We have received permission from Dan Farber to provide a transcript of a July 22, 2003 interview with Darl McBride, CEO and President of the SCO Group, by Dan Farber and Charlie Cooper of news.com. We have tried to be accurate, as always, and any mistakes are ours, not those of Mr. Farber, Mr. Cooper, Mr. McBride, or News.com. If you see any errors, do tell us, so we can perfect the transcript.
Our sincere thanks for allowing us to make this transcript available. You can find the video at this page, and then scrolling down. Thanks too, to the transcription group for working so hard to make this available.
By far the most interesting statement is McBride's characterization of Novell's copyright position, in light of later evidence that came to light, the Novell correspondence.
Because of what we now know, it's also telling to read the description of the vast copyright claims they began by asserting and comparing those claims to the very narrow and limited claims of today. I see that from day one SCO was talking not so much about Linux accepting stolen copyrighted code but about code SCO lays claim to under their derivative code legal theory, but presented in such a way as to make people think they were talking about Linus being asleep at the wheel. Oh, and their header file claims, which have been debunked thoroughly.
The understanding back then was that they were saying that there were flaws in the Linux code process. We all spent considerable time and effort to prove that there was no line-for-line stolen code in Linux, and that must have made them laugh, but in the end it has only strengthened the perception that the Linux open source system worked quite well indeed, with the result that there is no improper code, in the stolen code sense, and that the only way to attack it was to invent a new legal theory that no one else has ever gone by and then accusing Linux, or more precisely IBM, of violating this novel theory of what belongs to SCO.
The final interesting point from this interview is McBride's obvious misunderstanding of how the GPL works. When he makes the assertion that "Because we were distributing Linux does not mean we had donated code to Linux", he reveals he does not accurately grasp that their distribution triggered the GPL, no escape, no excuses.
Dan Farber (DF): Darl, Thanks for being here.
McBride: Hey, thank you guys. It's good to be here on your show today.
DF: You know it's been about 4 months since you launched your lawsuit against IBM. What precisely is the status of the lawsuit and what are those thousand-dollar-an-hour lawyers doing at this point?
McBride: Yeah, good question. The lawsuit was launched back in March. It was after a period of several months where we had found intellectual property violations and contract problems. We launched the lawsuit on March 7. We had a hundred-day notice in place at the time we put the lawsuit in place that if we did not have these contract issues resolved then we would in fact be canceling or revoking the AIX license. On June 13, we in fact did revoke the AIX license agreement as per our contract rights. With respect to the broader lawsuit that was filed on March 7, it is now moving into discovery and we're going to the next phases. So basically the legal teams on both sides are digging up information and, you know, we're going from there.
DF: Well, There seems to be some alleged code that's offending that is purportedly the property of SCO through all the rights they've gained in acquiring those versions of Unix. What precisely is the code that you're talking about? We've seen some stories where it's 80 lines here and it's just the comments. I mean how can you have a lawsuit like this and not be more forthcoming about what the offending code is?
McBride: Right, so when we started off this problem with IBM, it was very clear that IBM had donated things improperly into the open source community. That was the basis for our lawsuit against IBM, among other things, but that was the primary driver.
And during the period of time shortly after filing the lawsuit until recently when they came back and responded, we had a 60 day period there where we turned 3 different teams of code programmers loose on the codebases of AIX, Unix and Linux. And they came back with - independently - we had the three teams - one was a set of high-end mathematicians, rocket scientist, modeling type guys. Another team was based on standard programmer types. A third team were really spiffy on agent technology and how all of this technology was built in the first place. So the three teams came back independently and validated that there wasn't just a little bit of code showing up inside of Linux from our Unix intellectual property base. There was actually a mountain of code showing up in there. Now if you look at the types of code, we really see them in three different buckets.
DF: But, Where does all this code come from before. I mean Unix goes back to 1960. We're talking about code that could be as old as 40-some years old.
DF: So what's the prior art that you're going after at this point?
McBride: Good questions. So in 1996 our company paid, or in 1995 I guess it was, we paid more than 150 million dollars for the rights to Unix from Novell. So you could ask the same questions, what rights did we buy then? There very clearly are rights that are well beyond anything that's been out there in the public domain. You know, just because there have been versions or different iterations of Unix out there doesn't mean the proprietary or commercialized Unix has been in that bucket.
