Linux and other Free and Open Source Software (FOSS) enjoy a reputation for ubiquitous use in educational settings. While FOSS openness and low acquisition costs resonate with the approach and needs of academia, it's proving difficult to establish a clear adoption trend. Certainly there exists ample anecdotal evidence of adoption, school-by-school, department-by-department. Certainly a range of Open Source projects arise from and also target education. However, close investigation reveals a mixed reality for Linux and FOSS in education: perusing college course listings, at least in the United States, doesn't support the notion of near-universal Linux/FOSS use across curricula - either in computer science or as a platform across other disciplines. It's even more difficult to measure Linux and FOSS adoption in K-12 settings where course catalogs don't exist and where classroom IT closely tracks home computing use.

High Times or Hard Times?
In Hard Times, Charles Dickens introduces his monomaniacal schoolmaster Thomas Gradgrind as "a man of fact and calculations;" the character is most memorable for his constant exhortations to his charges of "fact fact fact!" in response to the slightest hint of fancy.

Now, what I want is, Facts. Teach these boys and girls nothing but Facts. Facts alone are wanted in life. Plant nothing else, and root out everything else. You can only form the minds of reasoning animals upon Facts: nothing else will ever be of any service to them. This is the principle on which I bring up my own children, and this is the principle on which I bring up these children. Stick to Facts, sir!

The goal of this article is neither to debunk notions of Linux and FOSS adoption in education nor to evangelize such adoption, but rather, in Gradgrindian fashion, to temper the often-unqualified enthusiasm that seems to surround the topic with sometimes sobering fact.

Let's start the investigation by breaking Linux/FOSS adoption in education into three focus areas:

• University computer science curricula
• Cross-discipline adoption
• University infrastructure

University Computer Science Curricula
For the past few years enrollment in computer science and related fields has been in decline in the U.S. This trend dipped further after the terror attacks of September 2001 and subsequent restrictions on student visas. Such attrition is having an impact on all kinds of IT competence, open and proprietary - fewer students mean fewer courses; fewer courses lead to scarcity of graduates to fill the IT and developer labor pool. While the impact is significant, computer science departments haven't shriveled up, just "right sized," and emphasized vocational offerings over computing theory.

In the U.S. published course descriptions belie the notion of Linux ubiquity in computer science curricula. Budding developers, IT managers, and ordinary users of productivity applications are more likely to find courses that explicitly target Microsoft Windows-family operating systems, Windows infrastructure (networking, administration, etc.), and Windows-hosted MS Office applications like Excel, especially in undergraduate settings or at junior colleges.

Table 1 illustrates the situation at six institutions of higher learning in Silicon Valley: two private schools (Santa Clara and Stanford), two California state schools (U.C. Berkeley and San Jose State), and two junior colleges (De Anza and Foothill). The table tallies the total number of classes offered by each CS department and summarizes the number of classes focusing on Windows, Unix, and Linux (based on examining course name, description, and when possible, syllabus). I chose the Valley because it's where I live and work, where I went to school, and a place whose technology culture is presumed to be well-attuned to Linux and FOSS.

Other findings included:

• Most CS classes are not OS-specific (at least they're not publicized as such)
• Universities emphasize theory while two-year college curricula veer toward pragmatic and vocational subject matter
• Operating systems classes tend to use non-commercial/non-standard platforms (not Linux, not Unix, not Windows); they may or may not examine Linux or other FOSS OSes
• A reasonable number of classes (at SJSU, Santa Clara, and De Anza) focused on Unix, but in most cases, Linux subject matter was promoted as "Unix/Linux"
• No course descriptions at any of the institutions highlighted "Open Source" or acronyms like FOSS

Beyond Course Catalogs: Finding Linux in Labs and Dorms
However, other anecdotal evidence at these six Bay Area schools indicates that study of core computer science subjects like C programming, operating system design, networking, security, Web applications development, and system management increasingly occurs on and for Linux. For example, several syllabi at Stanford made reference to assignments running on Linux; reading lists for computer security classes at multiple schools included texts centered on Linux-based firewalls; and data from university IT departments indicate that in many cases academic data centers that run on Linux also support it for individual students and cross-departmentally (see below).

As such, even undergrads are reasonably likely to find Linux in the lab and can install and use Linux on their own PCs and notebooks; this trend continues into graduate education, especially for supercomputing and other "exotic" disciplines. Owing to its openness, computer science graduate students today frequently conduct original research and write theses based on Linux itself or using Linux as an enabler, especially in areas like scalability, high availability, alternate scheduling schemata, queuing theory, etc. No metrics are available to substantiate such anecdotes.

The situation is reputedly different in Asia and across a range of developing countries. Conventional wisdom holds that in these countries by contrast, Linux explicitly forms the core of much official computer science curricula, especially in countries with under-funded public higher education and where governments see Linux as a means to local technology development and digital inclusion.

To test these assumptions, I visited the course listings for CESET - the Centro Superior de Educação Tecnológica at UNICAMP - the University of Campinas in Brazil, where IBM sponsors a Linux lab (I'm fluent in Portuguese, but not Mandarin Chinese!). There I found a listing of 45 courses and no references to any operating systems of any kind, open or closed, in the class listings and descriptions.

