
of the state of their product over time. Metrics also help potential
users of a piece of software; just knowing the total number of users
and developers provides an excellent way to analyze the health and
success of a component. The other nice thing about such a metric is
that it is a self-reinforcing feedback loop that causes the fastest
growing teams to grow faster.
Volunteers Leading Volunteers

A manager's task is to make the strengths of people effective
and their weakness irrelevant.
A manager sets objectivesA manager organizesA manager
motivates and communicatesA manager, by establishing yardsticks, measuresA manager develops people.
Peter Drucker
Many of Debian's leaders were grad students hanging out in

academia. One might wonder if someone working part-time can
effectively lead thousands of people, but if you add up all the time
corporate managers spend courting customers, creating budgets,
and attending the required training on Corporate Values, one might
find that very few spend a significant portion of their time thinking
about the long-term big issues. It is very easy to get caught up in the
day-to-day.
Because everything in volunteer organizations like Debians run
on autopilot, the leader has the opportunity to focus on the big problems and be effective, even while working part-time. However, leading volunteers is a task as difficult as herding cats as you must coax
each of them to do something.
217
There are benefits to a flat command structure: people are forced

to convince others to do something based on the merits of their
ideas. Because decisions are made on their merit, issues generally
resolve themselves, with the person doing the work making the ultimate decision in a Do-ocracy. Everyone wants to do their best, and if
they don't write correct code, bugs will come in, and developers will
be forced to confront their mistakes.
In addition to metrics, task forces are a great tool of leadership.
With small teams it is easier to get a consensus and produce a
coherent document describing a problem and the proposed solution.
Coming up with a proposal is easier than implementing it, but it can
serve as a rallying point.
For example, Debian uses a lot of different tools to manage the
history of the source code, with the number of tools increasing as
Debian has grown. A task force could determine the recommended
source control system. Such a recommendation wouldn't immediately change all the source control systems out there, but it would
serve as a good default choice in the future, and it would push
Debian, and the rest of the community, towards unification. Each
task force needs to establish credibility and validity on its own to
convince others in the loosely distributed meritocracy that is the
free software movement.
This book does not go much into specifics of how to create and
foster a healthy community around a piece of free software. However, many of the keys to success are embodied in old-school books
such as Dale Carnegie's 1936 classic, How to Win Friend and Influence People. For a more thorough treatment of this topic, I recommend Jono Bacon's Art of Community.
Must PC vendors ship Linux?

Many argue that hardware companies need to start shipping computers with Linux pre-installed before Linux will take off. While this
is helpful, it is not necessary, and not sufficient.
The best first step for the hardware vendors would be to ensure
that all of their hardware components have drivers in the Linux kernel tree. Once this has happened, it will mean that, in general, every
distribution of Linux will run on every model of their machines. In
many cases, different models use many of the same components, so
supporting the 20th machine is often no more work than supporting
the 19th.
218
Once the drivers are in the kernel tree, there is no need for a
hardware vendor to pre-install Linux because the user can do it
themselves. Windows is installed by hardware manufacturers
because the retail version is missing drivers, but this need not be
the case with Linux. (In fact, if your hardware is supported,
installing Linux takes less than an hour.)
Eric Raymond, in his recent essay World Domination 201,
argues that the free software community should only purchase PCs
from vendors that ship Linux. However, it is much simpler just to
convince hardware vendors to support Linux than to create new
multinational corporations.
Even if vendors don't provide support for Linux yet, they could
make sure it runs. Again, while people might debate whether free
software is good for software companies, it is inarguably a benefit
for hardware companies because it lowers the cost of a computer. In
fact, PC hardware vendors don't even need to do any work other
than demand Linux drivers from their component vendors. The
power easily lies within the PC hardware vendors to apply the necessary pressure.
Shipping computers with Linux is not sufficient for world domination because hardware vendors don't have the means to solve some
of Linux's problems. Dell cannot ensure that Apple's iTunes works
on Linux, or that OpenOffice can read your Publisher files.
219
The Desktop
The only strategy in getting people to switch to your product is
to eliminate barriers. Imagine that it's 1991. The dominant
spreadsheet, with 100% market share, is Lotus 123. You're the
product manager for Microsoft Excel. Ask yourself: what are
the barriers to switching? What keeps users from becoming
Excel customers tomorrow? Think of these barriers as an obstacle course that people have to run before you can count them
as your customers. If you start out with a field of 1000 runners,
about half of them will trip on the tires; half of the survivors
won't be strong enough to jump the wall; half of those survivors
will fall off the rope ladder into the mud, and so on, until only 1
or 2 people actually overcome all the hurdles. With 8 or 9 barriers, everybody will have one non-negotiable deal killer.
This calculus means that eliminating barriers to switching is
the most important thing you have to do if you want to take
over an existing market, because eliminating just one barrier
will likely double your sales. Eliminate two barriers, and you'll
double your sales again. Microsoft looked at the list of Lotus
123 barriers and worked on all of them:
Barrier
Solution
They have to know about Excel and

know that it's better
Advertise Excel, send out demo disks,

and tour the country showing it off
They have to buy Excel
Offer a special discount for former 123

users to switch to Excel
They have to buy Windows to run Excel
Make a runtime version of Windows

which ships free with Excel
They have to convert their existing

spreadsheets from 123 to Excel
Give Excel the capability to read 123

spreadsheets
They have to rewrite their keyboard

macros which won't run in Excel
Give Excel the capability to run 123

macros
They have to learn a new user interface
Give Excel the ability to understand

Lotus keystrokes, in case you were
used to the old way of doing things
They need a faster computer with more

memory
Wait for Moore's law to solve the problem of computer power
And it worked pretty well. By incessant pounding on eliminating barriers, they slowly pried some market share away from
Lotus
Joel Spolsky, Joel on Software
Free software developers have had a long row to hoe. As proprietary software has been the dominant model for decades, many
geeks have had to live with one foot in each world; Linux programmers today are often forced to use Macs, Windows and other proprietary software.
220
The good news is that a Linux PC does a reasonably good job of

interoperating with proprietary hardware and software, especially
Microsoft technology. Today, Linux supports the Microsoft Office file
format, Windows Media file formats and protocols, MSN Instant
Messenger protocols, Microsoft file systems, C#, and much more.
The implementations aren't perfect, but just the most popular 95%
of a standard is good enough for most. In fact, by tackling the easy
and more important portions, the codebases are smaller and simpler
than their Microsoft counterparts.
In general, when a free software application supports a file type
or protocol, the support is good, but there are still Microsoft formats
such as Publisher, Money, and Access whose lack of support is problematic today. Having worked on Microsoft Publisher, I can say it is
about 5,000 lines of code for OpenOffice to support reading these
documents, which is not a big deal. The work can be done by an
enterprising volunteer, or via a bounty.
As mentioned in the Apple section, one of Linux's most surprising
capabilities is its ability to run Windows applications. Given a complete operating system, implementing the Win32 syntax for similar
functionality is not difficult. Therefore, a backup strategy for supporting proprietary file formats is to run the proprietary Windows
programs on Linux.
I was able to switch to using Linux 100% of the time, and cannot
foresee a reason to switch back, but others are not so lucky. Work in
a number of areas needs to continue.
Approachability
The adversary she found herself forced to fight was not worth
matching or beating; it was not a superior ability which she
would have found honor in challenging; it was ineptitude a
gray spread of cotton that deemed soft and shapeless, that
could offer no resistance to anything or anybody, yet managed
to be a barrier in her way. She stood, disarmed, before the riddle of what made this possible. She could find no answer.
I believe the free software community ought to focus on a fully

functioning out of the box (OOB) experience. People need to be able
to take a Linux CD, stick it into a new, or old, computer, have it recognize the hardware, install everything, and migrate the data. A key
to Linux's success, and to the success of computers, is that it should
work with minimal user input.
221
Installing Windows, as many people know, can take a full day.8 To

beat Windows, Linux will have to do better. Fortunately, even with
all of its warts, Linux isn't that far off. I wrote a post on my blog in
April, 2006 entitled 10,000 Bugs Away from World Domination. In
it, I wrote that if we could fix about 10,000 selected bugs in the free
software stack (50 man-years of work, compared to Vista that had
5,000 man-years of effort), we could come very close to building an
operating system that someone as computer unsavvy as my mom
could install and maintain. Right now, setting up Linux can take ten
steps, but any issue you run into is likely to be arcane to fix.
Even if we never get there for everyone, we can definitely get
closer than we are now which is important because fixing computer
problems is never easy, even on Windows. In fact, a perk of Microsoft VIPs was that they had personal IT staff for computer setup and
maintenance! With their established IT infrastructures, large businesses should be able to move to Linux today, even with its warts.
In general, while the Linux kernel has a few remaining challenges, the biggest challenge for the next five years will be for the
rest of the free software stack to get up to the same level of quality
as the kernel. While Firefox and OpenOffice are rich and quite stable, few would argue that these popular and important applications,
or any other free software for that matter, has the quality, reliability
and elegance of the Linux kernel.
In some ways, applications are harder to build because unlike a
hardware driver, which is by definition concrete and fully specified,
a web browser and a word processor are applications whose full
requirements are unknown and unlimited. What should a web
browser do while disconnected? If speech recognition becomes the
de facto input method, how does this change the user interface of a
word processor? These are questions to which answers are elusive.
In the tools chapter, I discussed how modern tools will build more
reliable applications, but the important point here is that better
tools will make developers more productive, which will allow software to evolve more quickly as we better figure out the answers to
some very hard questions.
It is important that free software programs, and their codebases,
be approachable. Wikipedia's motto: Don't bite the newcomers is a
Assuming you don't run into a situation where you need a network card driver, but
you can't download it until you have a functioning network card!
222
great one for every free software organization. A big challenge for
the next few years is to make the entire free software stack easier to
use and more approachable for new users and programmers.
Many wonder whether Linux needs a killer application to beat
Windows. Linux's status today of being mostly compatible with, and
slightly more reliable than Windows might not be enough to overcome the inertia required for a worldwide switch, even to something
that is free to acquire.
It is true that if Linux were to enable robust continuous speech
recognition, or some other transformational feature, Linux could
more quickly take over, but even that OS would also need to support
all your hardware and file formats as well. However, Linux already
has what I think is a killer feature: the rich set of programs that
come with it although they could still use a bit more work. For
example, the most popular high-end graphics editor in Linux, poorly
named the GIMP, is comparable to Photoshop, suitable for professionals, and free! However, like Photoshop, it isn't particularly
approachable for a new person to jump in and start using. The first
time you use GIMP, it might take 20 minutes to figure out how to
crop an image. It takes time to learn any powerful tool, but a lesson
in engineering is: You only pay for what you use. Simple things
should be simple to do, and today this is not always the case. While
many of the free applications are good enough to convert people
from Windows, and as good as competing proprietary products, they
could still be dramatically better.
While the free software community is already doing good business, especially for websites and embedded applications, PC hardware and software are the biggest and final challenge. The minimal
requirements, the Web and Office, are almost met today, yet its true
potential, its Wikipedia-scale potential, is something much larger
and several years of work away.
Once we start to coalesce into a few very good codebases, then
the PhDs will be motivated to jump in. Nowadays, a graphics
researcher might use GIMP to manipulate images, but he would definitely not use GIMP's source code as the basis for his research.
Eventually computer scientists around the world will instinctively
realize the place to contribute is in the free software code, just as
there are linguists and computer scientists today using Wikipedia as
the basis for their research. Once the PhDs get involved, software
will get very interesting.
223
Monoculture
Next to the problem of code written in old programming languages, the second biggest challenge the free software community
faces is the amount of duplicate code that exists. There are several
dictionaries in Linux, and so when you add a custom word to one,
only some applications notice it. In Linux there are too many paint
programs, media players, download managers, RSS readers, programming environments, source control systems, etc. Each piece of
duplicate code is a missed opportunity for sharing and slows
progress. On the other hand, the free software community has gotten together on primarily one kernel, Linux, and one web server,
Apache.
There is an interesting debate in the software community about
the worry of a software monoculture. Like a monoculture in the biological world, one virus or other problem could destroy the entire
ecosystem, and some argue that we run similar risks in the software
world.
While a monoculture may be a risk, having different codebases
doesn't necessarily help: any new, previously unforeseen exploit is
able to cause damage to all codebases because none of them
designed for it. A castle's walls may stop men, but they were not
designed to stop cannonballs or helicopters.
While the little differences between codebases add an extra level
of variability that makes building a virus harder, software is different from DNA (today at least!) because of our ability to infinitely
recombine it. It is possible to make a product easy to use, and powerful, rich and reliable, fast and maintainable. Because software is
infinitely malleable, all the best ideas can be incorporated into a single product.
In fact, the monoculture risk only applies to the proprietary software world because in that model we are unable to infinitely recombine ideas. There are hundreds of millions of Windows users, but
there are only a few thousand people at Microsoft who can make
changes to it.
If the monoculture risk doesn't apply to the free software world,
we should focus more on working together. Progress in the free software world is slow where people are not working on the same codebases. As described in the tools chapter, switching to modern tools
will bring a 2x productivity improvement, and consolidating various
efforts into unified codebases will bring an additional 5x productiv-
224
ity jump. Unfortunately, agreeing to unified efforts is a much harder

challenge because each group needs to reach out to other groups,
and be willing to combine their efforts and potentially abandon their
existing software investments.
SourceForge.net has 1.9 million programmers, but when you
divide by 50 because they are volunteers, and then divide by 10
because of old tools and duplicate code, our 1.9 million-man army
has only 3,800 people. However, it is amazing what free software
has built when you realize that today it is on just a (relatively) small
number of PCs. Even with that tiny user base, and their inefficiencies, they still have had the resources to do everything they have
done. When you add to the 10x productivity improvement another
5x increase in users and contributors, free software in five years
could have what seems like 50 times as many programmers as they
do today.
225
Linux Dev Tools

