), a small Utah software company, claims Linux is trampling on intellectual property rights it inherited from Novell (NOVL
), which got them from AT&T (T
). In an e-mail interview with BusinessWeek Correspondent Jim Kerstetter, Torvalds explains why he thinks SCO is wrong. The following are edited excerpts from that interview:
Q: SCO claims that old Unix files it says it owns are now in Linux. Can you explain to me why you think that's wrong?
A: [A number of files SCO claims to own] were written from scratch for Linux.... SCO also doesn't hold any copyrights to the BSD code [software developed at the University of California at Berkeley that SCO says contained copyrighted material that was passed on to Linux], nor is it actually in [SCO's version of Unix]. So SCO is wrong.
Also, SCO has apparently several times mentioned how copyright notices have been removed. Just for the record: Original Unix doesn't have any copyright notices to remove. They were added after a lawsuit [between the Berkeley developers and AT&T, which was settled]. So SCO would be wrong again.
So basically SCO's arguments are just too wrong to even discuss rationally. SCO doesn't own the copyright on the files they are talking about -- the University of California at Berkeley does. But even if they did, the Linux files weren't even copied in the first place. And even if they had been copied, no copyright notices would have been removed, since they didn't exist in the original. There are literally several levels of SCO being wrong. And even if we were to live in that alternate universe where SCO would be right, they'd still be wrong.Q: In fact, I saw in a recent interview that you chided yourself for the quality of some of those files. Why so?
A: Hey, for some of the files they claim copyright ownership on I went back 12 years in the archives to see their original form, and the fact is, I was a young guy at university in '91, and I [made] mistakes that I simply wouldn't [make] anymore, and that are clear signs of beginner [programming].
And those mistakes show how the code wasn't copied -- it's a bit like how map makers used to introduce small errors in maps on purpose, so that if somebody copied them but claimed to have made their own, the original map maker could point to the error and say, "Well, how did you have exactly the same error, too?"
Except I can definitely state that I didn't make those mistakes on purpose. As a young student at the University of Helsinki, I definitely didn't have the kind of forethought required to foresee a company claiming my code as theirs 12 years later. If I had those kinds of powers, I'd never have gone into programming, I'd just play the stock market.Q: Do you think that any copyright or patent-protected Unix code has actually found its way into Linux?
A: Unlikely. There are now a number of people who have access to both Unix sources and Linux code, and literally written automated tools to find similarities. They found something like 30 lines from [Silicon Graphics, SGI
] that were dubious and that had been removed already. SGI wrote an open letter about their mistake. You can find it if you look for it.
As to patented algorithms, yes, there are a few examples of that -- IBM (IBM
) actually explicitly licensed some to Linux. That was a requirement on our side for even accepting the code in the first place. SCO doesn't own any patents, so they certainly can't be claiming ownership.Q: If there is protected code in Linux, is there a solution?
A: Oh, the solution to any patent/copyright dispute is licensing the [intellectual property] or not using it. It's that simple.
In this case, we would clearly just remove it, but SCO has been less than forthcoming about what the contested code would be -- and when they do mention code, we can prove they are [wrong]. But we've always said in public that if SCO can actually show that somebody has inserted SCO [intellectual property] improperly in the [core of Linux], we will remove it. SCO only needs to ask.
But of course, those two scenarios actually depend on the [intellectual-property] claim being valid. That's a much more fundamental issue for SCO. The validity of their claims has always been very shaky, even regardless of the fact that Novell (NOVL
) claims SCO doesn't own the Unix copyrights in the first place.
The SCO claims have been shaky from the start because they haven't actually been able to show any particular copied code. It's like me claiming copyright on some article you wrote for BusinessWeek [without being] able to specify which article and which part of it I would have written. The fact that Novell now contests the SCO copyright ownership just makes them even more shaky.Q: What do you think of SCO's lawsuit against IBM and its threats against customers?
A: Quite frankly, I obviously don't think much of it. But clearly we have to be careful.... Any business in the U.S. ends up having to worry about legal action, whether merited or not.
So I can't say that I'm ignoring it, because I'm not. But I do consider the thing to be a pretty clear-cut fishing expedition on SCO's part.... [But now,] they can't just back down, because that would show how [questionable] the original stuff was.
Nothing to lose is a bad situation to be in. They're a cornered rat, and quite frankly, I think they have rabies to boot. I'd rather not get too close to them.Q: What do you think is motivating them?
A: I think there was a fair amount of bad feeling when IBM dropped out of the Monterey project [a joint-development project with SCO]. That was a big deal for SCO, and they had a hard time with that. Never mind the fact that it had long since become clear that the project wasn't going anywhere, and IBM would have been crazy to continue with it.
So you have some pent-up anger at IBM, a failing business that was losing its market, and put it together with a greedy new CEO who has fought legal battles before, and what do you get?Q: Do you find this to be a personal affront to the work you've done over the years?
A: I do, and I don't. I get really upset every once in a while when SCO makes some new totally outrageous claim about the code I originally wrote and have maintained for over a decade. And I get frustrated at how this charade has been going on for something like eight months now.
But at the same time, I have obviously done all the development and the project management over the Internet for those 12-plus years, and the thing is, I've had to get a really thick skin and be able to laugh at myself, or I'd never have been able to do it.
Which means that I'm actually pretty good -- or at least I think I am -- at taking a step back, and looking at my own reaction to it pretty neutrally. And that not only calms me down a fair amount, it has actually made me appreciate the situation a bit: The lawsuit has made me realize just how personal these things get and how badly you can react to them.
So I still get angry, and sometimes I vent through e-mail, but at the same time, it has been an interesting experience. Not something I'd recommend to others, and quite frankly, if I never hear of another Linux lawsuit again, that will be too soon. But it's a bit like being at the dentist and instead of thinking about the irritating whining noise, you appreciate trying to figure out what it's like to be the man with the drill.
It doesn't make the situation any more pleasant, but at least you can get something out of it. Learning how things work, and how you yourself react to stressful situations, is worthwhile in itself, I think.Q: Finally, have you done a deposition for SCO yet?
A: No, they asked for the moon and the sun -- I've got tons of e-mails -- and they weren't specific enough for me to automate [searches of my files to find answers to] their questions, so the lawyers have been trying to pinpoint a set of automated queries that both sides are comfortable with. And it wouldn't be a deposition, they just ask for documents. In my case, that ends up being basically e-mail.
I don't have anything against making my e-mails available per se -- most of my e-mails are public anyway. But I don't want to make it all available, [such as] private stuff that has nothing to do with what SCO is asking for. And since I've got so much of it, I can't read through it manually either. I wrote some tools to automate the search, but they would require pretty specific search terms.