The obvious conclusion is that university curricula tend to emphasize theory and give minimal grounding in real platforms. Certainly my own undergraduate experience (decades ago) followed this tendency - at Cal I used versions of Berkeley Unix in most of my classes, but had zero training in the particulars of that OS; it was sink or swim.

This imbalance between theory in school and practice post-graduation doesn't bode well for Linux and FOSS. Graduates will arrive in the job market with ample theoretical knowledge of computer science, but scant formal exposure to Linux and FOSS. Without more visible focus on FOSS technologies and exposure to open community practices in colleges and universities, graduating seniors are much more likely to possess practical familiarity with the current dominant proprietary productivity OS - Microsoft Windows, even if they use Linux in some lab settings.

Linux and FOSS across Higher-Education Disciplines
If de facto Linux and FOSS use in the U.S. is occluded in computer science settings, it can be hard to see in disciplines further removed from the CS lab - but it's there. The situation is further complicated by the fact that Linux deployment across academic fields isn't predicated on the OS itself, but rather on the availability of domain-specific applications, both open and proprietary, hosted on the OS.

The good news is that Linux benefits from straightforward migration from legacy Unix, a path that ISVs and end users are taking in droves. As a result, we see the Open Source OS appearing in areas as diverse as library science to linguistics, from medicine to mathematics, and from industrial design to industrial control. The number of Open Source projects and applications addressing these fields speaks to the number of developers and users.

For example, Linux-based Open Source library catalog and search applications support library science curricula and school library operations, several from institutions in Brazil; AI and speech processing software built on Linux has emerged from schools like Carnegie Mellon, MIT, and the University of Pittsburgh. Medical and biology programs benefit from Linux-based gene-sequencing software that leverages low-cost Linux-based scalable clustering. Mathematical proofs and visualization also use software that runs on Linux. In fact, Linux has taken over most if not all legacy Unix workstation applications with ISVs porting their CAD, CAE, and CAM packages to lower-cost, more powerful commodity Linux hardware. And any and all disciplines that rely on embedding intelligence in complex systems, including automotive design, nuclear engineering, mechanical engineering, and robotics, today develop on Linux workstations and then proceed to deploy the OS in embedded form and require an introduction to it in preparatory education.

Linux and FOSS in University and College IT Infrastructure
OSDL has perhaps had the greatest direct exposure to the category of data center use by institutions of higher learning. In fact, OSDL membership roles boast a large contingent of Academic affiliates (see Table 2).

OSDL academic membership suggests a motivation for migration to Linux that echoes corporate IT - cost of acquisition, lower cost of management, and the ability to reuse or prolong the useful lifetime of legacy hardware.

Colleges and university IT staff doesn't look at Linux in a vacuum, but use Linux, the LAMP stack, and key tools like OpenLDAP to manage enrollment and school records and supply secure e-mail and Web access to their student and faculty populations. In this last area, Linux shines above and beyond competing solutions in terms of performance and scalability: even "enterprise-ready" Windows platforms and applications can't easily support bases of 50,000-250,000 user with a 25% turnover every 90 days!

One of the better-documented and longest-term university deployments of Linux and FOSS took place and continues to grow at Boston University. There the IT staff supports a mix of 40,000 students, faculty, and other staff by leveraging a mix of Open Source Linux and platform software sustaining a mix of open and "vendor source" middleware and applications. According to Gerard Shockley, BU assistant director of technical services, his IT department can provide services to a diverse user community and still engage in community-based Open Source development like OpenAIS, Tomcat, Kuali, and Eclipse. BU looks to a mix of resources to accomplish its goal of using Open Source for mission-critical applications: internal IT competence, some academic interface, a range of community resources, and IBM Global Services as a commercial supplier. Shockley emphasizes the benefits of the community interface - "Open Source," he says, "is like having additional staff."

Conclusion
One source for this article responded to the paucity of publicized courses by describing Linux and Open Source as "silent leaders" in academia. My own mostly positive view of the presence of Tux and friends on campus comes from interviewing and hiring dozens of highly qualified college interns for Linux and FOSS software development, QA, and technical marketing roles. However, many were called, few were chosen, and those chosen few brought their FOSS savvy mostly from self-directed study.

Warming to Linux and FOSS
The question of Linux and FOSS ubiquity in higher education resembles the on-going debate around global warming. Intelligent and otherwise well-informed sources proclaim that Open Source adoption, like average world temperatures, is on the rise. However, definitive evidence for such claims can be confusing or unavailable. Pockets of fervent adoption and development may or may not indicate a generalized climate that favors Linux and FOSS. Conversely, icy stares from CS faculty and tepid investment in curricula aren't proof that Linux and FOSS aren't HOT.

Rather than debating the issue of Linux and FOSS in academia, promoters of Open Source should instead return to their alma maters to ensure that adoption is occurring, and to evangelize and support adoption at their old schools through words and deeds and cajoling and (of course) corporate donations. And when your own children leave the nest for college, buy them a Linux notebook. Help them search beyond curt and cryptic course listings in computer science or any other chosen field of study. Help them find schools, departments, research centers, and curricula than include and emphasize Linux and FOSS. Encourage them to be Open-minded and Linux-literate. A mind is a terrible thing to waste on proprietary software.