Computer industry pundits have long speculated as to which company might be the first to overthrow the Microsoft juggernaut. The company has grown at an unprecedented pace since 1975, when Bill Gates started writing BASIC interpreters with Harvard University's computers during reading period (Parsons). With over 90% market share for its flagship product, Windows, the desktop personal computer operating system market has essentially belonged to Microsoft for years ("Direct Testimony of Franklin M. Fisher" 25). Predictions for its successor have ranged from traditional rivals, such as Sun Microsystems, to quirky upstarts, such as Apple Computer. Thus far, even the guesses of those on the front lines have proven incorrect; though Sun and Apple are still forces to be reckoned with in the computer industry, neither has severely damaged Microsoft's grip on the operating system market. Only in the past few years has speculation begun to rise about a new possible threat. However, instead of emanating from a software or hardware monolith as expected, Microsoft's nemesis has emerged seemingly out of thin air.
Named for its Finnish inventor, Linus Torvalds, yet owned by no one in particular, the Linux operating system has made Microsoft very scared indeed (Ricciuti 1). Out of his own interest in gadgetry and software, Torvalds innocently devised Linux as something to work on for fun. Now, Linux is at the forefront of the open-source software movement, defined by economic and moral principles that seem to defy conventional economics. More perplexed than anyone are the seasoned software executives who have made their fortunes by vigorously guarding their firms' intellectual property.
The ultimate goal of the open-source movement is to make all software freely available for the world to use, in the sense that the intellectual property core of software, source code, should be universally accessible. In the words of Richard Stallman, the head of the Free Software Foundation (FSF), "Free software is a matter of freedom: people should be free to use software in all the ways that are socially useful" (Kuhn 1). This well-intentioned philosophy, shared by Stallman's colleagues at the FSF, as well as by open-source enthusiasts worldwide, is only flawed in the sense that it indirectly advocates the destruction of some of American's most powerful corporations. If anyone can view the source code for a given program by simply downloading it from the internet, there is no need to pay for it. Without payment, there is no profit, and without profit, companies cannot survive.
Though the FSF might prefer a world without greedy software companies, its own movement is anything but pure. The FSF conveniently chooses to ignore the fact that much open-source development piggybacks off of the profits and resources of commercial software companies. If commercial software developers were to disappear due to intense competition from Linux, it would only be a matter of time before open-source software disappeared, as well. One of the best-known open-source projects, Mozilla, began as the open-source version of Netscape Communicator, which was subsequently acquired by America Online. America Online has continued to support Mozilla in order to reinvigorate the Netscape Navigator web browser, which became bloated and slow after repeated efforts to compete with Microsoft Internet Explorer (Zawinski 1). Apple Computer released an open-source version of its latest operating system, Mac OS X, under the code-name "Darwin." In turn, Darwin was based on a combination of technologies from FreeBSD Unix, itself an open-source project, and OPENSTEP, a closed-source Unix variant developed by Steve Jobs's NeXT, which Apple eventually acquired when Jobs became Apple's Interim CEO ("Apple - Public Source - Darwin FAQ" 1). phpMyAdmin, a user-friendly, web-based front end to the open-source MySQL database, was once a free-standing open-source project, but it is now has ties to Maguma, an Italian software developer that specializes in closed-source programming tools for open-source developers ("phpwizard.net" 1). More and more programmers are realizing that developing quality software requires labor and capital, and in a capitalist society such as ours, both cost money.
Fortunately for Microsoft, Oracle, Adobe, and thousands of other traditional software companies, fears of a future in which all consumers are open-source zealots harboring expectations of free software are largely unfounded. Though open-source software is at the forefront of today's industry hype, if the movement remains dedicated to serving the needs of technically-minded users and programmers alone, open-source software will never achieve the success that Windows has known in the home and business markets. At best, it will dominate the market for high-end server software, such as web servers. For even if the technology behind Linux is a decade ahead of Windows in terms of speed, power and reliability—all crucial factors in back-office software—the technology behind the user interface is still a decade behind. Linux may fragment the server market enough to give a future commercial competitor an edge against Microsoft, but it will not directly crush the company, as Oracle CEO Larry Ellison recently insinuated ("Friend or Foe" 60).