Simplicity is the ultimate sophistication.
Leonardo Da Vinci
Developers, developers, developers, developers, developers,
developers!
Developers! Developers! Developers! Developers! Developers!
Developers! Developers! Developers! Developers! Yes!
Steve Ballmer
Parody of Eclipse packaging, the most popular free developer tool, from the
funny folks at FarOutShirts.com. I took a few pages of notes of my frustrating experiences but decided that the picture above did a better job.
226
Over the years Microsoft has invested a lot in the development of

tools, and this generated much of its revenues in the early years. In
fact, Microsoft was a tools company before it was an operating systems company.
Next to Windows, the most important product Microsoft produces
today is Visual Studio. The tool has flaws, but it wins awards, especially for its ease of use. It is possible to create your first Windows
application or web site just minutes after installing it.
The closest thing to Visual Studio in the free software world is
Eclipse, which is despised by many for its excess complexity. Craig
Mundie, Microsoft's Chief Research and Strategy Officer, asserted in
a recent interview that the Linux server market share had
plateaued. If this is the case, the state of Linux's tools has to be a
big reason for this.
Eclipse is more powerful than Visual Studio, and can be used to
develop all kinds of applications, but it has never had the same
approachability. Eclipse has gotten better over the years, but it still
has a long way to go. The community needs to focus on building
powerful yet easy to use tools.
Backward Compatibility
As described in the Linux chapter, all of the source for the entire
software stack is publicly available so fixes can be made in the
proper place. This is an important part of what makes free software
simpler, more maintainable and more reliable over time.
The free software stack has traded simplicity for backward compatibility and so there is a downside. For example, Windows lets you
download drivers off the web that will work on multiple versions of
its OS, but Linux does not support this scenario because the internal
kernel architecture is constantly changing. If hardware vendors did
make Linux drivers available on the web, they would potentially
have different drivers for every version of the kernel. The best way
to get a new driver in Linux is to grab the latest kernel, which is
where all the new drivers go.
It is possible that if you buy a new piece of hardware, your old
kernel might not have support for the new driver. What then?
To be clear, this scenario is contrived in several ways:
Anyone can add a driver to the Linux kernel, so the need to

ever download code from a website is greatly reduced.
227
The vast majority of the hardware you would buy in a store

will have drivers that have been a part of the Linux kernel for
several years. Even new hardware typically uses old drivers.
Linux ships more frequently, and the code is free, so it is less
likely you will be running an old version of the Linux kernel.
This is not a problem that has ever bitten me, but it is theoretically possible given its development model. In this situation, it could
be that the simplest way to get your new hardware to work is to
install a new kernel. Unfortunately, most Linux distributions do not
allow you to just update the kernel as you must upgrade the entire
OS. This is a notable downside of today's Linux.9
However, this is a trade-off in the development process that
favors simplicity, reflects the dynamic nature of software, and so the
need to upgrade software periodically is a fact of life that needs to
be accepted rather than resisted. The alternative is a backward
compatibility quagmire that Microsoft wallows in.
There are other backward compatibility issues. OpenOffice's
native format is OpenDocument, but you can't use it until all whom
you might share documents with are also using that software. And
even when the documents are converted to the new format, the files
are often hardcoded with font names owned by Microsoft. Proprietary software's biggest advantage is its switching costs. Oracle's
proprietary database is powerful, yet unwieldy and old, and application vendors who have spent thousands of hours writing Oracle-specific code, have also created a situation where moving to any other
product, even something free, is expensive. Fortunately, these are
one-time costs.
One solution is to build an OS that continuously and smoothly upgrades itself to

the latest software, a feature that is not offered on many distributions today. I
think it would be great to always have the latest and greatest software on my
computer without having to lift a finger. This is primarily a testing challenge.
228
Standards & Web
STANDARDS & WEB

From:
Bill Gates
Sent:
Saturday, December 5, 1998
To:
Bob Muglia, Jon DeVann, Steven Sinofsky
Subject:
Office rendering
One thing we have got to change in our strategy - allowing
Office documents to be rendered very well by other people's
browsers is one of the most destructive things we could do to
the company.
We have to stop putting any effort into this and make sure that
Office documents very well depends on PROPRIETARY IE capabilities.
Anything else is suicide for our platform. This is a case where
Office has to avoid doing something to destroy Windows.
I would be glad to explain at a greater length.
Likewise this love of the standard DAV in Office/Exchange is a
huge problem. I would also like to make sure people understand this as well.
o use the Internet, you need software that supports two big
standards: TCP/HTTP and HTML. There is no HTML standard competing with an HTMM standard, as the idea is silly
on its face, yet such redundancies exist in many other areas in the
world of bits today. When you can't agree on a file format, your ability to exchange information goes from 1 to 0.
Free software has been a part of the Internet since the beginning.
In fact, a website needs to send you software, in the form of HTML
and JavaScript, in order for you to have something to look at and
interact with. It is easy to learn how a website does its magic as the
tags of an HTML document are self-describing and, on top of that,
there is an organization called W3C whose job is to fully describe
them.
In contrast, to display a Word or WordPerfect document, you had
to reverse-engineer a complicated binary file format! This prevented
their widespread use as the standard document format of the Internet.
Standards & Web
229
Digital Images
Unlike audio and video, in the realm of still images things are in
good shape. JPEG is an efficient, free, and widely-supported standard for compression of images.1 There might be a couple of standards better than JPEG out there, but the end is nigh. (There is a
JPEG 2000 standard based on wavelets2 which is 20% better than
JPEG, but it has higher memory and processing demands. Sometimes to get a little more compression, you have to do a lot more
work, and so you reach a point of diminishing returns.)
Microsoft in early 2007 announced a new Windows Media Photo
format, which is also 20% better than JPEG, but without higher
memory and processing demands like those required by JPEG 2000.
The new format is based on JPEG, but with nine small tweaks. The
spec is public, and there is even free public source code, but the
license explicitly excludes it from being used in combination with
copyleft licenses:
2. c. Distribution Restrictions. You may not modify or distribute the source code of any Distributable Code so that any part
of it becomes subject to an Excluded License. An Excluded
License is one that requires, as a condition of use, modification
or distribution, that the code be disclosed or distributed in
source code form; or others have the right to modify it.
Fortunately, reverse engineering is easily doable because the spec

is public and if you start with existing JPEG code you are already
very close a tiny, hidden, speed bump Microsoft has created for
the free software community.
Digital Audio
It is the mess of proprietary standards and patent restrictions
that are impeding the progress to digital audio. Proprietary software
companies have been pushing their standards down our throat. If
1
There is also PNG and GIF for lossless compression, but they are not suitable for
real-world images with continuous gradients like clouds, etc. I took a high quality
1.9 MB JPEG and converted it to PNG and it became 2.9 times bigger.
Interestingly, these lossless formats do a better job than JPEG for certain images
like screen shots because JPEG doesn't handle the sharp transitions from black to
white, etc. that you find on a computer screen. A JPEG of a screen shot is 2.2
times larger than an equivalent PNG. With a JPEG of the same size, you find it has
added gray display artifacts at the black/white transitions.
PNG was only created because after GIF became popular, Compuserve started
suing.
Wikipedia: A wavelet is a kind of mathematical function used to divide a given
function or continuous-time signal into different frequency components and study
each component with a resolution that matches its scale.
230
Standards & Web
you insert a music CD into a Windows computer, it wants to rip the

audio into WMA, a proprietary format which the Mac OS doesn't
support out of the box. If you insert that CD into a Mac, it rips the
music into AAC, a format that Windows doesn't support by default.
I'm not sure what should be done about this colossal mess. We
should keep trying to pick a standard format for audio, like JPEG has
become for still images. The best candidates are OGG and MP3.
MP3 is an old format and while it is not considered state of the art,
it is efficient enough. MP3's primary problem is that there are a
number of companies with patent claims against it. If the industry
cannot agree to end the patent hassles, it should adopt OGG or
some other fully free format and tell those MP3 license owners to
take a long walk off a short pier.
If we could finally agree on a free digital music format, we could
finally have digital music which will also open up many possibilities of a richer connection between artist and consumer. Whatever
format is chosen, we need to define the number of bits per second
required to achieve transparency.3 We also need to come up with a
standard streaming protocol and video format, but I won't even get
into that here; we must crawl before we can walk.
The Next-Gen DVD Mess

Like many other things in the computer industry, the next generation format for high definition DVDs was a mess for several years
because two standards were created: HD-DVD and Blu-Ray. Both
formats are high quality and their discs look the same, but the playback hardware is inexplicably incompatible.
Fortunately, in early 2008, HD-DVD was abandoned, although billions of dollars were wasted because the two camps bickered like
twin sisters, and couldn't agree on a few minuscule technical details
before they went to market. The bad news is that Blu-Ray adoption
is slow. In mid-2009, it was reported that HD-DVD players are 60%
more popular than Blu-Ray, even though it had been abandoned by
the industry. Furthermore, why pay $300 for a new player when all
of the movies we already own are not in that format? Why should we
pay full price merely to get a higher quality copy of something we
already own? Didn't we really mean to purchase a high quality copy
3
Transparency is defined as audio that is of such high quality it cannot be

distinguished from a CD. Transparency should mean passing the aforementioned
listening test, plus when you convert it to other transparent codecs and back,
even 100 times, the quality is not diminished. To achieve that we do not need to
go to lossless compression, which is five to seven times bigger.
Standards & Web
231
in the first place? The best way for faster adoption would be to create a service where you mail in your old VHS or DVDs, and they
mailed you back HD versions of them for a few dollars each. That
could be a huge, if low-margin, business.
Consumers would adopt new standards faster if they could immediately enjoy everything they already own in that new format at
something approaching the actual cost of producing a disc, instead
of the full retail price. This is another area where shrinking the
copyright expiration will help. It will make nearly free all those
things we paid a license fee for, but which we now only have lowquality reproductions of.
At some point, the industry should remove the Interpol and DHS
warnings about why we shouldn't have stolen what we are about to
watch. When you stick in a disc, there should be two buttons that
show up within 5 seconds: Play and Menu. I have a boxed set of
DVDs that display four minutes of introductory warnings and selfpromotion that I cannot even skip through, before playing the actual
content. I once put in the wrong disc and so had to repeat the
hassle. This may seem petty, but these encroachments get worse. I
have to accept the license agreement of my navigation system every
time I turn on my car!4 When people feel ripped off and treated like
a sucker, a culture is created where people decide not to pay for
what is sold. Respect is a two-way street.
Suppose I don't activate my nav system one day because I can't be bothered. Then
I get into an accident and die because the system wasn't helping me. It could
therefore be argued that the need to accept the license agreement played a factor
in the cause of my death. The good news is that it would create standing to sue!
Now, if only we could come up with a way for someone to die because they had to
sit through all that stuff at the beginning of a video so we could get standing for
that!
232
Standards & Web
MS's Support of Standards
The shapes of the Latin alphabet are mostly happenstance. Many details in
this world do not matter, it is only that we agree to them.
Microsoft had a mixed attitude towards standards. It played a key

role in the creation of many, such as: HTTP, HTML & XML, USB,
ACPI, Unicode, and TrueType, and also supported standards it didn't
play a part in creating, like TCP, SSL, SQL and SMTP.
However, in many cases, Microsoft had a strategy to create a
competing standard, and wasn't particularly interested in documenting its technologies into open standards. Before HTML and HTTP
took off, Microsoft had an effort, code-named Blackbird, which contained many web-like features, but all elements were specified in a
Microsoft-only world. Microsoft eventually canned Blackbird but, for
a time, it was a much bigger deal inside the company than Internet
Explorer, which started off as a skunkworks project based on the
acquisition of third-party code. In most cases, Microsoft supported
standards only when it felt commercial pressure to do so; one of the
mottoes in the early days was that Microsoft: sets the standard.
So many of the company's engineers wrote code all day long on
market-leading products that working with competitors to standardize arbitrary details seemed pointless. The protocol to share files
and printers between computers went undocumented for many
years; what would be the point when it was never a priority to get it
Standards & Web
233
working on Mac and Unix? Proprietary protocols are a special type

of bad standard because they tie you to two pieces of someone's
technology. Microsoft's Outlook uses a proprietary protocol to fetch
e-mail from Exchange; this protocol ties you to client and server.
The free software guys do not, and should not, have any compunction about implementing closed, undocumented or poorly documented standards, as long as the standard is popular. An open
standard is always better, but it is the source code that is the true
value of intellectual property, not whether the details are sanctioned
by a particular organization.
Free software supports more standards than proprietary software. If a standard is popular, it will get implemented. Linux by
default rips music into the free OGG, but it is easy to change it. The
Linux media player supports Microsoft's and Apple's media formats,
plus free standards. Linux's instant messaging applet supports MSN,
Yahoo, AIM, ICQ, GroupWise, Jabber, and others. In fact, proprietary
software is holding up the development and adoption of new standards. The BBC has created a free wavelet-based video codec known
as Dirac, but only Linux supports it out of the box today.
Since it is built by users who add features as necessary, free software gives you more choices than any proprietary vendor can motivate themselves to provide. Not only is there better support for
standards, but also you will never be locked into a corner because
support for a standard disappeared for strategic reasons. With
open formats and free software, you are building a future-proof platform.
The web is the most important standard for information
exchange, but one that is almost as important is the standard for
office documents.
234
Standards & Web
OpenDocument Format (ODF)

From:
Bill Gates
Sent:
Thursday, August 5, 1999
To:
Bob Muglia
Why would the Office Group be giving out the Office 2000 format to competitors? To me this sounds crazy.
Microsoft Office 2007 promotional screenshot (Got Blue?) Office includes

Word, Excel, PowerPoint, Outlook, Access, FrontPage, Visio, Project,
OneNote, web services, and tools to build Office extensions, an incredible
package of interconnected proprietary technologies.
Microsoft's relationship with document creators is older and more

entrenched than its relation with Windows users; I've been using
Windows since 1990, but have been using Word and Excel since
1986. For hundreds of millions of information workers, students,
writers, and government employees throughout the world, the primary tools for producing their intellectual property is inside Microsoft Office, and there are billions of documents out there. I have
heard that every corporate purchase in the UK involves the creation
of an Excel spreadsheet. A friend in the military told me that creating PowerPoint documents is a required skill for the upper echelons.
Standards & Web
235
Microsoft Office is a huge part of Microsoft's profits, and the best

reason to install Windows. Today, about 50 million people use the
free OpenOffice, while the rest of the world has paid $200 or $0 for
a copy of Office. (The name OpenOffice is trademarked by someone else, so their official name is actually OpenOffice.org.)
When you realize how much intellectual effort is expended inside
productivity tools, you realize how important it is that the file formats be documented. A company or a government would like to
know that they can keep these files around for decades like they can
with paper. This is, of course, a much harder challenge because on a
piece of paper all the digitized words, text layout rules, and other
formatting information are lost imagine if someone replaced all
your Word documents with screenshots of those documents; the
information would be readable, but not editable.
The Microsoft Word team didn't try to build a file format they
would be happy with for 20 years because they knew that engineering Word in 1993 to read the files of Word 2013 was an impossible
task. For the first 10 years, Word n could never even read the files of
Word n + 1. This is because for many years Office used binary formats, and it would have crashed if it tried. Inside a binary Word file,
you might find the following data:
05
bold
01
on
'0501' means turning on bold

