It's generally agreed that college and university students should learn the fundamentals of information technology, including the use of operating systems, office application software and the Internet. It's quite another matter, though, to pay for the necessary infrastructure--wired dormitories, industrial-strength servers, lots of PCs around campus, and pricey commercial software for student use. Now that Linux and open-source office applications such as AbiWord and Gnumeric are available for free, institutions of higher education can save big money in software costs, and more than a few campuses and university consortia are starting to take Linux seriously (see, for example, Robiette 1999). They're discovering what Linux users already know--namely, that Linux, compared to Microsoft Windows, offers an unbeatable combination of advantages, including a zero price tag, do-it-yourself flexibility, freedom from licensing headaches, stability, performance, compliance with public standards, interoperability with existing systems, and a design that reduces the threat of computer viruses (see Prasad 1999).

As I'll argue in this essay, there's much more at stake here than money. In what follows, I'll argue that open source software in general--and Linux in particiular--holds the key to the ability of colleges and universities to retain their traditions of scientific and scholarly excellence as they adapt to an increasingly computerized world. By establishing Linux as the international standard for academic computing, institutions of higher education can directly address challenges to the integrity of scientific research, do a better job of preparing students for a world of rapidly changing technology, and combat the growing and disturbing disparities in access to information technology. The following sections detail the case for Linux in higher education--a case that, in my view, amounts to a moral imperative.

Since science's earliest days, the enterprise has been based on a gift-economy notion very much like that underlying open- source software: scientists receive credit and prestige for their discoveries, but they do not receive ownership of them. On the contrary, scientists are expected to publish their findings in open, public journals, which are accessible to all. These journals print scientific articles only after a submission passes peer review, in which a scientist's peers scrutinize all of the assumptions and calculations that produced the conclusions. The journal's editor will publish a scientific article only when the peer reviewers conclude that the underlying methods are sound. To be sure, the system doesn't always work perfectly, but--like democracy--it is clearly superior to its alternatives.

Increasingly, scientists are beginning to see that their use of closed-source software poses a profound threat to the integrity of science (Kiernan 1999). Computer software is increasingly used to analyze research results or simulate real-world systems. However, scientists rarely make their software available to other scientists for scrutiny--and even if they did, they often used closed-source programs in which the underlying source code is protected by copyright and trade secrecy claims. But this practice strikes at the heart of science, namely, the notion of verifiability. To be accepted as valid, all calculations and assumptions that go into a given scientific assumption must be open to public scrutiny. Yet closed-source software makes such scrutiny impossible.

These are the simple facts, from which Dan Gazelter, a professor of biochemistry at Notre Dame University, draws the following, compelling conclusion: scientists are positively obligated to use open-source software, and what is more, the future of an increasingly computerized scientific enterprise may well depend on their decision to do so (Gezelter 1999; cf. Wilson 1999). Increasingly, scientists and university librarians are developing clearinghouses and large-scale development projects to create more open-source alternatives for use in higher education (see the Open Science Project and oss4lib).

But the use of open-source software is insufficient. If the future of science depends on scientists' use of open-source software, one can very well argue that colleges and universities are under a positive obligation to move away from closed-source computing infrastructures as well as closed- source software. Consider this: many of the instructions in computer programs do little more than issue directives to the operating system; this is done by means of the operating system's application programming interface (API). To verify scientific software fully, the scientific community may need to examine the program's interaction with the operating system. Yet Microsoft refuses to document the Windows API fully and regards the Windows source code as an immensely valuable trade secret. What is more, Microsoft has taken the lead in lobbying for proposed changes to the U.S. commercial code that would effectively criminalize reverse engineering.

It's not enough for scientists to use open-source software; they must also use an open-source operating system. Colleges and universities can help to assure the ubiquity of open-source software and operating system usage in science by moving to Linux as an international standard for academic computing.

Let's turn from professors to students and examine another area in which the use of commercial, closed-source software is rapidly growing: the computer literacy curriculum. Even if it is admitted that open-source software is needed at the research level, skeptics will counter that colleges and universities do not exist in a vacuum. Whether we like Microsoft or not, they will argue, we still need to teach students how to use the software they will encounter after they graduate--and that means Microsoft Windows and Microsoft Office. Among those pushing for this type of computer literacy instruction are local businesses, which hope to avoid paying the enormous costs needed to train their employees to use such software effectively. Increasingly, computer literacy instruction looks like a Windows and Office training seminar. Not surprisingly, Microsoft is taking advantage of this situation by cutting deals with colleges and universities that provide every enrolled student with licensed copies of Microsoft software. In response to this assertion, I argue that a focus on Windows and Office skills is the wrong type of computer literacy instruction at the college and university level. In addition, I argue, it fails to serve the needs of business.