Since Linux has been designed by and for programmers, by definition the user interface functions exactly how its users would like. The entire structure of Linux, from the file system, which allows an administrator to secure individual files at user, group and universal levels, to the software installation process, which usually involves compiling source code first, is designed to give the user ultimate control over the computing environment. The opportunity cost of creating such a system is that the user must know an entire dialect of Linux jargon beforehand. Most grandmothers, small business owners, and students are blissfully ignorant of terms such as "root," or crucial commands such as "ls," "df," "quota" and "gcc," some of which further require complex parameters to function properly. Few would guess that the most important Linux configuration files reside in the "etc" directory. While consumers of commercial software can usually find their favorite software application at the nearest computer store, open-source programs have neither the budget nor the marketing personnel necessary to find their way to store shelves. Instead, they must be downloaded from their creators' web sites, decompressed (usually twice, from compressed versions in the "gzip" and "tar" formats, respectively), compiled, and only then installed. After an application is finally set up on a Linux hard drive, users are strongly encouraged to spend time providing the authors with feedback about bugs or design issues; the best contributors can point out flaws in the source code directly. Faced with the prospect of memorizing yet another set of commands and procedures, after using Linux, the average Windows user might feel as though he had regressed back to the era of MS-DOS, or worse.
Even so, considering the incredible amount of hype in the media about open-source software, it seems surprising that more average users are not using open-source applications for their everyday needs. There are a number of possible explanations for why Linux has not yet replaced Windows. First, it is likely that consumers do not behave perfectly rationally, making decisions on price alone, as many economic theories assume. Consumers who thrive on rapid Schumpeterian "creative destruction," which characterizes the world of computer technology, are a mere subset of the whole; the rest are afraid of change. Given that using Linux often involves no vendor, no dedicated technical support phone number, no box, no manual, and no disk, there are a number of changes aside from the operating system itself that consumers must absorb. Second, it is also possible that Microsoft still uses questionable tactics behind closed doors to preserve its monopoly. Few hardware vendors, or Original Equipment Manufacturers (OEMs), have successfully managed to offer both Linux and Windows as options for consumers purchasing new desktop systems. Dell, which at the moment is generally regarded as the leading OEM, began offering Linux on its OptiPlex business systems in 1999, but withdrew the option only two years later (Kanellos 1, Spooner 1). According to one article, "Dell executives have suggested that the operating system has more potential for workstations and servers," due to the lack of common business software available for Linux (Spooner 1).
Despite the challenges associated with using Linux, open-source software is still attractive to many consumers because for the first time there appears to be a way out of the lock-in that Microsoft has so effectively enforced. The primary targets of the company's efforts to control the operating system market have been the distributors of hardware: OEMs, such as Dell and Gateway. In the United States of America v. Microsoft Corporation anti-trust case, Judge Thomas Penfield Jackson's decision that Microsoft had a monopoly in the operating system arena was heavily influenced by testimony from managers at Gateway, Compaq, Hewlett-Packard, and a number of other OEMs (2). In response to the question, "Is it your view as we sit here today that there is no choice but Microsoft for operating systems with regard to the Pavilion line?" John Romano, Hewlett-Packard's Operations Manager for the Home Products Division in Asia Pacific, stated, "Absolutely there's no choice." Bart Brown, Gateway's Vice President of Direct Marketing stated that even a ten-percent increase in the price of Windows would have no bearing on Gateway's decision to preinstall Windows. Several other OEMs cited Microsoft's exclusive dealing tactics as the reason why Windows was their only viable option, and in turn, their consumers' only choice at the point of sale ("Direct Testimony of Franklin M. Fisher" 23-24). Through Byzantine licensing agreements in which Microsoft would often charge different vendors different prices for Windows, at one point OEMs were forced to install Windows on each Intel-based computer they sold, with the alternative of answering to the company's massive legal department. With little choice but to continue selling the hardware and software solutions that consumers were demanding, the OEMs capitulated.