05
bold
00
off
'0500' means turning off bold

06
?!
??
?!
What if an old Word doesn't understand '06'? What next?
If you don't know what '06' means, you don't know what is coming next, and so you can't continue and must abort. Even if you
wanted to skip over it, you can't because you don't know how far to
advance.
Microsoft did not create binary formats to lock out other vendors.
Formats by all the word processors were binary for many years
because it is efficient and because a better solution hadn't been
invented.
236
Standards & Web
The answer to this conundrum, which has plagued computing

since the beginning, is for documents to be self-describing, and that
is what eXtensible Markup Language (XML) is all about. XML
defines how to create a document which any software can read and
write without crashing, even if it doesn't completely understand it.
I can guarantee that the following XML (fragment, trimmed for
clarity) will be readable by a word processor in 20 years:5
<?xml version="1.0" encoding="UTF-8"?>
<office:document-content office:version="1.0">
<office:body>
<office:text>
<text:p text:style-name="Standard">Hello, 2028!</text:p>
</office:text>
</office:body>
</office:document-content>
Hello, 2028! in Open Document Format (ODF)
XML is built upon Unicode, the standardization of all the world's

characters, and adds to it <brackets> which allow you to find the
end of every element, even those you don't understand.</brackets>
As tiny as it sounds, the key to being able to read old files years
later, or new files with old code, is simply to make the data selfdescribing, allowing you to figure out a piece of data's name and
length. The brackets, and the exact XML format, isn't important,
what is important is the fact that all types of information you'd like
to represent are doable in a way that is possible for humans to read
and for computers to manipulate. When every computer uses a standardized, self-describing format, we will have taken a big first step
in being able to exchange documents without causing a crash.6
XML is about having a standard, self-describing file format and is
one of the most important standards in the history of computing,
and just one of its uses will be the standard schema to represent the
billions of office documents.
Even if the schema does change in an incompatible way, it should be possible to

write a tiny program to update to the new schema try doing that with a binary
format. XML also allows you to keep but ignore information you don't understand,
whereas code that supports a binary format usually throws away things it doesn't
understand.
There could be other formats than XML which is not particularly efficient for
computers to parse. Efficiency was never a part of the design, as the code was
created after the spec. The XML people say that they can come up with more
efficient binary formats and I suggest we take them up on their boast. Perhaps
proprietary software is holding this up because a binary XML would require
changes to many XML codebases.
Standards & Web
237
The Office binary formats were not documented for many years,
and the license agreement for the documentation today says that
you can only use the information for products that complement
Microsoft Office. Is supplant the same thing as complement?!
Microsoft is not particularly interested in building an open standard because it will never perfectly represent their features, and
because an open standard makes it easy to switch tools. Right now
everyone buys Office because that is what you need to read the documents you receive today. The adoption of an open format for productivity tools is a mortal threat to Microsoft's Office profit margin.
Regrettably, there is a battle going on in the XML office document
space. Microsoft has for many years ignored and then resisted the
ISO standard called OpenDocument Format (ODF), and now they
have created their own competing standard called Office OpenXML
(OOXML). However, the whole point of a standard is to not have two
of them.
XML provides the structure for your files and guarantees that
applications should be able to parse everything, even parts they
don't understand, without crashing. Given that baseline, it should be
possible to create a format that can represent the features of office
productivity tools. Microsoft's OOXML specification, which provides
100% compatibility with Microsoft Office, is 6,000 pages, while the
ODF specification is only 1,000 pages because it doesn't re-use
many existing standards like SVG, SMIL, MathML and XForms.7
OpenXML is also filled with legacy bloat. At the top of 600 pages
of the VML specification is the following text:
Note: The VML format is a legacy format originally introduced
with Office 2000 and is included and fully defined in this Standard for backwards compatibility reasons. The DrawingML format is a newer and richer format created with the goal of eventually replacing any uses of VML in the Office Open XML
formats. VML should be considered a deprecated format
included in Office Open XML for legacy reasons only and new
applications that need a file format for drawings are strongly
encouraged to use preferentially DrawingML.
Stphane Rodriguez documents a number of defects in the OpenXML standard:

http://ooxmlisdefectivebydesign.blogspot.com/.
238
Standards & Web
Microsoft has since moved to deprecate the VML spec in favor of

the DrawingML format which is equally as proprietary, but Word
2007 generates VML, so deprecating the spec doesn't make the
work go away for apps that want to interop with Microsoft's. Google
wrote in their analysis of OpenXML:
Although OOXML may formally comply with Ecma, it was
clearly not designed with an open spirit. Comparing the current with the future situation, interoperability is likely to
become more difficult instead of easier. The implementation of
a fully compatible ODF importer (the current efforts
regarding .doc and .xls) is not an easy task, but it is dwarfed by
the implementation of a fully compatible OOXML importer,
which we estimate to take something between 50 500 person
years, or even longer. Therefore, although it is theoretically
possible to generate an OOXML document, this document will
probably only use a very small subset of the standard.
In sum, OOXML can be compared to Microsoft giving access to
a labyrinth to which it alone owns a map; moreover, certain tunnels within this labyrinth are not accessible without a key that
only Microsoft has, and that third parties would need to replicate first. (And, in doing so, these third parties would not know
whether they would violate any rights that exposes them to litigation).
All things being equal, leveraging existing standards is better

than re-inventing them, and in choosing between two standards, the
one which is smaller, because it re-uses other standards, is going to
be a much better choice for the industry. A standard which is hard to
support will be adopted slowly, or have buggy implementations.
Today, OpenXML is mostly only supported by Microsoft, unlike ODF,
which has broad industry support from companies like Red Hat,
Adobe, Computer Associates, Corel, Nokia, Intel, Oracle, Novell,
Google, IBM and Sun.
OpenXML is a proprietary format wrapped in XML. This is not a
standard suitable for use by many different types of tools over many
years. If you start on Day 1 with a lot of baggage, you are doomed.
Microsoft has aggressively lobbied for OpenXML support, recognizing that adoption of ODF could make it easy to switch away from
Office.
It is important to recognize that Microsoft Office is the most fullfeatured and popular productivity tools suite on the planet, and
therefore any open format should support important Office features.
My impression from reading through the specification is that the
OpenDocument guys have bent over backwards to ensure good
Microsoft Office compatibility, and they should be commended for
Standards & Web
239
their open-mindedness. For example, I was amazed, and slightly dismayed, to find references inside the spec to DDE, an obscure and
now mostly dead Microsoft-only technology. However, this technology became a part of the Microsoft OLE monikor format, which
specified how documents would embed portions of spreadsheets,
and became an important part of Microsoft's documents, thus ODF
supports it.
A robust, standard, self-describing file format for productivity
tools will allow people to archive their documents, confident that the
format will be readable many years into the future.8 In addition, like
everything we build in computers, standards can become platforms
for other standards. When ODF incorporates scenarios to encrypt
and digitally sign documents, to notarize and transmit legal documents, and support cross-company workflow, e-forms and e-government, it could become a lingua franca in a way that Microsoft
Word's DOC and PDF combined have never been.
Competing standards is a misnomer in my opinion, so perhaps it
would be best if everyone were to adopt ODF. Sun is building extensions to Office to support this format, although if people start using
the free OpenOffice, which uses this format as its native format,
there is no real need for Office. This book was written using OpenOffice, and while the application is far from perfect, it is far beyond
good enough for most users.
The state of Massachusetts was forward-looking in nearly adopting ODF, but after tons of lobbying by Microsoft, it reversed course
and now endorses either OpenXML or ODF. They either caved on the
idea of creating a standard, or they didn't really understand the
issue. Microsoft has attempted to confuse many on the importance
The tricky thing about building a standard is that new requirements can cause
ripples throughout the design of a system. Imagine two people sitting in different
countries collaborating on one file. The challenge is that the ODF file format was
not made to be incremental, it was meant to represent an entire document. Do
you send a new copy of the document over every time the user makes a change?
This is very inefficient and yet doesn't tell the user what changed. Another
solution is to use the undo stack, but it doesn't look like OpenDocument stores an
undo stack with the document. They could just send around XML diffs, but the
XML is not usually the in-memory representation of a document, in which case
XML isn't easily usable! One solution is to have an object model (with functions
like CreateTable) on top of the file format, and can be sent between computers,
but the OpenDocument committee has not attacked this yet.
I look forward to seeing how they solve this problem, or whether they decide that
while it is a doable feature, it is too hard and outside the scope of the
OpenDocument standard. Software is infinitely malleable, but that doesn't mean
you'll like what requirements have forced upon you!
240
Standards & Web
of standards as if saying: There are many standards out there, and

code should be able to work with all of them. It is true that there
are lots of standards, but each one should serve a different purpose!
Web
The first message ever to be sent over the ARPANET occurred
at 10:30 PM on October 29, 1969. The message itself was
simply the word login. The l and the o transmitted
without problem but then the system crashed. Hence, the first
message on the ARPANET was lo. They were able to do the
full login about one hour later.
Wikipedia article on Internet precursor ARPANET
Cartoon from xkcd.com
One could write an entire book about the web, but I did want to
include some ideas here.
From a technical perspective, HTML has always been lacking as a
text-formatting standard. It is important because of two reasons:
1. It is widely used: it's famous because it's famous.
2. It has an easy deployment model.
HTML was not designed in a rigorous way by text processing
experts and so it never had the sort of awe-inspiring respect that
Donald Knuth's TeX typesetting system has had. HTML was mature
before it even added support for the concept of a page, which is why
printing does not work well yet. Text processing is a hard problem,
but not incorporating basic features like styles for seven years
demonstrates that those guys were in over their heads and they
shouldn't have reinvented the wheel.
In addition, the uptake of the JavaScript language outside of the
web has been minimal. There is nothing web-specific about the language, and so it didn't even need to be created. It has just made the
programming language tower of Babel situation worse.
In spite of its limitations, HTML today is the best cross-platform
widget set and is supported by a vast array of tools, so enterprises
should be using it for as many corporate applications as possible. A
Standards & Web
241
Boeing airplane could even have its entire cockpit UI be a web site.
You can build something simple, reliable and pretty enough if you
pick the right subset of HTML. Right now an airplane has a mess of
buttons and knobs because each subsystem of the airplane has its
own set; the guys who build the flaps don't want to share any buttons with those who control the overhead lights.
The web today is still far from being an appropriate tool for building rich applications, and nearly everything about the web is harder
than building an equivalent functionality in a rich client application.
Google has recently announced the Chrome operating system in a
vision where all the apps run on the web, but even they have created a number of applications that presumably couldn't have been
built using HTML, such as Google Earth, Picasa, and Google Desktop. Google Docs is an impressive web-based engineering effort, but
it is slow, clumsy, feature-limited, doesn't work offline, has its own
authentication mechanism, is hamstrung by the web's limited printing capabilities, and poses no threat to Microsoft's Office business
any time soon. The most interesting capability of Google Docs is its
support for real-time collaboration, but you don't need to re-build an
entire application in HTML and Javascript to add this relatively
small feature.9
The missing features of HTML are reasons why Adobe's Flash has
become so popular. Flash started off as a failed client-based programming environment, but had a rebirth as a web plugin as a way
to fix limitations in HTML.
Adobe Flash
Flash is a primitive GC runtime whose primary advantage is that
it is cross-platform like the web. Its programming language is
known as ActionScript, which is based on the standard Ecmascript,
and which is similar but incompatible with Javascript and JScript. (It
is a mess.)
Flash is interpreted, somewhat buggy and not a standard, nor is it
available as free code. We should minimize the use of Flash because
it is a big black box to the web server, web browser, search engines,
and all existing HTML tools for building and managing websites.
Flash also doesn't enforce UI standards. Every flash site works and
looks different and sometimes I can't even tell what is a clickable
In fact, this could be a feature of the operating system that would let you share an
application with any number of other people.
242
Standards & Web
button! Creators of the website, who myopically live in their application, don't notice it, but users visiting the website for the first time
do.
Many websites built by non-technical people hire programmers to
write a Flash website because of the UI candy it allows, but then
they don't update their site for years because it would require hiring
a programmer again. It is possible to build pretty, interactive web
pages using only HTML and Javascript, as modern mapping websites have demonstrated. If you want a pretty website, use pretty
pictures!
Limiting Flash to specific portions of specific pages, as YouTube
does with its video player, is quite reasonable given certain limitations of the web and the mess of video standards, but building whole
websites in Flash is a mistake and a threat to the web.
Merging the Rich Client and the Web

In addition to its limitations, HTML has been stagnant since the
release of 4.01 in 1999. We need to continue to evolve HTML, and
push tools to keep up with the latest standard. Free software lowers
the friction of code distribution and makes this easier to do. I worry
that Internet Explorer will hold up progress on HTML because it is
so popular, yet it is years behind in certain standards already, and
Microsoft has disbanded the team several times.
XHTML and HTML 5 are nascent efforts to improve the web,
although they are still fundamentally limited because as standards
they define the limits of what a web application can be. My computer has thousands of software components that web applications
cannot call into because they are not part of the HTML standard.10
The Internet is transformational tool enabling all manner of collaboration, but using the HTML text format is not required.
The holy grail of computing is to find a way to merge the best of
the rich client and the web. Java tried but failed to build a solid
cross-platform set of widgets in an extensible runtime. Maybe the
programming language that replaces C/C++ on the Linux desktop
will revive this possibility. This is one of the interesting remaining
challenges in computing.
Web Etc.
The web needs more personalized content. I live in Seattle, and I
am a huge fan of the Seahawks football team, but this is the news
10 Even if they could call into them, they couldn't install them if they weren't already
on your computer.
Standards & Web
243
article ESPN is offering to me on a website today: Bengals decline

big Redskins offer for WR Johnson. If they want me to click, they
are going to have to do better than that.
Amazon.com has music samples I can listen to but that take 15
seconds to buffer before I can hear them. Music files on the web are
often recorded below 192 kbps MP3, which means they sound worse
than a CD and we are therefore going backwards from 1982 technology.
Pictures should be big, and scale down on small screens. Drugstore.com has pictures that are at best 300 x 300 pixels:
Can you read the text on this box, like you could in a real store?
This is an actual picture of Mt. Saint Helens found on a website:
This picture is not nearly as good as being there.
An immersive reading experience is missing one thing: 200 dpi

monitors. I saw such a monitor in 2002, and it was so pretty I could
not pull my eyes away, but they have inexplicably been taken off the
market.11 Reading is tiring because it takes more work for your eyes
to recognize letters when they are jaggy or blurry.
The web needs to continue to integrate with other media like TV.
The best way to implement interactive TV is to visually overlay
HTML on top of the TV signal. The nice thing about this is that you
could turn off the distracting crawling text along the bottom of the
screen if you wanted. If the TV industry wanted a few features to
11 I noticed that its high-resolution broke websites which work in pixels, and so
screw up when pixels get 4 times smaller, by doing things like displaying only 3
words per line.
A much smaller advance in monitor technology is to use LEDs as the backlight for
LCD monitors rather than fluorescent bulbs. LEDs are more durable and more
efficient than today's fluorescent bulbs. LEDs will eventually replace many uses of
incandescent and fluorescent light because they are 45 times more efficient than
incandescent, and 7 times more than fluorescent.
244
Standards & Web
make it look prettier than a web page, that would be very easy to do.
A French company called Free is pioneering this, using free software as part of this effort. (A cable provider is a good example of a
company which doesn't want to maintain a bunch of proprietary
software.)
Everyone who produces television shows should also create an
XML schema which has information like who the guests are, etc.
This extra information allows for a more personalized experience
similar to what I get with the web: I could set my cable box to
record whenever comedian Dennis Miller is on any channel. If a
show runs over, which often happens with sports, the cable box
would be smart enough to not stop recording; it needs just a tiny bit
extra information to do this. Interactive TV is simply waiting for
someone to create, and everyone to get behind, two simple standards, each of which would be less than 50 pages if they re-used
HTML and XML. (It would also be nice to be able to watch the Seahawks no matter what city I live in. Comcast only offers me 4 football games per week. People complain: There are 500 channels
with nothing on. In truth, we aren't to that point yet!)
Hardware
Lots of people worry about running out of Internet bandwidth, but
they are just being nattering nabobs of negativism. Nippon Telephone and Telegraph of Japan demonstrated sending 14 trillion bits
per second down a single strand of fiber or 2,660 music CDs in
one second. We are not running up against the limits of the laws of
physics yet!
Metering data sounds logical, although it would seem that the distance would provide a better measure. However, I counted the number of router hops required to get data from my home in Seattle to
various destinations:
Seattle to:
Router hops
google.com (Mountain View, CA) 23

msn.co.jp (Japan)
23
www.tmobile.de (Germany)
30
www.latviatourism.lv (Latvia)
17
www.google.co.uk (UK)
15
www.gws.com.tw (Taiwan)
23
Standards & Web
245
In other words, the distance to the destination was not correlated

with the amount of work required to route the packet to the destination. Therefore, routing based on the distance might not make
sense. However, if we are going to meter based on bandwidth used,
how does ten cents per gigabyte sound? For $20 per month, my
Internet service provider gives me 200 gigabytes of data transfer, an
IP address, a virtualization instance, ten gigabytes of redundant disk
storage, and 360 MB of RAM. I get ten cents per gigabyte, and all
that other stuff for free. 1 gigabyte for ten cents allows you to send
two copies of the text of Encyclopedia Britannica, which would be all
the web traffic many would need for a whole month.
Note that the cost of transmitting packets follows Moore's law
because a router is just a specialized computer. Therefore, the cost
of sending a packet is dropping exponentially. In 18 months, we
should ask for five cents a gigabyte. The nice thing about moving to
such a model is that it would create an incentive for people to offer
ways to use more bandwidth. Cable companies would offer more HD
content if they were getting paid more for the traffic.
246
Da Future
DA FUTURE
Phase II of Bill Gates' Career
The royalty paid to us, minus our expenses of the manual, the
tape and the overhead make Microsoft a break-even operation
today.
Bill Gates, Open Letter to Hobbyists, 1976
Bill to Steve:It was fun while it lasted.
n June 2006, we learned that Bill Gates would step down from
Microsoft in June 2008. This gave plenty of warning for the markets, but it also means that Steve Ballmer will remain as CEO
for a decade or more, the rumors of his demise having been greatly
exaggerated.
One can presume that Bill Gates doesn't believe his legacy is furthered by spending any more time at Microsoft. Given the twin
threats of free software and Google, history may judge Bill Gates
more as an Andrew Carnegie than a Michelangelo.
Microsoft succeeded because it was the company that exploited
Metcalfe's law to its greatest advantage. Microsoft got everyone
Da Future
247
using MS-DOS, which it used to suck customers into Windows,

