SCO to IBM: I’ve Shown You Mine, Now Show Me Yours

SCO has at last turned over a list of files and code snippets it claims violate its intellectual property rights.

If The SCO Group’s $3 billion lawsuit against IBM Corp. is a game of intellectual property (IP) poker in which—as some industry watchers argue—SCO has long been bluffing, the embattled vendor last week showed its hand when it handed Big Blue a list of the files and snippets of code that it claims violate its IP rights.

SCO’s disclosure came in response to a ruling from U.S. Magistrate Brooke C. Wells, who in early December found that IBM was entitled to see more specific evidence of the code misappropriations that SCO has charged it with.

Even so, SCO missed the thirty-day headline imposed by Wells. In a declaration that accompanied its response, SCO general counsel Ryan Tibbitts claimed that he was unable to coordinate information gathering efforts in time to comply with the request—even though he apparently didn’t begin his efforts until December 12th, or one week after Wells handed down her decision. “In light of the traditional holiday hiatus from work, and the unavailability of some of those directors during the holidays, I undertook the best reasonable efforts to gather information responsive to IBM's requests from SCO's outside directors prior to this Court's deadline of January 12, 2004,” the declaration says.

SCO’s disclosures were also incomplete, Tibbitts acknowledged. “All of the information has been gathered and is currently being processed for the remaining individuals. None of these individuals whose documents are not yet available are the ones that IBM identified as seeking the documents in an expedited manner,” the declaration says.

SCO claims that it will deliver the missing files to Big Blue before the next court date, scheduled for January 23.

Elsewhere in the declaration, Tibbitts claims that because SCO doesn’t have access to the most current version of AIX, it can’t adequately assess the extent to which IBM may have misappropriated code from AIX into Linux. “I have been informed by SCO's engineers and consultants that since the only version of AIX source code that was available for comparison purposes is several years old, and predates most of IBM's contributions to Linux, it was not possible to directly compare IBM's contributions to Linux with the most likely source of those contributions, namely the missing versions of AIX (including the most recent versions),” the declaration reads.

Nevertheless, Tibbitts purported to cite a variety of alleged code misappropriations, including:

  • "Read Copy Update" code—which SCO claimed IBM contributed to Linux 00 that “was copied substantially verbatim with only relatively minor changes from Dynix/ptx.”

  • Journaling File System code (JFS), which SCO also claimed was contributed to Linux by IBM, “was almost certainly copied and adapted for Linux from a version of AIX more recent than the one available for comparison.”

  • Enterprise Volume Management System—which, again, SCO alleges IBM contributed to Linux—“is based on the same architecture and data structures as are present in the AIX Volume Management System, and was therefore copied from AIX.”

In a couple of cases, Tibbitts’ declaration appeared to damn IBM by dint of association. “The AIO code contributed to Linux by IBM was written by an engineer who had a detailed knowledge and familiarity with the same area of technology in Dynix/ptx, and who likely used the same methods and/or structures in the AIO Linux implementation,” the declaration reads.

Since filing its $1 billion lawsuit last March, SCO has tuned the thrust of its legal argument to center on issues of so-called “derivative works,” in which SCO argues that it owns the rights to literally anything built on top of, or as a result of exposure to, its Unix System V source code and libraries. Participants in the community—where Tibbitts’ declaration has been posted in its entirety and exhaustively commented upon—note that the embattled vendor’s claims about IBM’s contribution of JFS to Linux could provide a test of this claim.

According to a mini-FAQ on JFS available from IBM, Big Blue claims that the version of JFS in Linux is based on a JFS2 port that it originally created for OS/2. ( In his declaration, Tibbetts requested access to “All versions of JFS, whether a part of AIX or not (including, but not limited to the AIX and OS/2 versions), together with documentation and programmer notes from the development process”—which suggests that SCO is aware of this lineage and will attempt to assert its derivative works claims in this regard.

About the Author

Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.