So, what we have is the main code base of Unix. We have the source tree here of System V if you will, that SCO owns, and then what you have are multiple branches off from that tree that go into many of the OEM vendors that have been selling Unix over the years. Unix last year was a 21 billion dollar marketplace. All of those versions of Unix tied back to SCO's branch. So what happens here is SCO owns the main source trunk and they have derivative rights control on the branches that are out there. So when you have an AIX or an HP-UX or Fujitsu or wherever the various version is coming from, they have the rights to go sell and promote their branches and the derivative works they do on those branches, but they don't have the right to donate that or give it away. They have to keep all of those works in confidence.
Charlie Cooper (CC): And how ... Darl let me ask you this, how then did you folks determine that this is worth one billion dollars plus in damages?
McBride: Yeah, there are a few different ways you get there. But let's go back to the basis of our claims which is there is a huge amount of code inside of the Linux kernel today that is improperly there that has come from system vendors that we have contracts with.
That... what that does is it creates copyright violations inside of there, some of it coming straight from our source tree. So yes, there are direct line-by-line codes.
DF: Well, Novell would say you don't actually own those copyrights fully.
McBride: Yeah well, the Novell thing. They came out and made a claim that held up for maybe four days and then we put that to bed. If you go back and talk to Novell I guarantee what they'll say, which is they don't have a claim on those copyrights anymore.
DF: Well, that's not exactly what they're saying at this point. But I think it's pretty clear that one of the issues here is saying that the open source community is purposely putting code into their open source Linux that is code that is not their own. Is that something that you're actually saying in your complaint, and will you end up suing the various Linux providers?
McBride: What we're saying right now is that we're seeing a lot of code. Let's go to the basis here. Linux 2.2 kernel goes up through the end of the millennium. 2.4 comes out early millennium. Since 2.4 has gone out, there have been a couple of million servers shipped. And if you look, if you do a comparative analysis, the code in 2.2 kernel is very simple, doesn't allow for a lot of scalability, not high reliability in terms of where this stuff runs. When you go to 2.4 and beyond, you're seeing Linux run inside of enterprises. Data-grade, carrier-grade Linux and the SMP - the symmetric multi processing, the ability to scale up to many processors in terms of a Linux box - is way beyond where it was in 2.2. One thing I will make clear here, Dan, is we're drawing a line on 2.2 and 2.4. The massive amounts of code we're seeing show up in Linux today that relates to our IP really came along the lines of 2.4 kernel.
DF: Let me just get a clarification here. Does that mean that you plan to levy a licensing fee on all commercial Linux users of 2.4?
McBride: We think that there are a lot of ways that we get justice. Our main thing right now - it's just a little bit like what's going on in the music industry, the film industry. It's very clear in the film industry when the video comes out two weeks ahead of the release date of the showing that there's been an infringement. It's very clear with the online music sharing there are problems. They're working through their issues in different ways. We very clearly have problems here. How we work through and get justice to that? We're open to as far as how that works out. Does it happen at an end-user level? Does it happen at an OEM level? Does it happen at a Linux distributors level? We're looking at all of those options right now.
CC: Darl, does it bother you at all that the company is now considered a pariah by the open source community?
McBride: Well, there's no doubt we're not winning the Miss Congeniality contest of the software industry now. But I think if you step back and look at it, in terms of if we really are right with what we're saying, and we're 100% convinced we are and the world is coming around to that. The few dozen people that have come in, taken the time to take a look at our code, take a look at our claims have unanimously walked out of our Lindon, UT offices with the same conclusion that, "Yeah, you guys do have some pretty powerful claims here." They then turn to, "What's next? What are you going to do next?"
So I believe that as our claims are now being validated and as we start to gain wins in this area, I believe that it will turn, it will come back round and people say "Well, OK, there were problems." In the end of the day though, what I would say is, we're not trying to kill Linux. I mean we're not attacking Red Hat right now which if we did, by definition, could end up shutting down Linux tomorrow.
CC: But what do you think the impact is on a CIO who reads that they may very well be vulnerable to litigation if they go ahead and deploy open source software?
McBride: I think the impact is, this situation potentially can get wrapped up a lot sooner than having this big cloud over the industry for what could take a couple years to go litigate with IBM.
DF: You mentioned that you have a lot of evidence as to your intellectual property that's being violated, but I so far haven't seen any of it. I do have a quote from someone who said allegedly that they did see it, an Aberdeen analyst, Bill Claybrook. He said he looked at the code, he saw some lines that looked similar, but it was not conclusive in any way.