Microsoft leveraged consumer demand for Windows by designing the operating system specifically to exploit the network effects that its software generated through everyday use. Bill Gates was quick to realize that network effects were the main economic principle responsible for Microsoft's continued success, and he frequently alluded to "positive feedback loops" and "locking the customer in" during daily conversations at Microsoft's headquarters in Redmond (Bank 24-28). Before the advent of the internet, when consumers depended upon floppy disks to transfer files from one computer to another, the operating system determined whether or not a disk could be read. Though Apple made efforts to ensure that its Macintosh line would be compatible with disks formatted for MS-DOS systems through a software extension called "PC Exchange," Microsoft never offered a similar add-on for Windows. It instead chose to set up a barrier to entry by isolating its network of users from those on other platforms. Third parties eventually did write software to bridge the gap between Windows and Mac OS, but consumers were forced to pay a premium in order for Windows systems to read disks formatted for other platforms. Thus, the option of buying another Windows-based computer became far more attractive than that of purchasing a similarly priced non-Windows-based computer, plus conversion software.
Operating systems also exhibit indirect network effects through service technicians. Especially in large enterprises where a number of personal computers are simultaneously in use, technical support staff must be familiar with the minutiae of a given software platform in order to be successful at repairing problems that might arise in the course of everyday use. The more people who use the same platform, the greater the benefit derived by each individual user, since technicians do not need to be re-trained.
Not surprisingly, Microsoft used this same argument of standardization reducing costs in its anti-trust case to defend itself against allegations by the Department of Justice that its monopoly tactics were harming consumers (Schmalensee 13). While Judge Jackson's recommendation that the company be split into two parts, for Operating Systems and Applications, was never carried out, it did demonstrate why policy-makers are often in a poor position to manage high-technology innovation. If the government attempted to restrict the use of Linux to promote competition, or promote Linux over commercial software to decrease costs, the results could be disastrous. Innovation takes place so rapidly in the software industry that even when a company breaks the law, the government can end up doing more harm than good by intervening.
Therefore, when Linux provided a low-cost alternative to Windows, effectively providing consumers a way out of Gates's feedback loop for the first time, Microsoft had good reason to be afraid. The fact that its backers were concerned not with profit, but with some vague notion of social welfare, presented an extraordinary problem that would have confounded any company trying to compete in the same marketplace. Open-source software was actually taking the savings associated with information goods' marginal costs of zero, and passing it along to consumers. The contributors, adept programmers who typically hold other full or part-time jobs for commercial enterprises, had no monetary incentive to work. The monetary costs of production were literally zero, since no programmer received any sort of compensation from whichever non-profit group or organization coordinated the development efforts for any given project. The ultimate products were freely available for public use, including commercial use, and generally had a following of other technical zealots who are often programmers themselves. Businessmen were rarely involved, since programmers could communicate instantly and effectively with each other via the World Wide Web and e-mail, coordinating code and schedules, and eliminating the need for middle managers.
Microsoft's official response to Linux was close to panic. In a 2001 interview with the Chicago Sun Times, Microsoft CEO Steve Ballmer boldly asserted, "Linux is a cancer that attaches itself in an intellectual property sense to everything it touches," evoking the eternal ire of a number of critics (Greene 1). Gates took jabs at the General Public License (GPL), the license under which much open-source software is released, referring vaguely in an interview to its "Pac-Man-like nature" (Ricciuti 1). According to a working paper by Casadesus-Masanell and Ghemawat at Harvard Business School, "In economic terms, the GPL license attempted to establish the "public good" character of the software that it covered. This copyright form quickly became known as ‘copyleft,' as it was diametrically opposed to traditional copyright forms...The copyleft license was offered on take-it-or-leave-it terms and without it, software was not considered open-source" (4). Microsoft was afraid that it would be perceived as a group of mercenaries trying to compete with the U.S. Army.