Is a focus on Microsoft Windows and Office skills the right kind of computer literacy instruction at the college and university level? If you accept the conclusions of a major national report that reflects a growing consensus among computer literacy educators, the answer is "No." The report, titled Being Fluent with Information Technology(National Research Council 1999), is the report of the U.S. National Research Council's Committee on Information Technology Literacy and represents the best thinking of the leading experts in the field. In brief, the report rejects a narrow focus on skills-based training for the following reasons:

Recognizing these facts, the authors of Being Fluent with Information Technology conclude that a computer literacy curriculum focusing on skills alone is insufficient. The ideal curriculum, they argue, would equip students with computer fluency, a "robust understanding of what is needed to use information technology effectively across a range of applications" (14). In addition to possessing the essential skills of software usage, computer-fluent individuals can apply information technology in novel situations--and what is more, they can understand the consequences of doing so. As the authors observe, "these capabilities transcend particular software and hardware applications" (17). Equally essential to computer fluency is the mastering of fundamental computer concepts, such as the difference between absolute and relative cell references in an electronic spreadsheet program.

If computer fluency is indeed a desirable goal, then it follows that colleges and universities can and should base their curricula on products other than Microsoft's. A student who learns the fundamental concepts of spreadsheet usage from Gnumeric, admittedly, may not know which key to press when confronted with Microsoft Office. However, computer literacy instruction should not focus on which key to press, but rather on the concepts that underlie the use of computer software. A student who fully understands the concepts of absolute and relative cell references will experience little difficulty learning Microsoft Excel; she will quickly learn which key to press. Indeed, asking students to move to a different vendors' spreadsheet application may well be the best way to test whether students have acquired the desired computer fluency. In contrast, a student whose computer literacy instruction emphasized Excel skills rather than the transcendent concepts of software usage may require a round of costly retraining when the next version of Excel is released.

From the foregoing argument, one can conclude that colleges and universities can well serve the goals of computer literacy education by moving to a Linux standard. We should teach the concepts of operating system and office software usage, and there is no reason to use expensive, commercial products for this purpose. At higher curricular levels, colleges and universities are arguably under a positive obligation to move away from closed source software and proprietary computing infrastructures (Vermeer 1998). Increasingly, it is not only scientists who must understand the details of operating systems and computing networks; advanced research in virtually every field of scholarship inevitably requires the type of intermediate to advanced understanding of information technology that was formerly possessed only by computer science graduates. In this context, open-source operating systems and networking infrastructures offer a significant advantage: they are open to dissection, analysis, and scrutiny in ways not possible with closed-source architectures.

Let's leave the campus entirely and consider the broader society. And what's going on is troubling. There's an increasing gulf separating the IT-literate "haves" and "have-nots." Over the past 35 years, the income ratio separating the world's richest from the world's poorest nations has nearly tripled (Watkins 1999). In general, familiarity with IT and access to IT are associated with the adaptability needed to cope with rapid social and economic change. With this adaptability comes greater earning power. Conversely, lack of familiarity and access are associated with flat or declining incomes. What is more, this disparity in IT literacy--the so-called "Digital Divide--tends to line up not only with ethnic divisions in the U.S., but also with the widening global gulf between the economies of the North and those of the South (Vee 1999).

If IT has helped to generate the problem, it may also offer a solution, in the form of Internet-mediated distance education. Using distance education, colleges and universities may be able to extend educational opportunities to precisely the areas that are poorly served by colleges and universities today, such as the inner cities and impoverished rural areas. But academia may not be able to succeed if distance education is hijacked by commercial vendors, who are bent on extracting lucrative profits from what they see as a huge and growing market. Already, they are pushing universities to adopt policies that rob professors of their right to the intellectual property produced in the classroom, so that this property can be packaged and sold to commercial distance education vendors (Noble 1998).

Distance education cannot succeed in impoverished areas of the U.S., let alone the Third World, if students and schools are forced to pay commercial software licensing fees and copyright fees in addition to the cost of computer hardware and network connections. The use of commercial operating systems and commercial applications for distance education is impossible to justify when stable, high-quality products are available from the open-source community. This is precisely the reasoning that led Mexican government officials to choose Linux and the GNOME desktop as the foundation for a new push to place computers in Mexican schools (Kahney 1998). The economies of Linux are no less relevant to underfunded schools in rural and inner-city settings in the U.S. (Dean 1999). Adopting commercial, closed-source software as the infrastructure for distance education amounts to a slap in the face to the poor.