Office, and everything else. Few other companies had this strategy,
or the resources.
Bill Gates has provided a lot of leadership to the computing industry over the last few decades. In areas from graphical user interfaces to integrated productivity tools, to software as a service, to the
web lifestyle, so his stepping down could be a greater loss to the
industry than even Microsoft's demise. Who else can provide such
vision, something that rallies the entire industry?
That said, Bill hasn't provided too much leadership in recent
years. The last time he was on the cover of Time Magazine, he was
talking about the XBox 360.
Why is Bill not smiling? Perhaps because the XBox 360 is a PC minus a keyboard, web browser and lots of other software.
248
Da Future
There may be several reasons why Bill hasn't remained an iconic

figure for the industry over the last few years. Part of it is that
Microsoft doesn't seem to have anything new to say. HTML has
changed very little in the last six years, and so Microsoft doesn't
have anything to talk about regarding the web.
Secondly, what happens with Microsoft doesn't matter as much to
the computer industry. Lots of people building websites are building
them in PHP, a free programming language. MySQL is the second
most popular database in North America, so all of those users don't
care about what's included in the latest Microsoft SQL Server, or
how it integrates better than ever with the latest Windows. Vista
was a major release, but it has no must-have features, and so the
excitement was muted. (Windows 95's major innovation was 32-bit
computing.) OpenDocument format is a specification almost as
important as HTML, but Microsoft doesn't support it, so Bill can't
talk about it.
Bill Gates signed the ultimate Faustian bargain in the history of
business: proprietary software made Microsoft the most valuable
company ever created, but it was destined to fail because it didn't
adopt an expansive licensing agreement that let its users contribute
back to the system.
If Windows NT had adopted GPL, there would have been no reason to invent Linux. Unix predates Microsoft DOS, and so perhaps
DOS and Windows wouldn't have been invented if those various
Unixes had been GPL from the beginning.
A lot of software has been proprietary since computers were
invented, but it is interesting to wonder where we'd be today if GPL
had been the standard license agreement from the very beginning.
The industry would be very different, and certainly a lot further
along.
Even though free software obviates the need for Microsoft's existence, history may remember Bill Gates for many other things separate from his role in the company. The Gates Foundation has $80
billion to spend which is enough to hire 20,000 skilled workers for
40 years. Spending $80 billion is much harder than it sounds, especially if it is invested in projects which create value, and in turn generate revenues! Bill has the resources to get involved in very big
efforts, including space exploration. (Comedian Dennis Miller said
that Bill Gates is just a monocle and a Persian Cat away from being
a James Bond villain.)
Da Future
249
Space, or How Man Got His

Groove Back
Midnight, July 20, 1969; a chiaroscuro of harsh contrasts
appears on the television screen. One of the shadows moves. It
is the leg of astronaut Edwin Aldrin, photographed by Neil Armstrong. Men are walking on the moon. We watch spellbound.
The earth watches.
Seven hundred million people are riveted to their radios and
television screens on that July night in 1969. What can you do
with the moon? No one knew. Still, a feeling in the gut told us
that this was the greatest moment in the history of life. We
were leaving the planet. Our feet had stirred the dust of an
alien world.
Robert Jastrow, Journey to the Stars
Management is doing things right, Leadership is doing the right
things!
Peter Drucker
SpaceShipOne was the first privately funded aircraft to go into space, and it
set a number of important firsts, including being the first privately funded
aircraft to exceed Mach 2 and Mach 3, the first privately funded manned
spacecraft to exceed 100km altitude, and the first privately funded reusable
spacecraft. The project is estimated to have cost $25 million dollars and
was built by 25 people. It now hangs in the Smithsonian because it serves
no commercial purpose, and because getting into space has never been the
challenge it has always been the expense.
250
Da Future
In the 21st century, more cooperation, better software, and nanotechnology will bring profound benefits to our world, and we will
put the Baby Boomers to shame. I focus only on information technology in this book, but materials sciences will be one of the biggest
tasks occupying our minds in the 21st century and many futurists say
that nanotech is the next (and last?) big challenge after infotech.
I'd like to end this book with one more big idea: how we can
jump-start the nanotechnology revolution and use it to colonize
space. Space, perhaps more than any other endeavor, has the ability
to harness our imagination and give everyone hope for the future.
When man is exploring new horizons, there is a swagger in his step.
Colonizing space will change man's perspective. Hoarding is a
very natural instinct. If you give a well-fed dog a bone, he will bury
it to save it for a leaner day. Every animal hoards. Humans hoard
money, jewelry, clothes, friends, art, credit, books, music, movies,
stamps, beer bottles, baseball statistics, etc. We become very
attached to these hoards. Whether fighting over $5,000 or
$5,000,000 the emotions have the exact same intensity.
When we feel crammed onto this pale blue dot, we forget that any
resource we could possibly want is out there in incomparably big
numbers. If we allocate the resources merely of our solar system to
all 6 billion people equally, then this is what we each get:
Resource
Amount
Hydrogen
34,000 billion Tons
Iron
834 billion Tons
Silicates (sand, glass)
834 billion Tons
Oxygen
34 billion Tons
Carbon
34 billion Tons
Energy production
64 trillion Kilowatts per hour
Even if we confine ourselves only to the resources of this planet,

we have far more than we could ever need. This simple understanding is a prerequisite for a more optimistic and charitable society,
which has characterized eras of great progress. Unfortunately,
NASA's current plans are far from adding that swagger.
Da Future
251
If NASA follows through on its 2004 vision to retire the Space

Shuttle and go back to rockets, and go to the moon again, this is
NASA's own imagery of what we will be looking at on DrudgeReport.com in 2020.
Our astronauts will still be pissing in their space suits in 2020.
According to NASA, the above is what we will see in 2020, but if

you squint your eyes, it looks just like 1969:
All this was done without things we would call computers.
252
Da Future
Only a government bureaucracy can make such little progress in

50 years and consider it business as usual. There are many documented cases of large government organizations plagued by failures
of imagination, yet no one considers that the rocket-scientistbureaucrats at NASA might also be plagued by this affliction. This is
especially ironic because the current NASA Administrator, Michael
Griffin, has admitted that many of its past efforts were failures:
The Space Shuttle, designed in the 1970s, is considered a
failure because it is unreliable, expensive, and small. It costs
$20,000 per pound of payload to put into low-earth orbit
(LEO), a mere few hundred miles up.
The International Space Station (ISS) is small, and only 200
miles away, where gravity is 88% of that at sea-level. It is not
self-sustaining and doesn't get us any closer to putting people on the moon or Mars. (By moving at 17,000 miles per
hour, it falls fast enough to stay in the same orbit.) America
alone spent $100 billion on this boondoggle.
The key to any organization's ultimate success, from NASA to any
private enterprise, is that there are leaders at the top with vision.
NASA's mistakes were not that it was built by the government, but
that the leaders placed the wrong bets. Microsoft, by contrast, succeeded because Bill Gates made many smart bets. NASA's current
goal is flags and footprints, but their goal should be to make it
easy to do those things, a completely different objective.1
I don't support redesigning the Space Shuttle, but I also don't
believe that anyone at NASA has seriously considered building a
next-generation reusable spacecraft. NASA is basing its decision to
move back to rockets primarily on the failures of the first Space
Shuttle, an idea similar to looking at the first car ever built and concluding that cars won't work.
Unfortunately, NASA is now going back to technology even more
primitive than the Space Shuttle. The consensus in the aerospace
industry today is that rockets are the future. Rockets might be in
our future, but they are also in the past. The state-of-the-art in
rocket research is to make them 15% more efficient. Rocket
The Europeans aren't providing great leadership either. One of the big investments of their Space agencies, besides the ISS, is to build a duplicate GPS satellite constellation, which they are doing primarily because of anti-Americanism!
Too bad they don't realize that their emotions are causing them to re-implement
35 year-old technology, instead of spending that $5 Billion on a truly new
advancement. Cloning GPS in 2013: Quite an achievement, Europe!
Da Future
253
research is incremental today because the fundamental chemistry

and physics hasn't changed since their first launches in the mid-20th
century.
Chemical rockets are a mistake because the fuel which propels
them upward is inefficient. They have a low specific impulse,
which means it takes lots of fuel to accelerate the payload, and even
more more fuel to accelerate that fuel! As you can see from the
impressive scenes of shuttle launches, the current technology is not
at all efficient; rockets typically contain 6% payload and 94% overhead. (Jet engines don't work without oxygen but are 15 times more
efficient than rockets.)
If you want to know why we have not been back to the moon for
decades, here is an analogy:
What would taking delivery of this car cost you?
A Californian buys a car made in Japan.
The car is shipped in its own car carrier.
The car is off-loaded in the port of Los Angeles.
The freighter is then sunk.
The latest in propulsion technology is electrical ion drives which
accelerate atoms 20 times faster than chemical rockets, which mean
you need much less fuel. The inefficiency of our current chemical
rockets is what is preventing man from colonizing space. Our simple
modern rockets might be cheaper than our complicated old Space
Shuttle, but it will still cost thousands of dollars per pound to get to
LEO, a fancy acronym for 200 miles away. Working on chemical
rockets today is the technological equivalent of polishing a dusty
turd, yet this is what our esteemed NASA is doing.
254
Da Future
The Space Elevator

When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states
that something is impossible, he is very probably wrong.
Arthur C. Clarke RIP, 1962
The best way to predict the future is to invent it. The future is
not laid out on a track. It is something that we can decide, and
to the extent that we do not violate any known laws of the universe, we can probably make it work the way that we want to.
Alan Kay
A NASA depiction of the space elevator. A space elevator will make it hundreds of times cheaper to put a pound into space. It is an efficiency difference comparable to that between the horse and the locomotive.
Da Future
255
One of the best ways to cheaply get back into space is kicking
around NASA's research labs:
Geosynchronous
Orbit (GEO)
Low Earth Orbit (LEO)

Space Shuttle &
Space Station
Scale picture of the space elevator relative to the size of Earth. The moon is
30 Earth-diameters away, but once you are at GEO, it requires relatively little energy to get to the moon, or anywhere else.
256
Da Future
A space elevator is a 65,000-mile tether upon which we can

launch things into space in a slow, safe, and cheap way.
And these climber don't even need to carry their energy as you
can use solar panels to provide the energy for the climbers. All this
means you need much less fuel. Everything is fully reusable, so
when you have built such a system, it is easy to have daily launches.
The first elevator's climbers will travel into space at just a few
hundred miles per hour a very safe speed. Building a device
which can survive the acceleration and jostling is a large part of the
expense of putting things into space today. This technology will
make it hundreds, and eventually thousands of times cheaper to put
things, and eventually people, into space.
A space elevator might sound like science fiction, but like many of
the ideas of science fiction, it is a fantasy that makes economic
sense. While you needn't trust my opinion on whether a space elevator is feasible, NASA has never officially weighed in on the topic
they haven't given it enough serious consideration.
This all may sound like science fiction, but compared to the technology of the 1960s, when mankind first embarked on a trip to the
moon, a space elevator is simple for our modern world to build. In
fact, if you took a cellphone back to the Apollo scientists, they'd
treat it like a supercomputer and have teams of engineers huddled
over it 24 hours a day. With only the addition of the computing technology of one cellphone, we might have shaved a year off the date of
the first moon landing.
Da Future
257
Carbon Nanotubes
Nanotubes are Carbon atoms in the shape of a hexagon. Graphic created by

Michael Strck.
We have every technological capability necessary to build a space

elevator with one exception: carbon nanotubes (CNT). To adapt a
line from Thomas Edison, a space elevator is 1% inspiration, and
99% perspiration.
Carbon nanotubes are extremely strong and light, with a theoretical strength of three million kilograms per square centimeter; a bundle the size of a few hairs can lift a car. The theoretical strength of
nanotubes is far greater than what we would need for our space elevator; current baseline designs specify a paper-thin, 3-foot-wide ribbon. These seemingly flimsy dimensions would be strong enough to
support their own weight, and the 10-ton climbers using the elevator.
The nanotubes we need for our space elevator are the perfect
place to start the nanotechnology revolution because, unlike biological nanotechnology research, which uses hundreds of different
atoms in extremely complicated structures, nanotubes have a trivial
design.
258
Da Future
The best way to attack a big problem like nanotechnology is to

first attack a small part of it, like carbon nanotubes. A Manhattan
Project on general nanotechnology does not make sense because it
is too unfocused a problem, but such an effort might make sense for
nanotubes. Or, it might simply require the existing industrial expertise of a company like Intel. Intel is already experimenting with nanotubes inside computer chips because metal loses the ability to
conduct electricity at very small diameters. But no one has asked
them if they could build mile-long ropes.
The US government has increased investments in nanotechnology
recently, but we aren't seeing many results. From space elevator
expert Brad Edwards:
Theres whats called the National Nanotechnology Initiative.
When I looked into it, the budget was a billion dollars. But
when you look closer at it, it is split up between a dozen agencies, and within each agency it's split again into a dozen different areas, much of it ends up as $100,000 grants. We looked
into it with regards to carbon nanotube composites, and it
appeared that about thirty million dollars was going into highstrength materials and a lot of that was being spent internally in a lot of the agencies; in the end theres only a couple of
million dollars out of the billion-dollar budget going into something that would be useful to us.
The money doesnt have focus, and its spread out to include
everything. You get a little bit of effort in a thousand different
places. A lot of the budget is spent on one entity trying to play
catch-up with whoever is leading. Instead of funding the leader,
theyre funding someone else internally to catch up.
Again, here is a problem similar to the one we find in software