Not all Microsoft employees were completely averse to the new operating system; a few wanted to put its technical superiority to work for Microsoft. Adam Bosworth, the lead developer on the Internet Explorer 4.0 team, was frustrated with his job after he lost an internal political battle over the inclusion of what he considered to be a mediocre technology in Windows 98, called Active Desktop. Microsoft ultimately decided to include Active Desktop in Windows 98 to bridge the gap between Windows and Internet Explorer—an action which the Department of Justice later used as ammunition against Microsoft in the anti-trust case ("Direct Testimony of Franklin M. Fisher" 45). Willing to give Microsoft one last chance, Bosworth asked Gates directly for special permission to use Linux in an experimental research project involving relational databases and XML (Bank 195). When Gates refused, Bosworth finally left the company (241).
Ironically, had Microsoft realized what would eventually happen to its existing competition in the server software space, it might have supported Linux with considerable ardor. The true victims turned out to be commercial vendors of Unix-based operating system software, since Linux was perceived to be a closer substitute for Unix than Windows ("Friend or Foe" 60). Vertically integrated manufacturers, who once built everything from microprocessors to high-end custom software around a Unix software core, suddenly found themselves out of favor with penny-pinching CIOs. Sun Microsystems, Silicon Graphics, IBM and Hewlett-Packard, which manufactured Solaris, IRIX, AIX and HP-UX respectively, were forced on the defensive. Each confronted Linux in a different way, with staggering effects across the board.
Sun, a darling of technology analysts since its amazing debut in the early 1980s, saw its stock (NASDAQ: SUNW) decline from a peak of approximately $80 per share in late 2000 to slightly above $3 per share, where it remains today ("Yahoo! Finance - SUNW" 1). In September, 2000 it acquired Cobalt Networks, a manufacturer of small, Linux-based web hosting servers "Sun Microsystems Acquires Cobalt Networks" 1). Sun is just beginning to offer Linux pre-installed on some of its high-end workstations ("Friend or Foe" 60). Another company which caters to the high-end computing market, Silicon Graphics, once offered consumers a choice between IRIX and Windows NT. It now offers a choice between IRIX and Linux (Shankland 1).
After an initial period of uncertainty, IBM embraced Linux with such zeal that it found itself in violation of city ordinances. In April of 2001, in an advertising campaign that ran simultaneously in several major cities, IBM was fined several thousand dollars for spray painting the sidewalks of San Francisco and Chicago with symbols representing peace, love, and Linux (Hechinger B7). One of the original pioneers of Unix, the Santa Cruz Operation (SCO), has recently entangled itself in a battle with IBM, in what skeptics are calling a last stand before the company self-destructs. SCO alleges that IBM integrated copyrighted material from the SCO UnixWare kernel into IBM's own open-source Linux kernel. Though SCO will not disclose the source code in question, it claims that the infringement is substantial enough to merit damages (Cooper 1).
Now that Hewlett-Packard has acquired Compaq, previously one of Microsoft's largest Windows customers due to its leading market share, HP is encouraging customers of Compaq Tru64 UNIX to transition to its own HP-UX, which itself is slowly becoming obsolete. Enterprises running on Tru64 UNIX will eventually have no choice but to transition to another system, since Digital Equipment Corporation's Alpha microprocessor, which was also acquired by HP in the deal with Compaq, has been discontinued. Given other firms' desires to purchase enterprise software from other enterprises, the battle between the various Unix flavors will ultimately favor the commercial distributors of Linux, and the one company that manufactures a non-Unix operating system designed for the business market: Microsoft.
At the same time, Linux will surely cut into some of the Windows server market, especially for companies purchasing servers for the first time that do not face any switching costs or backward compatibility issues. Total Cost of Ownership (TCO) has long been a hot topic among Chief Information Officers trying to decide which platform to settle on. According to statistics compiled by the Boston Consulting Group, Linux is more than ten times more cost effective than Windows for an organization supporting 1,000 users on a network. At a total cost of $14,349, only one Linux server is required, as opposed to four Windows servers at a total cost of $177,000. The only area in which Linux commands a higher price is in hourly labor for technical support, where Linux technicians demand $250 per hour, as opposed to Microsoft technicians, at only $125 (Casadesus-Masanell and Ghemawat 30). Given market pricing for information technology consultants in the Boston area, however, these figures are debatable.
