SCO Shares Linux Code

The SCO Group is taking its case against the Linux operating system and IBM on the road.

Last week, the company began showing code to U.S. analysts that, it claims, proves that the source code to the Linux operating system contains sections of code lifted directly from SCO's Unix code base.

SCO's presentation, which has been seen by analysts at Gartner and Aberdeen Group, features pieces of software that IBM contributed to the Linux code base, SCO said. That code was derived from SCO's Unix code, itself derived from AT&T's System V Unix code, and then distributed to the Linux community in violation of SCO's licensing agreement, the company said.

SCO will spend the month of June showing its evidence to analysts and journalists who are interested in seeing it, said SCO's general manager of SCOsource, Chris Sontag.

Copyright Concerns

"The one specific example that I'm showing right now is [Unix] code, line by line copied into Linux," said Sontag. He said that this code, which SCO considers to be a copyright violation, is identical in both the Linux and SCO code base.

Sontag would not say which specific parts of Unix SCO believes were copied, but he did say that the work appeared to have been done by a company other than IBM.

SCO sued IBM for $1 billion this March, charging the company with misappropriation of trade secrets, tortious interference, unfair competition, and breach of contract.

Making Its Case

SCO's case against IBM does not involve copyright law, which, Sontag said, is harder to prove in court. Instead, it revolves around contract law claims. Specifically, SCO claims that IBM's 1985 Unix license, originally signed with AT&T but subsequently transferred to SCO, prevents IBM from distributing software derived from the Unix code base.

IBM's contributions to Linux in the areas of non-uniform memory architecture (NUMA), symmetric multiprocessing (SMP), and a journaling file system all constitute breaches of that contract, Sontag said.

SCO plans to revoke IBM's Unix license on Friday, "as we have the contractual right to do, due to their breach of the contract that they have with us," said Sontag.

IBM has denied engaging in any wrongdoing in the case.

Burden of Proof?

Open-source advocates now say that the burden of proof is on SCO to show that the code was copied by Linux developers and not the other way around. "Something I strongly suspect is going to come up in the court case is, how do we know what the direction of transmission was?" said open-source advocate Eric Raymond.

The fact that SCO has been an active contributor to the Linux kernel means that SCO has a harder case to prove, Raymond said. "The burden of proof is on [SCO] to demonstrate that the transmission went from System V to IBM to Linux, rather than from System V to SCO's own kernel developers to Linux," he said.

SCO was able to uncover the alleged violations by hiring three teams of experts, including a group from the Massachusetts Institute of Technology math department, to analyze the Linux and Unix source code for similarities. "All three found several instances where our Unix source code had been found in Linux," said an SCO spokesperson.

"They're claiming that they can point to dozens of files, and that the code lines vary from a few chunks to big blocks of code," said Gartner analyst George Weiss, who was briefed by SCO last week.

Open-source advocates say that it is no surprise that Linux and System V share similar code, since they both include components of the Berkeley Software Distribution (BSD) code base. "It's clearly a fact that [Unix] contains a lot of BSD code, and Linux has some too, so there are bound to be...lines of code in common," wrote Linux creator Linus Torvalds in an e-mail interview.

SCO has offered to show the code in question to Torvalds, but he has not yet accepted the offer, said Sontag.