today: people playing catchup rather than working together. I don't
know what nanotechnology scientists do every day, but it sounds like
they would do well to follow in the footsteps of our free software
pioneers and start cooperating.
The widespread production of nanotubes could be the start of a
nanotechnology revolution. And the space elevator, the killer app of
nanotubes, will enable the colonization of space.
Da Future
Why?
William Bradford, speaking in 1630 of the founding of the Plymouth Bay Colony, said that all great and honorable actions are
accompanied with great difficulties, and both must be enterprised and overcome with answerable courage.
There is no strife, no prejudice, no national conflict in outer
space as yet. Its hazards are hostile to us all. Its conquest
deserves the best of all mankind, and its opportunity for peaceful cooperation may never come again. But why, some say, the
moon? Why choose this as our goal? And they may well ask why
climb the highest mountain? Why, 35 years ago, fly the
Atlantic? Why does Rice play Texas?
We choose to go to the moon. We choose to go to the moon in
this decade and do the other things, not because they are easy,
but because they are hard, because that goal will serve to organize and measure the best of our energies and skills, because
that challenge is one that we are willing to accept, one we are
unwilling to postpone, and one which we intend to win, and the
others, too.
It is for these reasons that I regard the decision last year to
shift our efforts in space from low to high gear as among the
most important decisions that will be made during my incumbency in the office of the Presidency.
In the last 24 hours we have seen facilities now being created
for the greatest and most complex exploration in man's history.
We have felt the ground shake and the air shattered by the testing of a Saturn C-1 booster rocket, many times as powerful as
the Atlas which launched John Glenn, generating power equivalent to 10,000 automobiles with their accelerators on the floor.
We have seen the site where five F-1 rocket engines, each one
as powerful as all eight engines of the Saturn combined, will be
clustered together to make the advanced Saturn missile, assembled in a new building to be built at Cape Canaveral as tall as a
48 story structure, as wide as a city block, and as long as two
lengths of this field.
The growth of our science and education will be enriched by
new knowledge of our universe and environment, by new techniques of learning and mapping and observation, by new tools
and computers for industry, medicine, the home as well as the
school.
I do not say the we should or will go unprotected against the
hostile misuse of space any more than we go unprotected
against the hostile use of land or sea, but I do say that space
can be explored and mastered without feeding the fires of war,
without repeating the mistakes that man has made in extending
his writ around this globe of ours.
We have given this program a high national priority even
though I realize that this is in some measure an act of faith and
vision, for we do not now know what benefits await us. But if I
259
260
Da Future
were to say, my fellow citizens, that we shall send to the moon,
240,000 miles away from the control station in Houston, a giant
rocket more than 300 feet tall, the length of this football field,
made of new metal alloys, some of which have not yet been
invented, capable of standing heat and stresses several times
more than have ever been experienced, fitted together with a
precision better than the finest watch, carrying all the equipment needed for propulsion, guidance, control, communications, food and survival, on an untried mission, to an unknown
celestial body, and then return it safely to earth, re-entering the
atmosphere at speeds of over 25,000 miles per hour, causing
heat about half that of the temperature of the sun almost as
hot as it is here today and do all this, and do it right, and do
it first before this decade is out then we must be bold.
John F. Kennedy, September 12, 1962
Lunar Lander at the top of a rocket. Rockets are expensive and impose significant design constraints on space-faring cargo.
NASA has 18,000 employees and a $16-billion-dollar budget. Even

with a fraction of those resources, their ability to oversee the
design, handle mission control, and work with many partners is
more than equal to this task.
If NASA doesn't build the space elevator, someone else might, and
it would change almost everything about how NASA does things
today. NASA's tiny (15-foot-wide) new Orion spacecraft, which was
built to return us to the moon, was designed to fit atop a rocket and
return the astronauts to Earth with a 25,000-mph thud, just like in
the Apollo days. Without the constraints a rocket imposes, NASA's
Da Future
261
spaceship to get us back to the moon would have a very different

design. NASA would need to throw away a lot of the R&D they are
now doing if a space elevator were built.
Another reason the space elevator makes sense is that it would
get the various scientists at NASA to work together on a big, shared
goal. NASA has recently sent robots to Mars to dig two-inch holes in
the dirt. That type of experience is similar to the skills necessary to
build the robotic climbers that would climb the elevator, putting
those scientists to use on a greater purpose.
Space debris is a looming hazard, and a threat to the ribbon:
Map of space debris. The US Strategic Command monitors 10,000 large

objects to prevent them from being misinterpreted as a hostile missile.
China blew up a satellite in January, 2007 which created 35,000 pieces of
debris larger than 1 centimeter.
The space elevator provides both a motive, and a means to launch

things into space to remove the debris. (The first elevator will need
to be designed with an ability to move around to avoid debris!)
Once you have built your first space elevator, the cost of building
the second one drops dramatically. A space elevator will eventually
262
Da Future
make it $10 per pound to put something into space. This will open
many doors for scientists and engineers around the globe: bigger
and better observatories, a spaceport at GEO, and so forth.
Surprisingly, one of the biggest incentives for space exploration is
likely to be tourism. From Hawaii to Africa to Las Vegas, the primary
revenue in many exotic places is tourism. We will go to the stars
because man is driven to explore and see new things.
Space is an extremely harsh place, which is why it is such a miracle that there is life on Earth to begin with. The moon is too small to
have an atmosphere, but we can terraform Mars to create one, and
make it safe from radiation and pleasant to visit. This will also teach
us a lot about climate change, and in fact, until we have terraformed
Mars, I am going to assume the global warming alarmists don't
really know what they are talking about yet.2 One of the lessons in
engineering is that you don't know how something works until
you've done it once.
Terraforming Mars may sound like a silly idea today, but it is simply another engineering task.3 I worked in several different groups
at Microsoft, and even though the set of algorithms surrounding
databases are completely different from those for text engines, they
are all engineering problems and the approach is the same: break a
problem down and analyze each piece. (One of the interesting
lessons I learned at Microsoft was the difference between real life
and standardized tests. In a standardized test, if a question looks
hard, you should skip it and move on so as not to waste precious
time. At Microsoft, we would skip past the easy problems and focus
our time on the hard ones.)
Engineering teaches you that there are an infinite number of
ways to attack a problem, each with various trade-offs; it might take
1,000 years to terraform Mars if we were to send one ton of material, but only 20 years if we could send 1,000 tons of material. Whatever we finally end up doing, the first humans to visit Mars will be
happy that we turned it green for them. This is another way our generation can make its mark.
2
Carbon is not a pollutant and is valuable. It is 18% of the mass of the human body,
but only .03% of the mass of the Earth. If Carbon were more widespread, diamonds would be cheaper. Driving very fast cars is the best way to unlock the carbon we need. Anyone who thinks are running out of energy doesn't understand
the algebra in E = mc2.
Mars' moon, Phobos, is only 3,700 miles above Mars, and if we create an atmosphere, it will slow down and crash. We will need to find a place to crash the fragments, I suggest in one of the largest canyons we can find; we could put them
next to a cross dipped in urine and call it the largest man-made art.
Da Future
263
A space elevator is a doable mega-project, but there is no

progress beyond a few books and conferences because the very
small number of people on this planet who are capable of initiating
this project are not aware of the feasibility of the technology.
Brad Edwards, one of the world's experts on the space elevator,
has a PhD and a decade of experience designing satellites at Los
Alamos National Labs, and yet he has told me that he is unable to
get into the doors of leadership at NASA, or the Gates Foundation,
etc. Setting aside the nanotechnology, we might need only 5,000
man-years of work (NASA alone has 50,000 employees) to accomplish this task, but no one who has the authority to organize this
understands that a space elevator is doable.
Glenn Reynolds has blogged about the space elevator on his very
influential Instapundit.com, yet a national dialog about this topic has
not yet happened, and NASA is just marching ahead with its expensive, dim ideas. My book is an additional plea: one more time, and
with feeling!
How and When

It does not follow from the separation of planning and doing in
the analysis of work that the planner and the doer should be
two different people. It does not follow that the industrial world
should be divided into two classes of people: a few who decide
what is to be done, design the job, set the pace, rhythm and
motions, and order others about; and the many who do what
and as they are told.
Peter Drucker
There are a many interesting details surrounding a space elevator, and for those interested in further details, I recommend The
Space Elevator, co-authored by Brad Edwards.
The size of the first elevator is one of biggest questions to resolve.
If you were going to lay fiber optic cables across the Atlantic ocean,
you'd set aside a ton of bandwidth capacity. Likewise, the most
important metric for our first space elevator is its size.
The one other limitation with current designs is that they assume
climbers which travel hundreds of miles per hour. This is a fine
speed for cargo, but it means that it will take days to get into orbit.
If we want to send humans into space in an elevator, we need to
build climbers which can travel at 10,000 miles per hour. While this
seems ridiculously fast, if you accelerate to this speed over a period
of minutes, it will not be jarring. Perhaps this should be the challenge for version two if they can't get it done the first time.
264
Da Future
The conventional wisdom amongst those who think it is even possible is that it will take between 20 and 50 years to build a space
elevator. However, anyone who makes such predictions doesn't
understand that engineering is a fungible commodity. Two people
will, in general, accomplish something twice as fast as one person.4
How can you say something will unequivocally take a certain
amount of time when you don't specify how many resources it will
require or how many people you plan to assign to the task?
Efficiency drops as teams gets larger, but with the Internet, a
superb tool of communication and collaboration, thousands of people can work together efficiently. Manufacturers are using the Internet to reinvent how they interact with their suppliers during the
design process shaving years off the design. Boeing's newest 787
Dreamliner airplane will go from project start to take-off in just
four years, which is half the time it took them to design the 777.
(Why Boeing offered their 30-year old 767 instead of the 787 to the
US Military, as the workhorse of their new refueling fleet, was an
insanely backward-looking mistake. If Boeing doesn't think their
improvements are worthwhile, why are they making them?)
Furthermore, predictions are usually way off. If you asked someone how long it would take unpaid volunteers to make Wikipedia as
big as the Encyclopedia Britannica, no one would have guessed the
correct answer of two and a half years. From creating a space elevator to world domination by Linux, anything can happen in far less
time than we think is possible if everyone simply steps up to play
their part. The way to be a part of the future is to invent it, by
unleashing our scientific and creative energy towards big, shared
goals. Wikipedia, as our encyclopedia, was an inspiration to millions
of people, and so the resources have come piling in. The way to get
help is to create a vision that inspires people.
In a period of 75 years, man went from using horses and wagons
to landing on the moon. Why should it take 30 years to build something that is 99% doable today?
Many of the components of a space elevator are simple enough
that college kids are building prototype elevators in their free time.
The Elevator:2010 contest is sponsored by NASA, but while these
Fred Brooks' The Mythical Man-Month argues that adding engineers late to a
project makes a project later, but ramp-up time is just noise in the management of
an engineering project. Also, wikis, search engines, and other technologies
invented since his book have lowered the overhead of collaboration.
Da Future
265
contests have generated excitement and interest in the press, they

are building toys, much like a radio-controlled airplane is a toy compared to a Boeing airliner.
I believe we could have a space elevator built in 7 years. If you
divvy up three to four years of work per person, and add in some
time to ramp up and test, you can see how seven years is quite reasonable. Man landed on the moon 7 years after Kennedy's speech,
exactly as he ordained, because dates can be self-fulfilling prophecies. It allows everyone to measure themselves against their goals,
and determine if they need additional resources.
If the design of the hardware and the software were done in a
public fashion, others could take the intermediate efforts and test
them and improve them, therefore saving further engineering time.
Perhaps NASA could come up with hundreds of truly useful research
projects for college kids to help out on instead of encouraging them
to build toys.
The Unknown Unknown is the nanotubes, but nearly all the other
pieces can be built without having any access to them. We will only
need them wound into a big spool on the launch date.
I can imagine that any effort like this would get caught up in a
tremendous amount of international political wrangling that could
easily add years on to the project. We should not let this happen,
and we should remind each other that the space elevator is just the
railroad car to space the exciting stuff is the cargo inside and the
possibilities out there. A space elevator is not a zero sum endeavor:
it would enable lots of other big projects that are totally unfeasible
currently. A space elevator would enable various international space
agencies that have money, but no great purpose, to work together
on a large, shared goal. And as a side effect it would strengthen
international relations.5
Perhaps the Europeans could build the station at GEO. Russia could build the
shuttle craft to move cargo between the space elevator and the moon. The Middle
East could provide an electrical grid for the moon. China could take on the problem of cleaning up the orbital space debris and build the first moon base. Africa
could attack the problem of terraforming Mars, etc.
266
Da Future
21st Century Renaissance

The great achievements of western capitalism have rebounded
primarily to the benefit of the ordinary person.
Milton Friedman
If we have learned one thing from the history of invention and
discovery, it is that, in the long run, and often in the short one,
the most daring prophecies seem laughably conservative.
Arthur C. Clarke, 1951
As I see it, I still live in the 20th century. History will remember
the 21st century as the time when man entered a new Renaissance,
and when this was not just a photo shoot.
If we could replace journalists with robots, the issue of media bias would
disappear, and they would become better looking.
Even in our still-primitive world of today, I would rather be making $30,000 a year than be promised $100 million dollars if I walked
through a door back into 1986, and I say that not just because of the
big hair. The Internet was still seven years away from its first web
page in 1986. We might not be able to go back in time, but in general we wouldn't want to.
Many people don't appreciate how much faster the world is moving every day with the creation of the Internet and other modern
Da Future
267
technologies. Some say: The only constant is change, but this is

wrong: the only constant is acceleration, an increasing rate of
change. People don't feel the acceleration yet because the world is
moving so slowly. Energy prices are high because we haven't built a
nuclear power plant in 30 years (blame the Democrats), and building nuclear power plants supposedly takes 15 years (blame the
bureaucrats). It took decades to roll out HDTV. Genetically modified
foods are treated as Frankensteins of nature. Drug companies hire
as many lawyers as researchers, and then we wonder why drug
costs are so high, and why it takes years to release new ones.
Devices are disintegrated and stupid. Waiting in line is considered a
normal part of life.
The gap between the rich and the poor doesn't cause social instability because Bill Gates doesn't live materially better than anyone
else. His contact lenses are no better than mine, and neither of us
owns a robot-driven car.
With better cooperation and better tools, the 21st century Renaissance is waiting for us. The world there will be characterized by a
gift economy, so resist the urge to hoard. Once we cross that line,
we will get our sanity and happiness back. Let's hurry!
Do not fear the future; man was created to solve problems and
gain enjoyment from that process. The question as to whether the
United States will be relevant in the 21st century depends on the
responses to these simple questions: will it have the most scientists,
and are they learning from each other, and therefore working
together?
The End
268
Da Future
Warning Signs From the Future
This is our scary future, and unfortunately mankind is crawling towards it.6
From http://lifeboat.com/ex/warning.signs.for.tomorrow, created by Anders Sandberg. I sent multiple e-mails to various e-mail addresses at the Lifeboat Foundation to try to get permission to use these images, but I never received a response.
Perhaps they are too busy with their mission of encouraging scientific advancements to help humanity survive existential risks to respond. So I just donated
$200 for their use here :-)
Da Future
We go forward with complete confidence in the eventual triumph of freedom. Not because history runs on the wheels of
inevitability; it is human choices that move events. Not because
we consider ourselves a chosen nation; God moves and chooses
as He wills. We have confidence because freedom is the permanent hope of mankind.
When our Founders declared a new order of the ages; when soldiers died in wave upon wave for a union based on liberty;
when citizens marched in peaceful outrage under the banner
Freedom Now they were acting on an ancient hope that is
meant to be fulfilled. History has an ebb and flow of justice, but
history also has a visible direction, set by liberty and the Author
of Liberty.
US Presidential Inaugural Address, 2005
Genius is eternal patience.
The greater danger for most of us lies not in setting our aim too
high and falling short; but in setting our aim too low, and
achieving our mark.
The true work of art is but a shadow of the Divine perfection.
Michelangelo
Most of the things worth doing in the world had been declared
impossible before they were done.
Louis Brandeis, US Supreme Court Justice
We have only one alternative: either to build a functioning
industrial society or see freedom itself disappear in anarchy
and tyranny.
Peter Drucker
The mind is not a vessel to be filled but a fire to be kindled.
Plutarch
269
270
Afterword
AFTERWORD
US v. Microsoft
The political principle that underlies the market mechanism is