Therefore, the belligerent attitude promulgated by Microsoft executives is not entirely unfounded, at least from a financial perspective. In its most recent Securities and Exchange Commission 10-K filing, Microsoft stated, "To the extent the open source model gains increasing market acceptance, sales of the company's products may decline, the company may have to reduce the prices it charges for its products, and revenues and operating margins may consequently decline" (Galli 1). Large consulting firms have issued reports echoing Microsoft's sentiments. In "Fear the Penguin," a recent Goldman Sachs report (whose title refers to the mascot of the Linux movement), analysts suggest that in the next three years, the Windows market share in corporations will decline due to the new open-source alternative. Forrester Research also predicts that in the future, corporations will increase their dependence on Linux (Markoff C3).
Microsoft's 2002 Annual Report tells a different story about the company's health: revenues almost doubled from $15.2 billion in 1998 to $28.4 billion in 2002. Net profits also increased, from $4.5 billion to $7.8 billion in the same timeframe ("MSFTAR Form 10-K" 1). In recent years, Microsoft has been able to extend its reach into numerous other areas aside from software, such as television (MSNBC), video games (Xbox), and travel (Expedia.com). Given that the momentum behind Linux was growing simultaneously with Microsoft's diversification into so many distinct markets, one must question whether open-source software has really had the stifling effect that the company has so frequently invoked for pity.
Even if Linux did have a damaging effect on sales of Windows server software, Microsoft could easily adjust its pricing strategy (not to mention its general strategies in other markets) to compensate. Microsoft has already begun this process; in an interview, Gates admitted that the company has been forced to lower prices for some versions of Windows (Ricciuti 1). Assuming this trend continues, and that Linux actually does pose a threat to Microsoft, the price of Windows will eventually fall below the perceived switching cost of transitioning to Linux. At that point, consumers will be compelled to remain with the operating system that they have already been using for years.
Clearly, the switching costs involved with transitioning to Linux are not trivial. For the typical user, they are so large, in fact, that they more than make up for the $0 price tag that Linux and its associated applications carry. Most users accustomed to the drag-and-drop simplicity of operating systems including Windows XP and Mac OS X probably question the value of learning a system derived from the insights of Bell Labs engineers in the 1960s. Though graphical user interfaces such as GNOME and X11 are available for download free of charge, knowing which to use, let alone how to install the software once it has been downloaded, presents a further series of challenges. To a large extent, the market has already reached an equilibrium. The main concern for Microsoft and other closed-source companies should not be whether or not Linux-based computers will dominate the personal computing world, but whether or not the niche market for server software is crucially important.
Based on the company's past actions, Microsoft believes the server market is important, and it seems unlikely that Microsoft would simply let Linux take a significant amount of market share from Windows without first exercising its legal rights ("Direct Testimony of Franklin M. Fisher" 42). The open-source community, which might have evaded legal action by virtue of its dispersed nature, has nevertheless managed to provide excellent targets for Microsoft's attorneys. Upon realizing that firms would be more willing to adopt Linux if they could turn to a commercial vendor capable of offering business-class technical support, a number of for-profit Linux vendors came into existence. Today, in addition to the efforts of individual programmers, much of the innovation that takes place on the Linux platform is due to the efforts of Red Hat, SuSE, VA Linux and Debian. These companies, generally lacking high profits in a market where identical product offerings are available at no charge, are perfect targets for Microsoft's legal team.
With commercial Linux distributors, the FSF doctrine, which grew out of dissatisfaction with commercial software enterprises, comes full-circle. Compared to "pure" open-source communities, even the most successful of these companies appears incredibly wasteful. Multiple pay levels for programmers, managers and salespeople, not to mention employment benefits, computing equipment, office space and other capital expenditures all contribute to relatively elaborate cost structures. Open-source purists point out that these costs are largely unnecessary, avoidable by simply using a few on-line tools to synchronize the efforts of volunteer programmers around the globe.
