I live and work in Austin, an IBM city of sorts, and have spoken with many former IBM workers about the SCO claim and after the tenth person laughed it off I started to look into it a little further. It’s not simply improbable, it really does seem impossible for this to happen as claimed.

The claim is pretty simple. IBM licensed some elements of UNIX from Novell. SCO thinks it bought that license. SCO thinks that IBM let code go from the licensed AIX UNIX product into the public-sourced LInux product it was helping to develop (since both run on and are shipped with IBM servers).

The problem with this is that IBM is an old company. They aren’t careless and they aren’t oblivious to the importance of contracts and intellectual property. They’ve been around for decades longer than Microsoft, Novell, and certainly longer than SCO. They’ve done this before.IBM helped out with the Samba code in the past as well and I’ve spoken with several people about how that went down. As it was told to me, each person that joined that team lost all access to the AIX source when they joined the efforts to help Samba. More so, every single code change that left IBM had to be approved by several individuals, one of which was a lawyer. An engineer helping with any public-source project was overlooked overseen heavily and every effort was made to prevent anything proprietary, especially licensed, from leaving the company.

With that kind of oversight for something as unique as Samba, I’ve no personal doubts that IBM is innocent of these silly charges. I also do not believe some claims that Unix took from LInux. I honestly feel that Linux did, in fact, take from Unix but I find the most plausible cause yet was SCO actually putting the SCO source up for download for anyone interested. Let me say that again: SCO gave the source code away for free to anyone who clicked on a link. (To be completely fair, there were some pretty clear terms one had to agree to, and one of those was that the source and derivative products would not leave the computer it was downloaded onto.)

That rather strikes me as reasonable doubt and I’m sure IBM sees it the same way. The same code they licensed (or close enough) was available to the public at-large for years (the archive is about three years old). There’s no telling who added it … except via CVS. Which explains why SCO will not divulge the code, even to IBM. If you know the file, and you know the code, the CVS log is public information and they can simply, and quite easily, go and see who put it in there.

And it likely was not IBM. If it was, SCO would have pointed to the CVS logs already.