unanimity. In an ideal free market resting on private property,
no individual can coerce any other, all cooperation is voluntary,
all parties to such cooperation benefit or they need not participate. There are no values, no social responsibilities in any
sense other than the values and responsibilities of individuals.
Society is a collection of individuals and of the various groups
they voluntarily form.
If an exchange between two parties is voluntary, it will not take
place unless both believe they will benefit from it. Most economic fallacies derive from the neglect of this simple insight,
from the tendency to assume that there is a fixed pie, that one
party can only gain at the expense of another.
Milton Friedman
Microsoft got its tush handed to it in the DOJ trial, but that was
because it lost credibility. For example, Bill Gates argued that he
wasn't worried about Netscape. If so, why did Microsoft VIPs say
they wanted to smother Netscape and cut off their air supply?
Judge Jackson wrote that Microsoft's witnesses in the trial: proved,
time and time again, to be inaccurate, misleading, evasive, and
transparently false. However, what they say at a trial has nothing
to do with how they behaved in the marketplace.
Fisher, the government's economist argued that it was a joke
that Windows was challenged by the Macintosh or Linux. With just a
few years hindsight, and the tremendous potential of free software,
it becomes very clear that their economist was wrong. Furthermore,
in the trial he admitted that he could find no specific harm that
Microsoft had done. Microsoft might be a hard-charging competitor,
but the software they wrote has been invaluable to the world, and
they provided it at a cost much lower than many of their competitors
like Sun, IBM and Oracle.
In addition to the Mac and Linux, the web was and is a huge
threat to Microsoft the PC development community used to
revolve around Windows apps, but now it is all about web apps.
The government's major argument was that bundling a web
browser into the operating system was illegal. However, including a
browser with an operating system is a good idea for consumers.
Without a web browser, you wouldn't be able to surf the web or send
Afterword
271
e-mail. Piecing together an operating system as a jumble of parts is

a lot of work and should not be done by the end user. Every operating system today includes a web browser. If the government had its
way, progress in Microsoft's software would be frozen with early
1990's technology.
Furthermore, the government never had a clear proposed remedy.
The trial court's plan was to split Microsoft up into Windows, Office,
and other parts, which would have be an unprecedented intrusion
into a private company, and yet wouldn't have decreased the popularity of Windows or Office.
The fawning and shallow press focused more on David Boies brilliance and ability to quote obscure legislation from memory than the
important facts and issues of the case he was making. Almost everything Microsoft did was done by other people in the industry. It was
only because Microsoft became the biggest company on the block
that it became illegal. Microsoft was a monopoly, and therefore was
behaving badly, even when it was acting the same as everyone else.
And, no one could describe when Microsoft had crossed that line.
Microsoft succeeded because it built the best products. That Sun,
Netscape, Apple and other companies, who at the time were just as
proprietary as Microsoft, complained is disingenuous. They were
jealous and afraid, and used the government to help them. As
Richard Stallman points out, all proprietary technologies are monopolies; Microsoft was just the most popular one.
Free software, better programming languages, standard formats,
and the Internet have long been a threat to Microsoft, it just has not
caused Microsoft's revenue to decrease, yet. Milton Friedman wrote
that the only kind of monopoly that is a problem is a coercive
monopoly which are created by government. In this case, there is
no need to take an offending company to trial, but to remove the
government regulation which created the monopoly in the first
place. It would have been interesting if Microsoft would have made
only this argument, and then rested.
272
Afterword
Microsoft as a GPL Software

Company
Prediction is very difficult, especially about the future.
Niels Bohr
In most organizations, change comes in only two flavors: trivial
and traumatic. Review the history of the average organization
and youll discover long periods of incremental fiddling punctuated by occasional bouts of frantic, crisis-driven change.
Gary Hamel
I wrote this book under the assumption that if Microsoft continues on its current path and never adopts copyleft licenses for its
code, what Wikipedia did to Encarta will likewise happen to Windows, Office, Internet Explorer, SQL Server, Visual Studio,
Exchange, MSN, etc.
A tiny fraction of the software Microsoft produces runs on Linux,
so Linux world domination means the end of Microsoft as we know
it. Free software will force Microsoft to choose between licensing
revenue and relevance.
A lot of people that I meet in the software industry hate Microsoft, but have valued its leadership. Many people recognize that
Office is more polished than OpenOffice and pioneered many innovations. Internet Explorer was universally recognized for a significant period of time as a better browser than Netscape's. Microsoft's
SQL Server is considered easier to use than Oracle's. Word beat Ami
Pro and WordPerfect in the reviews.
Let us presume that Microsoft were to embrace the proposition
that free software software is a good idea. There would be no reason
to call Microsoft evil anymore, because copyleft would prevent it.
When something is free, no one can control or monopolize it anymore. In adopting free software, they would have two possibilities
available to it, however, both involve a dramatic change in what
their employees do, and a significant drop in revenue.
Microsoft could release all of its code as GPL, learn to build this
software as a globally distributed effort, and compete and better
interoperate with other existing free software. I think this would be
a mistake because Microsoft's codebases are too old, much bigger
and more complicated than existing free software codebases, and
the details today are not understood by the existing globally distributed software community.
Afterword
273
If Microsoft were to release its code as GPL it isn't clear if the

existing free software community would switch, and would create
confusion and further fragmentation, an inefficiency which is
already a huge problem. Therefore I will say little more about this
possibility.
Even more dramatic is for Microsoft to adopt Linux and other free
software. Microsoft could easily become the dominant Linux player.
Mark Shuttleworth invested only a few million dollars to create
Ubuntu and become the most popular Linux distribution, so imagine
what Microsoft's billions and army could do? Microsoft could even
develop mechanisms to make it easy for people and enterprises to
move from Windows to Linux, a challenge they are uniquely able to
tackle.
This transition would involve a complete upheaval of the codebase almost every Microsoft employee works on, and would require
they find an alternate source of revenue for their most profitable
products. But they would likely find a way to employ all their existing engineers and using just their cash on hand, they could hire
every one of their engineers for 5 years; in other words, they would
have many years to sort out the revenue situation during the transition.
As there is no free vision recognition, speech, or search engine
with any critical mass, Microsoft could lead those efforts and lead
the future even more dominantly than it owns the present! With the
rest of the world helping out, the value of these offerings would be
much greater than the proprietary set that Microsoft is currently
offering today.
One way Microsoft could jumpstart this transformation would be
to purchase Novell or Ubuntu, or just by hiring Debian developers to
help seed their existing teams and to cross-pollinate the culture. In
early-2008, Novell's market capitalization was $2.2 billion, a mere
5% of the value of their recent acquisition target Yahoo!, and this is
the most expensive way to jumpstart work on Linux.1
My one concern about this idea is that after criticizing Microsoft
for: building a kernel inferior to Linux, for shipping too infrequently,
for not building software with the same reliability that Boeing builds
into their airplanes, for not sufficiently embracing standards, for
being crippled by backward compatibility that maybe it wouldn't
1
One advantage of buying Novell is that Microsoft would get Mono, the free .Net
runtime, and it could use it as a base to build the next-generation programming
language, supplanting Sun. However, this would eventually involve merging it
with their own .Net codebase.
274
Afterword
work out that well. Furthermore, I made a switch to zero Microsoft

code smoothly and I believe the rest of the world can as well. Therefore if Microsoft made many mistakes, and we don't need them,
should we entrust the future to them?
However, if Microsoft were to embrace the GPL, better tools, shipping more frequently, having one tree the world can collaborate in
to evolve more smoothly, and with a greater emphasis on open standards created in parallel with code, their next stack would be a lot
better than their current one.
These changes would allow Microsoft to not necessarily accumulate the baggage they've acquired in their current software, but I
would just like to repeat that the most common reason why code is
not reliable is that it is too complicated.2 Whether that complexity is
caused by backward compatibility, tools, Intelligent Design, shipping
infrequently, too much of a focus on performance rather than code
cleanliness, age, or the number of customers, I cannot say, but it is a
problem that the rest of the free software community should learn
from. Complexity is not only the enemy of reliability, it is the enemy
of progress.
This is idle speculation if Microsoft does not adopt free software.
For example, SQL Server has almost an entire OS inside it with its own memory
manager, cache manager, file system, synchronization mechanisms, threads, interprocess communication, code loading, security subsystem, etc. and it still has
multithreading bugs. Huge portions of .Net are written in C++ because they
thought it would be a few percent faster, but this also has added complexity and
bugs and slows progress.
Afterword
275
The Outside World

In 1929, cars were unreliable and belched all sorts of noxious
fumes into the atmosphere. Computers did not exist. Airplanes
were slow and unsafe. Movies had just recently gotten sound.
TV and FM radio did not exist. The iPod and cell phone were
misspellings. DNA was just three letters and genetics non-existent. The stock market crashed and we were on our way to a
depression.
In 2009, cars are reliable and produce little noxious gases.
There are billions of computers in the world, especially when
you count the ones in cars, thermostats, cell phones, etc. Airplanes are fast, safe, and affordable. Home entertainment centers have surround sound and quality so good you think you are
in the movie theater. TV is digital and FM is everywhere. Close
to 200 million iPods and billions of cell phones have been sold.
The human genome has been sequenced multiple times...and
the stock market crashed. Many argue that we are headed for a
deep recession, if not depression. It seems more than obvious
to me that we need to import some of the methods of knowledge transfer from the technologists into the rest of our life
structures.
Seymour Friedel
Congress, the press, and the bureaucracy too often focus on
how much money or effort is spent, rather than whether the
money or effort actually achieves the announced goal.
Donald Rumsfeld, 1974
Just as unskilled manual workers in manufacturing were the
dominant social and political force in the 20th century, knowledge technologists are likely to become the dominant social
and perhaps also political force over the next decades.
Peter Drucker
The challenge for my generation was to provide an intellectual
defense of economic liberty. The challenge for the current generation is to keep it.
Milton Friedman
Eben Moglen says that one of the goals of the free software movement is to expand opportunities for billions more people out there
to quit throwing away most of the brains on earth. This book is
about free software, but I'd like to end this afterword with a few
ideas on a free press, free markets and several other issues. If it
weren't for our scientists and engineers, we'd still be picking our
noses is caves, and the lawyers in divorce proceedings would be fervently arguing about how to divide up the rocks. Our scientists can
tell us that one pound of Uranium generates the same amount of
276
Afterword
power as three million pounds of Coal, but whether we choose to

use that fact is a decision of the Lawyers. It's the outside world
which determines whether we will put our scientists to work, sue
them, or starve them. High taxes might have the benefit of taking
money away from the evil corporations, but they also leave less
money to spend on R&D.
Powering a free society

After a quarter-century of gas tax hikes, a ban on drilling for oil
and a complete destruction of the nuclear power industry in
America, I guess liberals can declare: Mission accomplished!
In response to skyrocketing gas prices, they say, practically in
unison, We can't drill our way out of this crisis.
What does that mean? This is like telling a starving man, You
can't eat your way out of being hungry! Finding more oil isn't
going to increase the supply of oil?
It is the typical Democratic strategy to babble meaningless slogans, as if they have a plan. Their plan is: the permanent twilight of the human race. It's the only solution they can think of
to deal with the beastly traffic on the LIE (Long Island Expressway).
Liberals complain that as Barack Obama put it there's no
way that allowing offshore drilling would lower gas prices right
now. At best you are looking at five years or more down the
road.
This is as opposed to airplanes that run on woodchips, which
should be up and running any moment now. Moreover, what
was going on five years ago? Why didn't anyone propose
drilling back then?
Say, you know what we need? We need a class of people paid to
anticipate national crises and plan solutions in advance. It
would be such an important job, the taxpayers would pay them
salaries so they wouldn't have to worry about making a living
and could just sit around anticipating crises.
If only we had had such a group let's call them elected representatives they could have proposed drilling five years
ago!
But of course we do pay people to anticipate national problems
and propose solutions. Some of them we'll call them Republicans did anticipate high gas prices and propose solutions.
Jim Bangs
I went through thick and thin in the nuclear plant licensing
process. The Shoreham, New York, plant was in operation. It
got up to well over 1% power, and then they shut it down.
Democrat Governor Mario Cuomo made 'em drill holes in the
reactor vessel so it could never be used again. And it was the
Afterword
277
first plant that had been built under the newest and safest regulations, and it took them years. It cost us well over a billion
dollars, and they just wrote it off.
Jim, Shoreham Nuclear Regulatory Commission Project Manager
The Alvin W. Vogtle Nuclear Generating Station in Burke
County, Georgia, has two 1,200-MW reactors sitting on the
Savannah River, directly across from the federal nuclear processing facilities in South Carolina. Now Southern Nuclear,
which owns Vogtle, wants to build two new 1,000-MW reactors
as part of the nuclear renaissance.
Environmental groups have immediately taken up the challenge, arguing that dredging the Savannah River to allow barge
delivery of reactor parts will damage the river. The Savannah
was dredged regularly for more than a century until the Army
Corps of Engineers gave up in 1980 because nothing much was
happening on the river. Now environmental groups say a
renewal will ruin the environment. The Nuclear Regulatory
Commission has nodded agreement and will require an environmental impact statement before early site clearance can begin.
That will probably add three years to the project.
William Tucker
Leonardo Cavallaro, a man who designed nuclear power plants

until the mid-70s when the industry collapsed, told me that Energy
changes everything.3 Each invention which decreased the amount
of energy man himself had to expend to accomplish a task increased
his quality of life. The use of the animals, water wheels to mill grain,
steam, electricity and batteries have each made our lives easier. Few
things would further increase our quality of life more than making
energy ten times cheaper than it is today.
One of the biggest reasons for offshoring of manufacturing jobs
outside of the US is the cost of energy. It takes a lot of energy to
bend steel! The idea that it was okay to offshore manufacturing
because America is moving to a service economy is one of the dumbest ideas ever uttered because it doesn't recognize that it isn't until
you do something that you figure out how to do it.
An interesting article explaining how nuclear power got killed in the United
States is located at http://tinyurl.com/WhoKilledNuclearPower.
278
Afterword
Every step that you take both solves today's problems, and sets
you up to take on new problems. At Microsoft, we had a phrase
Crawl, Walk, Run and it applies to other sectors as well as software:
Make TVs -> Make robots which make TVs -> Make robots
which make your bed
99 cent plastic toy -> materials sciences for toys -> materials
sciences changing all building materials
Food production -> genetically modified food -> genetically
modified everything
In 1996, President Clinton vetoed legislation to increase drilling

for oil inside the US because he said it wouldn't do anything for 10
years. 10 years later, with record gas prices, many in Congress said
they are against it because it won't do anything for 10 years. Either
people in our government are incapable of thinking long-term, or
they have been bribed by special interests and treat us like fools.
Afterword
279
Free Markets
Congress creates the problem, blames the free market, and uses the crisis
as an excuse to create more government.
A man's admiration for absolute government is proportionate to
the contempt he feels for those around him.
Alexis de Tocqueville
Nobody spends somebody else's money as wisely as he spends
his own.
A major source of objection to a free economy is precisely that
it gives people what they want instead of what a particular
group thinks they ought to want. Underlying most arguments
against the free market is a lack of belief in freedom itself.
Everybody agrees that socialism has been a failure. Everybody
agrees that capitalism has been a success...yet everybody is
extending socialism!
Spending by government in 1989 amounts to about 45 percent
of national income. By that test, government owns 45 percent
of the means of production that produce the national income.
The U.S. is now 45 percent socialist.
Milton Friedman
280
Afterword
I predict future happiness for Americans if they can prevent the
government from wasting the labors of the people under the
pretense of taking care of them.
Thomas Jefferson
While recovering from an emergency ruptured appendix operation in a Libyan state-run, universal health-care clinic in Tripoli,
an exasperated doctor lamented to me that the fellow who was
mopping the floor beside the bed by fiat made exactly what he
did.
Victor Davis Hanson
If I were designing a health care system from scratch, I would
probably go ahead with a single-payer system.
Barack Obama
So you think that money is the root of all evil? Have you ever
asked what is the root of money? Money is a tool of exchange,
which can't exist unless there are goods produced and men
able to produce them. Money is the material shape of the principle that men who wish to deal with one another must deal by
trade and give value for value.
Where freedom is real, equality is the passion of the masses.
Where equality is real, freedom is the passion of a small minority.
Eric Hoffler, American philosopher
There is a correlation between the free market and free software,