One must wonder, however, if commercial-quality software can really be achieved using a loosely coupled network of programmers, given the radical differences between the two approaches to development. The most notable omission from the FSF open-source approach is the lack of face-to-face human interaction. The typical exchanges that take place within the offices or corridors of office buildings simply do not exist in the virtual realm of the typical open-source project. Closed-source developers have two advantages as a result: interactions between programmers, and interactions between programmers and customers. Debugging code can be a tedious and exhausting process for a single programmer who has spent hours thinking about a single feature or function. It is often beneficial to have another person nearby to proofread for errors, which often turn out to be misplaced punctuation marks. While programmers can still look over code on-line, the process is made more difficult by coordinating across time zones and languages, which span all the way across the globe. In the case of interaction between programmers and average customers, open-source developers barely give the subject thought. Most "customers" are also developers or advanced technical users who enjoy the challenge of decoding complicated user interfaces. In contrast, closed-source developers might allocate entire budgets to implementing detailed usability studies. Without the budgets to back them, these studies are rarely conducted in the open-source world, contributing to the difference in overall usability.
The exception to this rule is Michael Robertson, a Linux advocate and entrepreneur who founded MP3.com during the dot-com boom. After giving the problem a considerable amount of thought, Robertson started a new company called Lindows, yielding what is so far the best compromise between the worlds of Windows and Linux. Lindows has introduced a low-cost computer running a customized version of Linux, with the added twist that the GNOME graphical user interface has been pre-configured to resemble the look-and-feel of Microsoft Windows. Lindows also includes software to read popular Windows file formats, such as Microsoft Word. Consequently, a novice user familiar with the Windows desktop and icon descriptions (such as "My Computer," "My Documents," etc.) should feel right at home using a Lindows PC. To facilitate distribution, Robertson secured a deal with Wal-Mart in order to sell the machines through the massive retail chain. Microsoft, not surprisingly, has expressed concern that the Lindows name infringes on its Windows trademark (Becker 1).
From a consumer perspective, Lindows represents an option that previously was unavailable the last time two products that both exhibited strong network effects dueled. In their 1986 paper, Katz and Shapiro examined the VHS versus Betamax case, when consumers were forced to choose between two incompatible technologies that exhibited strong network effects. They reached the conclusion that under certain circumstances, usually involving a social planner, it was possible for both technology A and B to peacefully co-exist. If Katz and Shapiro were writing the same paper today, they would have to consider what might happen if a third product C were introduced capable of using the proprietary formats of technologies A and B. In other words, the Lindows of the 1980s could have been a VCR that could play both kinds of tapes by using a mechanical adapter.
The open-source debate tends to lean excessively to one side or the other, thanks to the zealousness with which Microsoft protects its intellectual property and the eccentricities of open-source developers. Neither side is willing or able to accept the other's arguments because both stand to lose incredible amounts. If Microsoft were to release the source code to its products, it would risk going out of business, as anyone would be able to copy, modify, and improve on its products without ever turning over another cent. Likewise, if open-source developers were suddenly responsible for compensating the authors of the software that they used on a regular basis, the financial burdens would be enough to force many of them into personal bankruptcy. Meanwhile, the behemoth remains in an essentially insurmountable position in the software industry, with hundreds of products, domination of internet programming standards via more than 90% market share for its web browser, Internet Explorer, and ubiquitous use of its cornerstone product, the Windows operating system. Whether this will be the case in ten years remains to be seen. If Microsoft falls, however, it will not be because of Linux.
"Apple - Public Source - Darwin FAQ." Updated Unknown. <http://developer.apple.com/darwin/projects/darwin/faq.html>. Cited 6 May 2003.
Bank, David. Breaking Windows: How Bill Gates Fumbled the Future of Microsoft. New York:
The Free Press, 2001.
Becker, David. "Lindows heads for store shelves." Updated 17 December 2002. <http://news.com.com/2100-1040-978065.html>. Cited 6 May 2003.
Campbell-Kelly, Martin. From Airline Reservations to Sonic the Hedgehog: A History of the
Software Industy. Cambridge, Massachusetts: The MIT Press, 2003.