Were colleges and universities to move to a Linux standard for academic computing, they would soon become full participants in a growing movement to use open-source software as a means of achieving social justice worldwide. They would become engines of open-source software development -- and the results could make a genuine difference in helping to remedy international inequities in access to information technology.

If you're skeptical of this claim, consider the Littlefish project (Frazer and Brown 1999). In brief, Littlefish is an open- source project that seeks to develop license-free patient information software for use in rural and Third World settings. In such settings, the use of commercial software is impossible, and not just because of the cost; the vendors of commercial patient information software have little interest in supporting users in remote or impoverished locales. Accordingly, part of the Littlefish project's goal is not only to create high-quality, open-source patient information software, but what is more, to create a worldwide community of practitioners who possess and are eager to share the expertise needed to implement effective patient-tracking systems, even under conditions of extreme poverty and geographic isolation. Does this effort matter? The answer is found in one simple statistic: 97 percent of childhood deaths occur in developing countries. Effective patient information systems could help to reduce this mortality significantly.

Arguably, open-source software holds the key to addressing significant issues of social justice and economic development worldwide (Vee 1999). Accordingly, colleges and universities should do all they can to foster open-source software development, and this purpose would be admirably served by adopting Linux as the international standard for computing in higher education.

I'm fully aware that moving to a Linux standard would pose new and difficult challenges for colleges and universities; for one thing, Linux isn't the easiest operating system to learn, and maintenance costs could soar as students meddle with system configurations and wind up with unbootable systems. Still, Linux distributors are working hard to make Linux easier to install and use. As the GNOME and KDE desktops reach maturity, they will open the use of Linux to a much broader audience. But most importantly, the very nature of Linux as an open-source operating system will enable colleges and universities to create and distribute customized Linux distributions (for an example, take a look at CAEN Linux, a version of Red Hat 6.0 that's customized for University of Michigan e-school students). These "educational versions" of Linux will include pre-configured system and network settings that are designed to work seamlessly and transparently with the campus computing network, eliminating the need for students to acquire system and network administration skills.

Perhaps the best argument for moving to the Linux standard, however, comes from a consideration of what may happen if current trends continue. Microsoft is making increasing inroads into the academic computing market, largely on the strength of multi-million dollar deals that make Microsoft software available to all registered students. Microsoft has all but taken over academic computing, save on the server end. Colleges and universities have become yet another stepping-stone to Microsoft's stranglehold on the world market for Intel-based operating systems and application software.

I am well aware that many academics who are sympathetic to Linux are put off by what they see as "gratuitious Microsoft-bashing" by Linux advocates, but the bashing is far from gratuitous; indeed, there is ample evidence--supplied by none other than Microsoft itself--that the firm is considering measures that would drive Linux out of the marketplace, just as Microsoft has similarly destroyed earlier competitors. The so-called Halloween documents (Harmon and Markoff 1998), released to the Internet by an unknown source within the company, disclose a plan to counter Linux by "de-commoditizing" the public protocols that currently form the basis of campus computing networks and the Internet and will in the future provide the infrastructure for distance education.

What is meant by "de-commoditizing" public networking protocols? Currently, such protocols, such as the Domain Name System (DNS), are "commodities", in Microsoft's terminology, in the sense that they are standardized and publicly available. By adding proprietary extensions to these protocols, Microsoft hopes to make the use of non-Microsoft software more costly to users, even as the use of Microsoft software becomes more convenient. Microsoft's internal documents make it clear that the firm intends to introduce such extensions, not because doing so is in their customer's best interest or would improve their products, but because such extensions could prove effective in pushing Linux out of the marketplace.

To be sure, Microsoft disavows these documents and rejects the assertion that they accurately characterize the firm's intentions. However, Microsoft's critics argue that the company has played this game for years, and only the naïve would believe that it will no longer continue to do so. The company stands accused in a U.S. Federal Court of using similar tactics against its competitors in the past; if it wins, which seems increasingly likely, the firm's attorneys will no doubt advise the company's executives that it is free to pursue such tactics against new competitors as well--and currently, Microsoft's most vigorous competition stems from the open-source community.

