Oh, That's Why They Are Saying That
I figured out finally why SCO is making the wild claim about terrorism. It's just legal strategy: They can't revoke an irrevocable, perpetual license unless they can say that IBM violated the contract. (I don't know if they can say it even then, not having read the contract, but this is what SCO is claiming they can do.) It must have been hard to find a violation, so they came up with the idea of saying IBM violated the law, and so in that way, it broke it's contract. Specifically, they are claiming that IBM violated export laws by making multiprocessor operating systems technology available "'for free distribution to anyone in the world,' including residents of Cuba, Iran, Syria, North Korea and Libya, countries to which the United States controls exports. The open-source technology IBM released 'can be used for encryption, scientific research and weapons research,' the suit said. " Shankland has updated his article.
"SCO also detailed one element of technology that it asserts IBM copied, the Remote Copy Update ( RCU ) system, for relieving some memory bottlenecks on multiprocessor servers.
"The amended complaint includes an IBM copyright on the RCU technology that names the an engineer as the author, with work "based on a Dynix/ptx implementation by Paul Mckenney (sic)." Dynix/ptx was Sequent's version of Unix for servers with multiple Intel processors.
"It appears that RCU indeed stems from work in Dynix/ptx. In a paper on his Web site, IBM's Paul McKenney says RCU was included in Dynix/ptx in 1994. And the Linux Scalability Effort's Web site says that RCU patches are "based on original DYNIX/ptx code (released by IBM under GPL)"--the GPL referring to the General Public License that covers Linux. Torvalds accepted RCU into the Linux kernel in October 2002."
The document also attacks Linus Torvalds by name, saying he is to blame and that IBM knew he wasn't doing a good job of keeping proprietary code out of the kernel. Linus just announced he will take a year off from Transmeta to work on the kernel full-time. This is, from a PR point of view, a good thing. Now he will be completely focused on the kernel. I think businesses will be relieved and reassured. It's not hard to figure what kind of work that will be, and I hope he finishes before this case even goes to trial.
Most civil cases of this kind take about 18-30 months to get to trial. With some of IBM's affirmative defenses, I'm guessing it'll take longer. I also believe they will be adding some counterclaims. So even when they arrive at the first day, it'll be a lengthy trial in all likelihood.
Poor Linus. This probably isn't the fun part.