Casadesus-Masanell, Ramon and Ghemawat, Pankaj. "Linux vs. Windows: Modeling Competition between Open-Source and Closed Software." Working Paper. Harvard Business School, 2003.
Cooper, Charles. "Unix's war on Linux." Updated 28 March 2003. <http://news.com.com/2010-1071-994463.html>. Cited 6 May 2003.
"Direct Testimony of Franklin M. Fisher." Updated 5 January 1999. <http://www.usdoj.gov/atr/cases/f2000/2057.pdf>. Cited 6 May 2003.
Freiberger, Paul and Swaine, Michael. Fire in the Valley: The Making of the Personal Computer. 2nd ed. New York: McGraw-Hill, 2000.
"Friend or Foe." The Economist 12 April 2003: 60.
Galli, Peter. "Microsoft Warns SEC of Open-Source Threat." Updated 3 February 2003. <http://www.eweek.com/article2/0,3959,857638,00.asp>. Cited 6 May 2003.
Greene, Thomas C. "Ballmer: "Linux is a cancer." Updated 2 June 2001. <http://www.theregister.co.uk/content/4/19396.html>. Cited 6 May 2003.
Hechinger, John. "IBM's Sidewalk Art Draws Unappreciative Reaction in Chicago." The Wall Street Journal 26 April 2001: B7.
Kanellos, Michael. "Dell to sell PCs with Linux." Updated 19 March 1999. <http://news.com.com/2100-1001-223292.html>. Cited 6 May 2003.
Katz, Michael L. and Shapiro, Carl. "Technology Adoption in the Presence of Network Externalities." Journal of Political Economy 94, no. 4 (1986): 822-41.
Kuhn, Bradley. "Philosophy of the GNU Project." Updated 25 April 2003. <http://www.fsf.org/philosophy/philosophy.html#AboutFreeSoftware>. Cited 5 May 2003.
Markoff, John. "Microsoft Tries to Conquer the Corporate Data Center." The New York Times 24 April 2003: C3.
"MSFTAR Form 10-K." Updated 2002. <http://www.microsoft.com/msft/ar02/financials/highlights.htm>. Cited 6 May 2003.
Parsons, Philip. "Access to Harvard: A Portrait of the Faculty of Arts & Sciences." 1995. [CD-ROM]
Ricciuti, Mike. "Gates wades into open-source debate." Updated 19 July 2001. <http://news.com.com/2100-1001-268667.html>. Cited 6 May 2003.
Shankland, Stephen. "SGI begins high-end Linux push." Updated 6 January 2003. <http://news.com.com/2100-1001-979372.html>. Cited 6 May 2003.
"phpwizard.net." Updated 2003. <http://www.phpwizard.net>. Cited 6 May 2003.
Schmalensee, Richard L. "Executive Summary." Updated Unknown. <http://www.microsoft.com/presspass/trial/schmal/ExecutiveSummaryWeb.asp>. Cited 6 May 2003.
Spooner, John G. "Linux slips off Dell computers." Updated 2 August 2001. <http://news.com.com/2100-1001-271006.html>. Cited 6 May 2003.
"Sun Microsystems Acquires Cobalt Networks." Updated 19 September 2000. <http://www.sun.com/smi/Press/sunflash/2000-09/sunflash.20000919.2.html>. Cited 6 May 2003.
Torvalds, Linus and Diamond, David. Just for Fun: The Story of an Accidental Revolutionary. New York: HarperBusiness, 2001.
United States of America v. Microsoft Corporation. Civil Action No. 98-1232 (TPJ).
"Yahoo! Finance - SUNW." Updated 6 May 2003. <http://finance.yahoo.com/q?s=SUNW&d=c&k=c1&a=v&p=s&t=my&l=on&z=m
&q=l>. Cited 6 May 2003.
Zawinski, Jamie. "Fear and Loathing on the Merger Trail." Updated 23 November 1998.
<http://www.mozilla.org/fear.html>. Cited 6 May 2003.