McBride: Well, another report that I saw that Bill wrote basically said from what I saw there seem to be claims that seem to be very legitimate and if everything I saw was correct I think then the industry has a problem. So I think depending on which press report and how they want to twist his words, you come up with different conclusions. But very clearly the two or three I've seen and the report that he in fact wrote on it was very strongly weighing in on our side.
DF: It sounds like there's a lot of confusion over precisely, you know, is it a contract dispute, is it an IP dispute? Is there evidence, clear evidence that you've been able to present, which I haven't seen, that says that IBM in particular or the Linux community had violated any copyright or patent that SCO might have? When is that evidence going to be produced?
McBride: Well, so these were the issues that were coming up in the May time frame. Right after we sent out our letters to the CIO's, we had a lot of demand coming in saying, "We'd really like to see the code." We opened up our offices in Lindon, UT for code-viewing sessions. As I said, we've had dozens of folks that made the trek out to Lindon and we haven't had anybody yet that's walked away not a believer. So to the extent that you wanted to come and see it, you're more than welcome to. Anybody else that wants to come in. What we are doing to open it up more to the world, to more of an open session: We have our annual event that's coming up in Las Vegas on August 18, and we will in fact be doing a full show-and-tell session at that point in time.
DF: Let me ask you to comment on something that Linus Torvalds said. He said none of the SCO accusations have anything to do with IP rights. It's all about contracts between SCO and IBM and that it's blathering by SCO and it's a kind of holier-than-thou stuff.
McBride: The part with Linus that I would agree with him on is that it started out as a contracts case about IBM - and Linus and I have shared emails along the way and I guess we've chosen to agree to disagree. The point that we fully disagree on at this point is whether it is an intellectual property problem or not. Although it started off as a contracts problem, as we've gone in and done our full analysis of the code base, it is very clear that we have a copyright problem with the code that's in Linux today.
Now one thing that is late-breaking and fresh news for you and all the listeners out there is that last week, middle of the week, we did receive our registered copyrights back on the Unix code base. And the copyright registrations are interesting because they do in fact give you the ability to go seek injunctive and damages relief from end users.
CC: Well, clarify something for me. SCO has been a distributor under the GPL which provides for free distribution of the code. So how can SCO claim rightly that its IP has been violated if Linux customers are using the code that was distributed?
McBride: Right, so what happens here is there's a difference between a distribution of a code base and a donation of code. Because we were distributing Linux does not mean that we had donated code to Linux. In fact the GPL is very specific that it protects users or code developers who have had their code improperly donated into Linux. That is in fact the part that shuts down the GPL. If you have tainted code that's in there you have to in fact stop shipping, if we came out and put a claim against one of the Linux distributors. That's why we haven't done that at this point. If you look at the code that is out there. So we just found out about this a couple of months ago and when we did, immediately we came out and we suspended our shipment of Linux until this gets resolved.
So we're really protected on two fronts. The GPL actually protects us on this front and then if you look at copyright law, copyright law is just as explicit in terms of saying a copyright owner cannot accidentally give their rights away. You have to actually sign your rights away and we have never done that.
CC: If SCO prevails, what's the impact then on the future of Linux?
McBride: Well, I think the thing that is clear here is that we have code that is massively showing up inside of Linux 2.4 today. There's a couple million servers out there in the marketplace, and if you go back to early 2000, maybe 2001, and go forward, you have a lot of versions of Linux that are running that are tainted.
So I think the implications are pretty simple. We either get square with users that are using our code improperly and true up on that front, or, well, in any case we do that. But then I think in terms of the go forward is the real question.
The real question is, does the Linux community want to roll the codebase back?
I mean I actually agree with the points that Linus is making around "show us the code, we'll go back and fix this." The problem is we are talking about hundreds of thousands of code that have come from 25-year veterans of putting symmetrical multi-processing together.
When you take that code out, you know, to the extent that you can redo it with your own workings and have it there ready to go tomorrow, great. Most likely what happens is this rolls back to a pre-2000 state on the Linux codebase.
On the other hand if we want to leave the code in there and keep going and there's a way for SCO to get compensated for the damaged goods that are out there, then we're fine with that too.
DF: Did the symmetrical multi-processing actually come from SCO or its licenses, or did it come from IBM or some other organization?