In my opinion, the very fact that someone inside Microsoft considered such tactics is reason alone to argue that Microsoft software has no place in academic computing. I realize the software marketplace is a rough-and-tumble world, and hardball tactics are commonplace. Still, Microsoft doesn't seem to know where the line lies between aggressive competition and reprehensible, potentially illegal actions that result in the annihilation of anyone with the temerity to compete with the company--and at incalculable cost to consumers and the general public. For example, Microsoft is taking the lead to lobby for new legislation that will rewrite copyright laws--and the results threaten the very existence of well-established conceptions of fair use and reverse engineering, practices on which university research and education are deeply dependent. Given that Microsoft is involved in a variety of activities that threaten the very existence of open intellectual exchange in colleges and universities, they have every reason to do nothing to advance the interests of Microsoft, but in contrast, to do everything they can to advance the development of Linux.

In sum, the survival of open-source software in general, and Linux in particular, may prove essential to the preservation of the integrity of science, the effectiveness of computer literacy instruction, and the reduction of the digital divide. Colleges and universities are under a positive obligation, which I believe amounts to a moral imperative, to reject the growing role of commercial software in academic computing, and to transform themselves by the thousands, worldwide, into vigorous centers of open-source software development.

Bryan Pfaffenberger is a professor in the University of Virginia's new Media Studies program, where his responsibilities include developing and teaching UVa's new University-wide computer literacy course (Media Studies 110). His works on Linux include Linux Clearly Explained (Morgan Kaufmann) and Mastering GNOME (Sybex).

Dean, Katie. 2000. "Open Source Opens Education," Wired News (March 13, 2000). Available on-line at http://www.wired.com/news/culture/0,1284,34807,00.html.

Frazer, C., and S.M. Brown. 1999. "The Littlefish Project: Open Source, Open Health." Available on-line at http://www.paninfo.com.au/papers/hics%2099%20presentation.htm.

Kahney, Leander. 1998. "Mexican Schools Embrace Linux," Wired News (November 6, 1998). Available on-line at http://www.wired.com/news/technology/0,1282,16107,00.html.

Gazelter, J. Daniel. 1999. "Catalyzing Open Source Development in Science," paper presented at a conference entitled "Open Source/Open Science," Brookhaven National Laboratory, October 2, 1999 (slides available on the Web at http://www.openscience.org/talks/bnl).

Harmon, Amy and John Markoff. 1998. "Internal Memo Shows Microsoft Executives' Concern over Free Software," New York Times (November 3, 1998). Available on-line at http://www10.nytimes.com/library/tech/98/11/biztech/articles/03memo.html (requires site registration).

Kiernan, Vincent. 1999. "The 'Open Source Movement' Turns Its Eye to Science," Chronicle of Higher Education (November 5, 1999). Available on-line at http://www.chronicle.com/free/v46/i11/11a05101.htm.

National Research Council, 1999. Being Fluent with Information Technology. Report of the Committee on Information technology Literacy, Computer Science and Telecommunications Board, Commission on Physical Sciences, Mathematics, and Applications. Washington, D.C.: National Academy Press. Available on-line at http://books.nap.edu/books/030906399X/html/R1.html.

Noble, David. 1998. "Digital Diploma Mills: the Automation of Higher Education," First Monday , available on-line at http://www.firstmonday.dk/issues/issue3_1/noble/index.html.

Prasad, Ganesh C. 1999. "A Practical Manager's Guide to Linux," OsOpinion.com. Available on-line at http://www.li.org/li/resources/papers/1999-pracmgr/Manager's-Guide-to-Linux.html.

Robiette, Alan. 1999. "Value for Money Considerations in Software Strategies for Higher Education," JISC Technology Applications Program (JTAP). Available on-line at http://www.jtap.ac.uk/reports/htm/jtap-029.html.

Vee, Danny. 1999. "Development, Ethical Trading, and Free Software" (available on-line at http://www.anatomy.usyd.edu.au/danny/freedom/ip/aidfs.html).

Vermeer, Martin. 1998. "Unix as an Element of Literacy," Linux Today. Available on-line at http://www.geocities.com/SiliconValley/Garage/9032/articles/unixasli.htm.

Watkins, Kevin. 1999. Education Now: Break the Cycle of Poverty. Oxfam International. Available on-line at http://www.caa.org.au/oxfam/advocacy/education/report/index.html.

Wilson, Greg. 1999. "A Natural Home for Open Source," Dr. Dobb's (October 8, 1999). Available on-line at http://www.ddj.com/articles/1999/9975/9975q/9975q.htm.