as both are powered by innovation created by millions of disconnected individuals, acting in their own self-interest scratching
their own itch. The collaborative process in free software which
weeds out the bad ideas is very similar to what happens in the free
market. Innovation might be disruptive, but the alternative is stagnation, bad in both software and life.
Healthcare is the political topic du jour, and many advocate the
government involvement in creating a public option to increase
choice and competition in healthcare. But in the Microsoft
antitrust trial, the government never argued the way to increase
competition in software would be for the government to get into the
business, and if they had, they would have been laughed out of the
courtroom! What America the world's only superpower in 200 years
was that it had the freest economy. History has demonstrated
repeatedly that companies innovate much faster than the government; the free market is us. Government doesn't innovate because it
isn't set up to do that; every expense is a pre-planned line item in
the budget. No government employee can go to work one day and
Afterword
281
ask their boss for some extra money so they can develop a new idea
to make more money. In fact, working for such organizations must
drain the humanity from its workers.
Many think that the free market is evil because the spoils are
unequally distributed. What they are missing is that is something
that Milton Friedman wrote:
Industrial progresses, mechanical improvement, all of the great
wonders of the modern era have meant relatively little to the
wealthy. The rich in Ancient Greece would have benefited
hardly at all from modern plumbing: running servants replaced
running water. Television and radio, the Patricians of Rome
could enjoy the leading musicians and actors in their home,
could have the leading actors as domestic retainers. Ready-towear clothing, supermarkets all these and many other modern
developments would have added little to their life. The great
achievements of Western Capitalism have redounded primarily
to the benefit of the ordinary person. These achievements have
made available to the masses conveniences and amenities that
were previously the exclusive prerogative of the rich and powerful.
Today, we are told to despise many types of corporations, from

energy to health insurance. This ignores the social utility they create in employing people and providing value to customers. All transactions in a free market are voluntary; whereas if you don't like a
product or service your government is providing, there is no way to
opt out of paying for it. It is inexplicable that so many, especially
geeks who are by nature anti-establishment, want to their hand
their their property and liberty over to unaccountable bureaucrats
who run a monopoly that Microsoft would jealous of.
Free market capitalism is also a human rights issue: it is about
the right to make and purchase what one wishes and to keep the
fruits of one's labor. Even the most egalitarian programmer would
never suggest that Linus should earn the same amount as a newbie
programmer. Profits have to be earned in a free market, and it generates more innovation and competition. A recent newspaper headline proclaimed that there is a shortage of family doctors in
America. The free market can handle this: those doctors can raise
their fees, which will also encourage more doctors to go into this
specialty. This is the invisible hand described in Adam Smith's The
Wealth of Nations, written way back in 1776.
Greed is not bad it is what gets people out of bed in the morning. The dream of making a million dollars for someone else doesn't
282
Afterword
provide the same motivation as making it for yourself! Supporters of

the free market aren't against welfare nets, but they they are
against a welfare state and punishing success.
Here is a quote from Mark Levin's Liberty and Tyranny:
The Statist is constantly agitating for government action. And
in furtherance of that purpose, he speaks in the tongue of the
demagogue, concocting one pretext and grievance after
another to manipulate public perceptions and build popular
momentum for the divestiture of liberty and property from its
rightful possessors. The industrious, earnest, and successful
are demonized as perpetrators of various offenses against the
public good, which justifies governmental intervention on
behalf of the endless parade of victims. In this way, the perpetrator and the victim are subordinated to the government's
authority the former by outright theft, the latter by dependent existence.
The Statist veils his pursuits in moral indignation, intoning the
injustices and inequities of liberty and life itself, for which only
he can provide justice and bring a righteous resolution. And
when the resolution proves elusive, as it undoubtedly does
whether the Marxist promise of the workers' paradise or the
Great Society war on poverty the Statist demands even
more authority to wring out the imperfections of mankind's
existence. Failure is not the product of his beliefs but merely
want of power and resources. Thus are born endless rationalizations for seizing ever more governmental authority.
Paul Krugman, who won a Nobel Prize in economics, wrote: If

the rich get richer, the poor and middle-class must get poorer as a
matter of arithmetic. It is stunning that he doesn't understand that
innovation brings new wealth to a society. The sad part about
today's intelligentsia is that they do not accept even decades-old
economic theories like how the free market generates the most
prosperous society.
Nothing would improve our world more than having more people
familiar with the benefits of the free market, a concept as subtle as
free software. Anyone who thinks their government ought to do
more than enforce the border, create an army and a provide a basic
welfare net should read Milton Friedman's Free to Choose. His writings suggest profound implications for our government which is
steadily intruding itself into every aspect of our lives from cradle to
grave. We citizens should read his work, and demand our elected
representatives implement the ideas contained therein. If a medical
doctor didn't follow scientific advancements of the last 40 years, he
would be sued for malpractice, but somehow in government, economic malpractice is allowed!
Afterword
The Legislature
The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design.
Before the obvious economic failure of Eastern European
socialism, it was widely thought that a centrally planned economy would deliver not only social justice but also a more efficient use of economic resources. This notion appears eminently
sensible at first glance. But it proves to overlook the fact that
the totality of resources that one could employ in such a plan is
simply not knowable to anybody, and therefore can hardly be
centrally controlled.
Frederick Hayek, The Fatal Conceit
Knowledge is one of the scarcest of all resources in any economy. Even when leaders have much more knowledge and
insight than the average member of the society, they are
unlikely to have nearly as much knowledge and insight as exists
scattered among millions of people subject to their governance.
Thomas Sowell, Basic Economics
To provide for us in our necessities is not in the power of Government. It would be a vain presumption in statesmen to think
they can do it.
Edmund Burke, 1795
Of all tyrannies, a tyranny sincerely exercised for the good of
its victims may be the most oppressive. It would be better to
live under robber barons than under omnipotent moral busybodies. The robber baron's cruelty may sometimes sleep, his
cupidity may at some point be satiated; but those who torment
us for our own good will torment us without end for they do so
with the approval of their own conscience.
C.S. Lewis
Suppose you were an idiot. And suppose you were a member of
Congress. But I repeat myself.
Mark Twain
A nation trying to tax itself into prosperity is like a man standing in a bucket and trying to lift himself up by the handle.
Winston Churchill
A government big enough to give you everything you want, is
strong enough to take everything you have. My reading of history convinces me that most bad government results from too
much government.
Thomas Jefferson
The government solution to a problem is usually as bad as the
problem.
Milton Friedman
283
284
Afterword
We don't have a trillion-dollar debt because we haven't taxed
enough; we have a trillion-dollar debt because we spend too
much.
Ronald Reagan, 1989
No government ever voluntarily reduces itself in size. Government programs, once launched, never disappear. Actually, a
government bureau is the nearest thing to eternal life we'll ever
see on this Earth.
A young man, 21 years of age, working at an average salary
his Social Security contribution would, in the open market, buy
him an insurance policy that would guarantee $220 dollars a
month at age 65. The government promises $127. Now are we
so lacking in business sense that we can't put this program on a
sound basis? Barry Goldwater thinks we can.
Ronald Reagan, 1964
Only a crisis, real or perceived, produces real change. When
that crisis occurs, the actions that are taken depend on the
ideas that are lying around. That, I believe, is our basic function: to develop alternatives to existing policies, to keep them
alive and available until the politically impossible becomes
politically inevitable.
Milton Friedman
The problem with America is the high cost of energy, healthcare,

education, regulation, litigation, and taxation. All decrease the rate
of progress and push jobs offshore. All of these problems can be
fixed, but it is the US Congress who must fix it. They could pass legislation to fix all these problems in a small number of years if the citizens demanded it. Plans for all manner of reform have been sitting
around for decades, waiting for the political will. Ronaldus Magnus
talked about the need to reform the Social Security ponzi scheme in
the 1960s, and today we are still merely talking! America is a few
votes away from so much progress.
Before legislation can be voted on in the Senate, it needs a 60
vote requirement to end debate. This is a good check and balance
because it can allow a minority to keep out a bad government solution, which can be worse than the problem. However, for the longstanding problems that have existed in the United States, it is
entrenched interests in the Senate that have stalled progress. A lot
of people blame President Bush for the problems in the world
because his party controlled both houses of Congress for several
years, but this ignores the 60 vote requirement necessary to pass
any non-trivial legislation.
Afterword
285
Unfortunately, lots of what the Congress does involves so many

details that get lost on the public, which is why the media play such
an important role in educating or even shaping public opinion.
Education of a free society

Education is the transmission of civilization.
Will Durant, American historian
Give me four years to teach the children, and the seed I have
sown will never be uprooted.
Vladimir Lenin
The philosophy of the school room in one generation will be the
philosophy of government in the next.
Abraham Lincoln
I spent four years in the 1990s working at the centrist Brookings Institution and for the Clinton administration and felt right
at home ideologically. Yet during much of my two decades in
academia, Ive been on the far right as one who thinks that
welfare reform helped the poor, that the United States was
right to fight and win the Cold War, and that environmental regulations should be balanced against property rights.
All these views commonplace in American society and among
the political class are practically verboten in much of
academia. At many of the colleges Ive taught at or consulted
for, a perusal of the speakers list and the required readings in
the campus bookstore convinced me that a student could probably go through four years without ever encountering a right-ofcenter view portrayed in a positive light.
A sociologist I know recalls that his decision to become a registered Republican caused a sensation at his university. It was
as if I had become a child molester, he said. He eventually quit
academia to join a think tank because you dont want to be in
a department where everyone hates your guts.
Daniel Klein of George Mason University and Charlotta Stern of
Stockholm University looked at all the reliable published studies of professors political and ideological attachments. They
found that conservatives and libertarians are outnumbered by
liberals and Marxists by roughly two to one in economics, more
than five to one in political science, and by 20 to one or more in
anthropology and sociology.
I doubt that legions of leftist professors have set out to purge
academia of Republican dissenters. I believe that for the most
part the biases conservative academics face are subtle, even
unintentional. When making hiring decisions and confronted
with several good candidates, we college professors, like anyone else, tend to select people like ourselves.
286
Afterword
Unfortunately, subtle biases in how conservative students and
professors are treated in the classroom and in the job market
have very unsubtle effects on the ideological makeup of the professoriate. The resulting lack of intellectual diversity harms
academia by limiting the questions academics ask, the phenomena we study, and ultimately the conclusions we reach.
Robert Maranto, Associate Professor of Political Science, Villanova University
Sometime in the 1960s, Higher education abandoned their role
as advocates of American values critical advocates who tried
to advance freedom and equality further than Americans had
yet succeeded in doing and took on the role of adversaries of
society.
English departments have been packed by deconstructionists
who insist that Shakespeare is no better than rap music, and
history departments with multiculturalists who insist that all
societies are morally equal except our own, which is morally
inferior.
This regnant campus culture helps to explain why Columbia
University, which bars ROTC from campus on the ground that
the military bars open homosexuals from service, welcomed
Irans president Mahmoud Ahmadinejad, whose government
publicly executes homosexuals.
What it doesnt explain is why the rest of society is willing to
support such institutions by paying huge tuitions, providing tax
exemptions and making generous gifts.
Michael Barone, American political analyst
Thirty years from now the big university campuses will be
relics. Universities won't survive. It's as large a change as when
we first got the printed book. Do you realize that the cost of
higher education has risen as fast as the cost of health care?
And for the middle-class family, college education for their children is as much of a necessity as is medical carewithout it the
kids have no future. Such totally uncontrollable expenditures,
without any visible improvement in either the content or the
quality of education, means that the system is rapidly becoming
untenable. Higher education is in deep crisis.
Peter Drucker
I would rather be governed by the first 2,000 names in the Boston telephone directory than by the 2,000 members of the Harvard faculty.
William F. Buckley Jr., Rumbles Left and Right, 1963
A study demonstrated that 90% of the money donated by Harvard

professors in the 2006 election went to the Democrats. Blogger and
alumnus Steven M. Warshawsky tells me that there is one recognized conservative professor at Princeton, Robert P. George. I once
asked a political science student at a local Seattle university
Afterword
287
whether there were any Republican professors, and he merely

laughed. Higher-ed today focuses on diversity of skin color, but not
on diversity of thought. They are factories of liberalism which have
infected journalism and government. For the institutions that are
supposed to be preparing the next generation, this is a disaster,
especially because no one ever thinks they've been brainwashed.
Free Press
Promote then as an object of primary importance, institutions
for the general diffusion of knowledge. In proportion as the
structure of a government gives force to public opinion, it is
essential that public opinion should be enlightened.
George Washington Farewell Address, 1796
Journalism naturally draws liberals; we like to change the
world.
Washington Post Ombudswoman
Tonight we have put the best child care system in the world on
the American Agenda. That is to say, the system which is
acknowledged to be the best outside the home. Its in Sweden.
The Swedish system is run and paid for by the Swedish government, something many Americans [such as me] would like to
see the U.S. government do as well.
Peter Jennings, Anchorman of ABC News, 1989
I thought from the outset that Reagan's supply-side theory was
just a disaster. I knew of no one who felt it was going to work.
Tom Brokaw, Anchorman of NBC News, 1983
The collapse of Fannie and Freddie was completely preventable. The party that blocked any attempt to prevent it was:
the Democrat Party. The party that tried to prevent it was: the
Republican Party.
I have no doubt that if these facts had pointed to the Republican Party or to John McCain as the guilty parties, you would be
treating it as a vast scandal. Housing-gate, no doubt. Or Fannie-gate.
Instead, it was Senator Christopher Dodd and Congressman
Barney Frank, both Democrats, who denied that there were any
problems, who refused Bush administration requests to set up a
regulatory agency to watch over Fannie Mae and Freddie Mac,
and who were still pushing for these agencies to go even further in promoting sub-prime mortgage loans almost up to the
minute they failed.
Yet when Nancy Pelosi accused the Bush administration and
Republican deregulation of causing the crisis, you in the press
did not hold her to account for her lie. Instead, you criticized
Republicans who took offense at this lie and refused to vote for
the bailout!
288
Afterword
And after Franklin Raines, the CEO of Fannie Mae who made
$90 million while running it into the ground, was fired for his
incompetence, one presidential candidate's campaign actually
consulted him for advice on housing.
If that presidential candidate had been John McCain, you would
have called it a major scandal and we would be getting stories
in your paper every day about how incompetent and corrupt he
was.
But instead, that candidate was Barack Obama, and so you
have buried this story, and when the McCain campaign dared to
call Raines an adviser to the Obama campaign because
that campaign had sought his advice you actually let Obama's people get away with accusing McCain of lying, merely
because Raines wasn't listed as an official adviser to the Obama
campaign. You would never tolerate such weasely nit-picking
from a Republican.
Orson Scott Card, science fiction writer (Democrat)
The Republican debate provided red meat for conservatives:
anti-gay, pro-Jesus, anti-abortion and no gray matter in
between.
Brian Williams, Anchorman of NBC News, 2000
In a world without truth, freedom loses its foundation.
Pope John Paul II
In America the President reigns for four years, and Journalism
governs forever and ever.
Oscar Wilde
There never was an age of conformity quite like this one, or a
camaraderie quite like the Liberals'. Drop a little itching powder in Jimmy Wechsler's bath and before he has scratched himself for the third time, Arthur Schlesinger will have denounced
you in a dozen books and speeches, Archibald MacLeish will
have written ten heroic cantos about our age of terror, Harper's
will have published them, and everyone in sight will have been
nominated for a Freedom Award.
William F. Buckley Jr., National Review, 1955
The skillful propagandist has the power to mold minds in any
direction he chooses, and even the most intelligent and independent people cannot entirely escape their influence if they
are long isolated from other sources of information.
Frederick Hayek, The Road to Serfdom
Afterword
289
Poster of a smirking face found on many walls in Edinburgh in 2005.
A free society requires educated citizens for good governance.

