Penguin on Thin Ice?

Why IBM Should Win in the Fight to Save Linux

Anupam Chander

June 26, 2003

Earlier this year, SCO, a Utah-based software company, filed a multi-billion dollar unfair competition lawsuit in Utah federal court against IBM, one of the world's leading information technology companies.

The suit arose because IBM has made a strong push towards using the ever more popular Linux operating system for computers. (It's not alone; Linux is used by millions of people the world over, in companies, government offices, and private homes.) But SCO claims - astonishingly - that Linux is an illegal derivative of software SCO acquired from others years ago.

Though IBM is the only defendant, SCO's suit threatens Linux as well. Thus, the suit has captured the attention of the computing industry. Indeed, it has even inspired comedy - in the form of a parody of SCO's CEO [ http://lwn.net/Articles/34848/ ] set to a Gilbert and Sullivan tune. But at the same time, it has also generated FUD--"fear, uncertainty, and doubt"--among Linux users.

SCO has benefited from its decision to file the suit. During the last year, its share price had dropped to as low as $0.60 a share. Since the day its suit was filed in March, SCO's stock has traded up, currently trading above $10.00 a share.

But that rise in stock price may be illusory, for it seems unlikely that SCO's suit has any true merit. Rather than unfair competition, what likely occurred between SCO and IBM was healthy competition - the kind that helps consumers, and the economy, by pushing quality up and prices down, and increasing efficiency.

A Brief History of Operating Systems, from AT&T to the Penguin

To understand this suit, it's necessary to review a bit of computing history that now seems quite ancient.

Three decades ago, AT&T created a computer operating system called UNIX to run its telecommunications network. (An operating system is a computer program that allows one to run other, more specialized, computer programs. Microsoft Windows is today's popular example.) UNIX was widely adopted by corporations and other institutions. For the most part, AT&T and subsequent groups developed UNIX in a proprietary form--claiming exclusive rights to it.

Then, about a decade ago, a young Finn named Linus Torvalds introduced an operating system (named Linux, after its creator) that did some of what UNIX did. Linux proved remarkably successful. The community of Linux users adopted a mascot--the penguin - as a sort of trademark.

Unlike UNIX, Linux was open and free. Anyone the world over could examine its code and improve on it. A key to Linux's success was that its license (dubbed a "copyleft" instead of a "copyright") required users also to give away their own modifications to the software for free.

The joint collaboration of programmers around the globe, most of whom would never meet, created a powerful operating system.

In 2000, rather than fight Linux, IBM decided to embrace it, committing one billion dollars to developing the software. It installed Linux (not Microsoft Windows) on many of the personal computers it sold and Linux (not UNIX) on its large mainframe computers.

Meanwhile, over time, SCO acquired certain rights to other companies' versions ("flavors" in the lingo of techies) of the original, proprietary operating system UNIX. And even as it was developing UNIX itself, IBM licensed some of these flavors from these companies.

Finally, to further complicate matters, IBM also collaborated with SCO on yet additional enhancements to UNIX. But in 2001, IBM decided to abandon that collaboration.

SCO's Suit Against IBM: The Proof of the Pudding Is In the History

All these facts provide the background for SCO's current lawsuit.

SCO's theory is as follows. When IBM originally licensed UNIX, it signed contracts with companies such as AT&T. But IBM violated those contracts, later, by using information it gained working on UNIX to improve Linux.

Why does SCO care if IBM violated its contracts with AT&T and other companies? Because SCO claims that it is a "successor in interest" under those contracts. While it did not write UNIX, it claims to have acquired rights to the operating system over time from other companies.

Is SCO correct? Only close scrutiny of all the relevant history and relevant contracts can determine that question. Meanwhile, several legal principles make the situation even more complex.

Several Key Legal Principles May Undermine SCO's Case

The first principle - which is also common sense - is that one cannot sell (or give away) a proprietary interest that one no longer possesses.

So even if companies purported to give UNIX rights to SCO, their supposed grant of interest would be void if they earlier had given the same rights to IBM or if they had otherwise relinquished them. You can't sell what you no longer own.

The second principle is that a party's rights can be affected by its later conduct - which can constitute a "waiver," giving away rights. Until recently, SCO was a willing player in the Linux movement, releasing code under the open source ("copyleft") license. Everything that happened to Linux was in the open. Yet SCO delayed in suing.

That delay triggers not only the waiver doctrine, but also similar equitable doctrines such as laches. Indeed, SCO may run afoul of the relevant statutes of limitations as well.

The third principle is that what's good for the goose is good for the gander. Here, IBM is not the only one who has to abide by its contracts; SCO does too. This may cause SCO problems.

