Torvalds Slams SCO

In a hard-hitting interview with eWEEK, the father of Linux slams SCO's 'evidence' of illegal Unix code in Linux.

Linus Torvalds, the creator of the Linux operating system, this week responded to a range of allegations made by The SCO Group at its conference in Las Vegas. In a hard-hitting interview with eWEEK Senior Editor Peter Galli, Torvalds disputed what SCO presented this week as evidence of illegal Unix System V code in Linux.

eWEEK: What are your views on the lines of "offending" code that SCO showed this week and the thousands of other lines of code it claims are also illegally in Linux?

Torvalds: The code SCO showed represents an algorithm that can be used to manage a computers memory...Not a very interesting piece of code in itself, this is very basic "allocate a smaller chunk of memory out of a list of bigger chunks." The function is described in a lot of places, and exists in original Unix code and is apparently written by Ken Thompson himself. It shows up in the Lion book (a commentary on the traditional Unix), and the code is described in [Maurice J.] Bachs "The Design of the Unix Operating System." In other words, its not only 30 years old; its actually been documented several times. Its also part of BSD Unix, which was shown to not be a derived work of the AT&T copyrights 10 years ago.

Its part of the "original Unix" archives that Dennis Ritchie has made available, and from a legal perspective (and also of ironic interest), its also part of all the code that Caldera made freely available back when they still remembered that they were a Linux company and had made all their money on the Linux IPO. Ironically, the piece of code that [SCO demonstrated this week] had already been removed in [the Linux kernel] 2.6.x—and not because of copyright issues, but because developers complained about how "ugly" it was. So not only is the code available under the BSD copyright, it had been removed in new versions of Linux even before SCO made it public.

But what I find interesting is how it shows that the SCO people are having such a hard time with the truth. Theyve said several times that the code they have found is not "historic Unix" code and "not BSD" code (which they know you cant infringe, since BSD has been shown to be independent, and Caldera itself released the historic code in 2002). To counter the open-source peoples contention that any shared code is likely of BSD or "ancient Unix" origin, [SCOs] claimed several times how its "modern System V" code that they have clear ownership of. Thats despite massive proof to the contrary, going back three decades.

eWEEK: There is also a movement afoot in the open-source community, spearheaded by Eric Raymond and Jeff Gerhardt, to get SCO to allow the community to see the code under a less restrictive NDA so that any offending code can be removed. Do you support such a move?

Torvalds: Absolutely. If SCO can actually show code that is truly infringing, I (and a lot of other people) are going to figure out where it came from and remove the offending code. That goes without question. However, I clearly dont expect that to be the case. We expect to see more of the same: BSD code (or other code that is just commonly available to both parties, like the ancient Unix archives), or code that just looks similar because it is based on public standards.

So the main reason we want to see the allegedly infringing code is that we think its likely that its not infringing at all. And Im certainly willing to back that up with a promise to remove any code they point to that we cant show is ours or open.

eWEEK: For its part though, SCO has said that there are so many lines of code, and a variety of applications and devices that use that code, that simply removing the offending code would not be technically feasible or possible and would not solve the problem. Do you agree?

Torvalds: They are smoking crack. Their slides said there are [more than] 800,000 lines of SMP code that are "infringing," and they are just off their rocker. The SMP code was written by a number of Linux people I know well (I did a lot of the SMP IRQ scalability myself, personally), so their claims are just ludicrous. And they claim they own JFS [journaled file system technology] too. Whee. Theyre not shy about claiming ownership of other peoples code—while at the same time beating their breasts about how they have been wronged. So the SCO people seem to have a few problems keeping the truth straight, but if there is something they know all about, its hypocrisy.


eWEEK: SCO and its lawyers also say that even if that were a workable solution, they would still want damages for the illegal use of their code in Linux until the "fix" was implemented. They say Unix code has been in Linux since 2001 and that vendors and end users have been profiting from this since then, and they want to be compensated for that. Who, they ask, would compensate them under this scenario. Your thoughts on this?

Torvalds: Hey, until they can be bothered to show something real, I dont think its even worth discussing.