The world we live in is very complex, which is why the media have
such an important role. Unfortunately, the mainstream media is
mostly filled with liberals who push a big-government agenda, and it
has been that way for many decades. The press in America are
mostly free of government funding and control, but if they advocate
for government-provided childcare, higher taxes, and other trappings of socialism, what is the difference? The reason we don't have
nuclear power plants in America is not simply because the Congress
has obstructed this progress, but because the media indelibly
etched the Three Mile Island nuclear accident as a catastrophe into
the consciousness of America, even though no one died.
While there are conservatives in the media such as at Fox News
and The Wall Street Journal, nearly all the rest, from the big three
TV networks of ABC, CBS, and NBC, to the cable networks of CNN
and MSNBC, to all of the major newspapers from The Washington
Post, to The New York Times, and The LA Times, lean Democrat.4
Jack Cashill wrote that The Kansas City Star, which serves a centerright community has created a product, both in its reporting and in
its editorial, much better suited to the residents of say, Boston or
Seattle, than of Kansas City.
4
The funniest (if you like sarcasm) and most convincing case for media bias is Ann
Coulter's best-selling Slander. She is undeservedly vilified, even by conservatives.
She might say controversial things, but doesn't Jon Stewart?
290
Afterword
Not only is the media biased in the US, it appears to be biased

around the world as well. There was a recent study commissioned
by the BBC which found itself to be guilty of a left-wing bias, and a
bias of omission. This study came out 1.5 years ago; a company
could reinvent itself in that time, but I wonder if the BBC has made
any changes? The bias in the news filters to the outside world: a
study found that late-night comics skewer Republicans by a ratio of
7 1.
Everyone has personal biases, but if The New York Times is an
intellectually honest newspaper, why hasn't it endorsed a Republican for President in the last 60 years? If Ronald Reagan was the
great president as he is now generally accepted to have been, why
did they miss two chances to endorse him? Surely there are countless other things, with hindsight of only a few years, that they have
gotten wrong.
The key to successful propaganda is two things: repetition, and lie
by omission. A good example is how the media have repeatedly
reported that President Bush is lazy, using the fact that he has taken
a record number of vacation days:
President Bush recently spent his 879th day at his ranch in
Crawford, Texas, breaking former President Reagan's record
for taking vacations from the White House.
Note how a vacation day is defined as one not at the White House.
However, it leaves out the fact that a President takes the resources
of the White House with him wherever he goes, and President Bush
has had meetings with foreign leaders, and with his various national
security and economic teams at his Texas ranch. The media leave all
these facts out, and just report again and again that they have the
numbers to prove that Bush is lazy even though he gets up at 5
am and exercises every day.
It is with repetition and the ability to lie by omission that the
media can mold minds in any direction they choose. I have Russian
friends from Microsoft who lived in the Soviet Union during the days
of Pravda, and yet they don't seem to imagine that those same propaganda techniques could and do exist here. I believe the bias of the
media is one of the greatest ongoing scandals of our age.
How can the media be leftward in this center-right country? Why
doesn't the free market fix this? The answer is that the barriers to
entry for a newspaper or TV station are very high. If you live in New
Afterword
291
York, you read The New York Times, and even those who don't like
its extreme political bent read it for other reasons such as its Arts
section.
Like creating a newspaper, creating a TV network is also very difficult; CBS, NBC and ABC have been around for almost 70 years.
Even though there is now cable news, the big 3 TV networks have
10 times as many viewers.
While Fox News has higher ratings than CNN and MSNBC combined, the rest of the media have colluded to discredit it as a mere
propaganda arm of the Republican party. I've talked to a fair number
of fellow software engineers who scoff at the idea of Fox as a legitimate news organization; the simplest way to squelch other viewpoints is to discredit them like this. Fox News was long-anchored by
Brit Hume, who worked for ABC News for 23 years without getting
fired for being a nutcase, but now it is presumed that he is.5 Pundit
Charles Krauthammer wrote: Rupert Murdoch and Roger Ailes are
geniuses: They found a niche market half of America.
A video demonstrating how screwed up the news media in America is can be found at: http://bit.ly/BrokenMedia. In another era, this
interview would have been a scandal, but this sort of malfeasance
takes place every day. The election of Barack Obama is a an excellent case study of media bias.
British journalist Nile Gardner wrote: Television news in America has for decades
been dominated by a left-of-centre oligopoly that has not reflected public opinion.
That smug arrangement was shattered when Fox opened for business in the mid1990s.
Fox News has succeeded spectacularly in racing ahead of its rivals in the cable
news market, notably CNN and MSNBC. Its evening shows such as the OReilly
Factor, Glenn Beck and Hannity pull in several million viewers compared to just
hundreds of thousands on Foxs competitors. Fox offers a highly opinionated, fastpaced and entertaining brand of political debate that includes all sides of the
political aisle. The top hosts may be largely conservative (though not necessarily
Republican), but the guests frequently are not, creating an adversarial and combative arena that until recently was a rarity in American news coverage. Fox is
unashamedly pro-American, a breath of fresh air in an age when US foreign policy
is increasingly weak, muddled and confused.
292
Afterword
Barack Obama
Halloween 2008, by Sam Ryskind

Media bias in 2008 was the most disgusting failure in our business since the Iraq war. It was extreme bias, extreme proObama coverage. I think it's incumbent upon people in our
business to make sure that we're being fair. The daily output
was the most disparate of any campaign I've ever covered, by
far.
Mark Halperin, Time Magazine (not a right-wing magazine)
The mainstream media were not to blame for John McCain's
loss.
Washington Post Ombudsman (not a right-wing newspaper)
But Obama deserved tougher scrutiny than he got, especially of
his undergraduate years, his start in Chicago and his relationship with Antoin "Tony" Rezko, who was convicted this year of
influence-peddling in Chicago.
Another gaping hole in coverage involved Joe Biden, Obama's
running mate. When Gov. Sarah Palin was nominated for vice
president, reporters were booking the next flight to Alaska.
Some readers thought The Post went over Palin with a finetooth comb and neglected Biden. They are right; it was a serious omission.
Washington Post Ombudsman, the previous week
Afterword
293
As for the Democrats who sneered and howled that Palin was
unprepared to be a vice-presidential nominee what navelgazing hypocrisy! What protests were raised in the party or
mainstream media when John Edwards, with vastly less political experience than Palin, got John Kerry's nod for veep four
years ago?
Camille Paglia, liberal feminist
During this election, the media in the United States of America
was worse than the media in communist Russia. The anchormen and anchorwomen were reading from the same script.
They might have had different haircuts and they might have
had different outfits, but they were reading from the same
script.
Orly Taitz, attorney
The swooning frenzy over the choice of Barack Obama as President of the United States must be one of the most absurd waves
of self-deception and swirling fantasy ever to sweep through an
advanced civilization. At least Mandela-worship its nearest
equivalent is focused on a man who actually did something.
You may buy Obama picture books and Obama calendars and if
there isnt yet a childrens picture version of his story, there
soon will be. Proper books, recording his sordid associates, his
cowardly voting record, his astonishingly militant commitment
to unrestricted abortion and his blundering trip to Africa, are
little-read and hard to find.
Peter Hitchens, Daily Mail (conservative)
Do not blame Caesar, blame the people of Rome who have so
enthusiastically acclaimed and adored him and rejoiced in their
loss of freedom and danced in his path and gave him triumphal
processions. Blame the people who hail him when he speaks in
the Forum of the more money, more ease, more security, and
more living fatly at the expense of the industrious. Julius was
always an ambitious villain, but he is only one man.
Cicero
Barack Obama is intelligent and has some good ideas, but he won
the presidency because journalists fomented anger towards Bush
over his 8 years, and advocated for Obama's victory. The media
ignored Obama's failed efforts at the Annenberg foundation, his liberal voting record, lack of bipartisan accomplishments, evidence
that Ayers ghost-wrote Obama's first memoir, and more. (I am no fan
of John McCain!) I'm not arguing that these things are all necessarily true, simply that they weren't even discussed.
294
Afterword
Here is a conversation between Tom Brokaw of NBC and Charlie

Rose of PBS that took place just a few days before the Presidential
election:
Charlie Rose: I don't know what Barack Obama's worldview is.
Tom Brokaw: No, I don't, either.
Charlie Rose: I don't know how he really sees where China is.
Tom Brokaw: We don't know a lot about Barack Obama and the
universe of his thinking about foreign policy.
Charlie Rose: I don't really know. And do we know anything
about the people who are advising him?
Tom Brokaw: Yeah, it's an interesting question.
Charlie Rose: He is principally known through his autobiography and through very inspirational speeches.
Tom Brokaw: Two of them! I don't know what books he's read.
Charlie Rose: What do we know about the heroes of Barack
Obama?
Tom Brokaw: There's a lot about him we don't know.
A study commissioned by the pollster Zogby of Obama voters

found that 86% knew about the scandal that Sarah Palin had spent
$150,000 on clothes, and 94% knew that Palin had a pregnant
teenage daughter, but only 40% knew that the Democrats controlled
Congress, only 17% knew that Obama won his first election by getting all of his opponents kicked off the ballot, and only 18% knew
that Biden quit a previous presidential campaign because of a plagiarism scandal. While many millions of people around the world
know many details about Sarah Palin, they nothing about how Joe
Biden was wrong during the Cold War, voted against the first Gulf
War, etc. Even though they've only learned half the truth, they don't
even notice! All of these examples show how a bias of omission can
shape public opinion.
In spite of the fact that I was not an Obama supporter, I believe
he could have a great presidency if he tackles some of the longstanding problems in our country. However, as it isn't clear he
appreciates the power of the free market, I don't know if it will happen.
Conclusion
I can easily envision a world where free software has completely
taken over, but where The New York Times, et al, are still advocating against the policies of a free society. Let's build both!
How to try Linux
HOW
TO TRY
295
LINUX
People wanted me to include a Linux CD with the book so that

readers could try it out. However, this would add significant costs,
makes the book annoying to hold, and free software is evolving so
quickly that the CD would become obsolete immediately.
Instead, I suggest you burn a Linux Live CD which will run
Linux right off the CD-ROM. This gives you a chance to play with
Linux and see how it works without touching your existing operating
system. I thought it would be hard to make the switch given my
many years deep experience with Windows, but within a few months
I was using Linux 100% of the time.
Another way to see if you aren't sure if you are ready for Linux,
try OpenOffice.org and Firefox, which also run on Windows and the
Mac. Once you are using free formats and free applications, you are
ready to switch. Demand your hardware and software suppliers support Linux so it will be there when you are ready. Demand free software and it will happen even faster.
If you'd like more detailed information, please visit my website:
http://keithcu.com/SoftwareWars.
296
Dedication
DEDICATION
Writing books is the closest men ever come to childbearing.
Norman Mailer, novelist
Art is never finished, only abandoned.
Leonardo Da Vinci
Edsger Dijkstra, who wrote about the crisis in software in 1972,

died in 2002. John McCarthy, who created garbage collection in
1959, is 80. Ronald Reagan, who campaigned on Social Security
reform in the 1960s, died in 2004. Milton Friedman, who laid out
the principles of a modern free market in the 1960s, died in 2006.
Arthur C. Clarke, godfather of the Space Elevator, died in 2008.
Gene Roddenberry, creator of Star Trek, died in 1991.
This book is dedicated to giants who long ago showed us what
needed to be done, but who won't be around to see it happen. This
book presents an optimistic perspective of man's future, once we get
off our butts.
Acknowledgments
I would like to acknowledge my family, friends, teachers, colleagues, reviewers, etc. The total list would be very long and I would
likely leave out or misspell names that would mean nothing to my
readers, so I won't even try. I am not great about staying in touch,
but that doesn't mean I don't cherish our time together!
The cover art was created by Nils Seifert and the cover was
designed by Alex Randall.
If you enjoyed a free version of this book, and you want to send
me a donation as thanks for my two years of labor, that would be
appreciated! You could purchase a paper copy, or go to keithcu.com
and click on the PayPal button. Any amount donated over $5 will be
given to worthy efforts in free software. If you'd like to contribute
money towards a particular area of free software, but don't know
how, I can help!
Keith Curtis
[email protected]
twitter: @keithccurtis

SoftwareWars 200911 PDF

Uploaded by

Copyright:

Available Formats

SoftwareWars 200911 PDF

Uploaded by

Document Information

Original Title

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

SoftwareWars 200911 PDF

Uploaded by

Copyright:

Available Formats

Draft: 1.

Copyright 2009 by Keith Curtis

Free Software Battle..........................................................................1

Free Software Battle

Free Software Battle

iven the technology that's already available, we should have

Free Software Battle

Free Software Army

Glenn Reynolds, in his book Army of Davids, talks about how

Free Software Battle

IBM's Blue Gene (pun intended) supercomputer runs a lightweight Linux on

Free Software Battle

Microsoft stock price, 1990 2000

A friend installed Linux in the mid-90s but he gave up because his

Free Software Battle

college summer intern, I had planned for a potential encounter and I

I joined Microsoft in 1993 when it was hitting its stride. It had

I asked him about the performance of Microsoft Exchange's database storage

Free Software Battle

sophisticated technologies, I didnt really believe it would take off in

Free Software Battle

cient at things, I could do it in one. I came to understand that

Graph of the number of entries in the Wikipedias of various languages. This

Encyclopedia Britannica ad from 1913

Imagine that you wanted to research a topic like nuclear power.

those who enjoy a few seconds of perverted pride in temporary

By making these policies an integral part of the culture,

In short, Wikipedia is an evolving relationship between people

The Linux mascot, Tux, created by Larry Ewing

he kernel of an operating system (OS) is the central nervous

While cloud computing, the movement of increasing number of applications and

50% of NT's code

Windows NT kernel architecture block diagram. Cutler had a Windows 95

Unfortunately for Microsoft, the original kernel lived on through

Microsoft's corporate culture was based on the theory that software

Linux has achieved enormous gains against Windows, even while,

Linus' primary job is to provide technical expertise. He once said

Linux Kernel Superiority

1. Refactored Code (Reliability)

backward compatibility constraints are one of the biggest reasons

In essence, the Linux kernel is a bundle of device drivers that

If you look at the code changes required to make a bugfix, in the

Here is a graph of all the function calls into the OS required to

System call graph in Microsoft's proprietary web server, IIS.

Diagrams provided by SanaSecurity.com

2. Uniform Codebase (Reliability, Maintainability, and Usability)

Linux engineers have found a way to run the same codebase on a

doors to the NSA's computers. I will discuss in a later chapter why

3. Frequent Ship Cycles (Maintainability

By contrast, in a free OS, software components depend only on

4. Lower Development Costs (Maintainability)

Some components contain multiple versions to allow for a transition period.

5. Security (Reliability and Maintainability)

Attempting to compare the security of operating systems is a

Another big difference between Linux and Windows is that Linux

tem files, which make it very difficult to remove. On Linux, I can

Microsoft has greatly increased its focus on security in the past

6. Linux has learned from Windows

The Feature Race