When it used Linux in software it released, SCO - like everyone else - had to agree to Linux's "copyleft" licenses. In doing so, SCO arguably gave up prior rights: "Copyleft" licenses, as noted above, gave everyone the right to copy code freely, and to make improvements on it.

The fourth principle is that a suit must not only establish a right, it must also show the defendant violated it. The court must ask: Even if SCO had extensive rights to UNIX and never waived them, did IBM actually violate those rights by supporting Linux?

Again, the answer lies in the history and contracts, but a certain degree of skepticism seems warranted. Would IBM really have agreed to give up developing other operating systems, in exchange for being able to work on UNIX? IBM is, after all, the pioneer of such operating systems as OS/2.

Could Linux Only Have Come From UNIX Know-how? It Seems Unlikely.

Meanwhile, at the crux of SCO's suit is its claim that its predecessors - the companies from which it says it purchased rights to UNIX - had crucial know-how. Indeed, it says this know-how was so crucial that it was necessary to Linux's success.

In its complaint, SCO claims, in essence, that without its UNIX contracts, IBM was nothing, and without IBM, Linux was nothing. But both of these claims, like SCO's allegations, are dubious.

First, let's look at Linux: Was it nothing before IBM? Of course not. Engineers collaborating through the open source movement created extremely sophisticated work even without a major corporate patron.

SCO alleges that "as long as the Linux development process remained uncoordinated and random, it posed little or no threat to SCO...." But in truth, Linux was always coordinated - just by many different hands. Linus Torvalds himself calls the Linux development process "directed mutation." [ http://kerneltrap.org/node.php?id=11&PHPSESSID=79f4b8bb7a33a40f5d09ecc5fc318335 ]

Second, let's look at IBM: Was it nothing without the know-how it gained through its UNIX development contracts? Hardly.

IBM is hardly a novice when it comes to systems engineering. It has extensive experience with Intel-based microcomputers (SCO's original specialty). Indeed, more than two decades ago, IBM introduced the personal computer based on the Intel microprocessor. Plainly, IBM's engineers are quite capable of invention and advancement, even without SCO's assistance.

From A Policy Perspective, Closed Software Is Perilous

How does the suit fit into the open software versus closed software debate? I believe the answer is that it demonstrates the virtues of open software.

Some would disagree, however. On its website, SCO quotes a suggestion that computer "[s]ystem administrators must be admonished to submit open-source code to inspection for potential violation of patents."

And Microsoft has recently been arguing that the SCO case proves that open source software is bad policy for software corporations. According to Microsoft, open source software may lead to repeat patent violation: After all, as the SCO suit shows, you can never be sure where a piece of code, contributed by a volunteer, came from.

But in truth, it is closed, proprietary software that is the real villain here. After all, partnering with SCO and its predecessors on UNIX projects was what got IBM in trouble in the first place. It's the proprietary claims over UNIX that form the basis of SCO's suit.

More generally, companies trying to derive more revenue from their intellectual property portfolio may lash out at licensees. But licensees of open source software distributed under a permissive license do not have to worry about this possibility.

For instance, might Microsoft someday claim that software using Windows "derives" from Windows and unfairly competes with Microsoft's business? Because Windows software is proprietary, the possibility never can be fully ruled out.

Proprietary interests, especially in intellectual property, tend to breed confidentiality - as anyone who ever signed a nondisclosure agreement with a fledgling dotcom with a "brilliant business idea" well knows. And confidentiality, in turn, breeds conspiracy theories, and allegations of theft - or unfair competition, or breach of contract, or the like.

In contrast, if there is something amiss in open code, it will be more difficult to hide. Thus, Linus Torvalds retorted, "I allege SCO is full of it and that the Linux development model is the most transparent process in the whole industry." Torvalds himself has provided the most compelling reason SCO's suit must fail: Had information truly been misused in Linux's open source software, everyone and his brother - and his foreign penpal - would have been well aware of it.

The final policy argument in favor of open source software is, of course, societal. At some point, information that is widely studied in universities, reprinted in college textbooks, and advanced through academic scholarship must be considered public domain.

For this reason, SCO's claims that its intellectual property rights extend to basic computing features of large operating systems cannot be allowed to stand. Otherwise, there will be no such thing as truly open, free software - and as a consequence, there will effectively be an economy-dragging tax on information technology.

Anupam Chander is Professor of Law at the University of California, Davis, School of Law. A graduate of Yale Law School and Harvard College, he specializes in cyberlaw and international law. Chander refers readers interested in following the SCO saga further to Slashdot.org - itself a demonstration of the power of dispersed individuals working together.