McBride: When you look at the types of code that are in Linux today that are violating, there's really three types. There's line-by-line code that came right out of our System V source tree. There's derivative works code that came from vendors that we have license protections against them donating. And then there is basically non-literal implementations where they've munged code or obfuscated it to make it look like it's not.
The biggest concerning areas are the direct line-by-line and these derivative works code. The derivative works is the main area that IBM has been in violation of.
DF: Isn't it difficult to prove derivative works in the sense that Unix System V is derivative. There's things that predated it. There's BSD, I mean, is the Mac OS still going to be in your gun sights as well?
McBride: No, this is a little bit different than that. We're not aiming at BSD. We're not aiming at pre-2000 Linux. We're not trying to go after University code. We're basically saying in the late 2000's or pre-2000, you had two codebases that had the NUMA technology - non uniform memory architecture - that allows high end scalability inside of them. It was Unixware and it was Dynix that Sequent held. OK, so IBM buys Sequent while we're doing Project Monterey. Within the last 18 months, two years, they have in fact donated the Sequent NUMA code, the Sequent RCU code into Linux. OK, so part of it is code that we had as part of our Unixware product. Other parts such as RCU were things that were clearly derivatives. I mean if you look at the RCU code it could not be more clear that it's a derivative because in the copyright statements where IBM donated it into open source -- you can go out and look on sourceforge yourself -- it lists out that in fact RCU is a derivative work of Dynix, and Dynix is in fact a derivative work of System V.
CC: OK, one quick question.
CC: You said in the past that you intend to sue other hardware vendors. What's the status of that?
McBride: Well, I think what I've said in the past is I don't want to have any more lawsuits. If people press me on it and say if they don't come clean, does that mean you won't sue them? No. I mean, if we have to use the legal resources we have available to us, we will do that, but our preference would be to work things out with whether it's at vendor, user, or at a distributor level, or at an end-user level.
CC: I'll take that as a yes.
DF: I have one final question, Darl.
DF: Speaking about IBM, you were quoted as saying, "Those guys know what it's going to come out in discovery and you hear a lot of rumors on the street that they're going to buy us out. Well, I bet that's exactly what they want to do. The last thing they want to do is hear the testimony that's going to come out." Is your strategy really to get acquired by IBM, to basically save your company from extinction?
McBride: No, our strategy is not that. Our strategy is to take this business that we have and get it back on track and moving again. It was rolling along very nicely at the end of the millennium. IBM approached us and said, "Hey, let's do this thing called Project Monterey. Let's go to market together, file this thing with the Justice Department. Let's create 54% of the market share going into the new millennium." It was all set up and then after the project was all done from a technology standpoint, IBM backed off and didn't go to market with us.
And at the same time they did that, they jumped in bed with Red Hat and went off and started distributing Red Hat.
Now, just because they did that doesn't mean -- I mean at one level, yeah, we can be upset and we can whine and moan about it -- but that doesn't create technically a legal violation. What creates the violations are when they actually go out and take our code, contribute that into open source that in fact does boost Red Hat. It does boost the other distros. And at the same time our revenue comes down from 230 million down to 60 million. At the same time, the Linux marketplace is just booming.
DF: So didn't it really, didn't you really just miss the Linux market? Just kind of timing, that Unix being displaced by Linux in a lot of these scenarios? And that's a lot of the enterprise, not kind of in the smaller businesses where necessarily SCO has been doing a lot of business in the past.
McBride: I think you have to look at it real clearly. What is Linux? It is a Unix-like operating system on Intel. For 20 years, SCO had the value proposition of Unix on Intel. We all remember in the valley what it was like in the 90's, early 90's moving on. It was a boom time for the SCO group, the Santa Cruz Operation. It had the kind of brand Red Hat has today. When you go into the new millennium, what is the Red Hat value proposition? It is Unix on Intel.
CC: So you're saying if Linux hadn't come up, SCO Unix would have conquered the world?
McBride: A lot of the people that I talk to today talk about the proposition of Linux has more to do with cheap hardware than it does the operating system. I think absolutely SCO was positioned to be in the driver's seat for this burgeoning market of Unix on Intel.
DF: Well, the operating system wouldn't be in the enterprise if it were just because of cheap hardware.
McBride: I'm sorry, say that again?
DF: The operating system Linux wouldn't get into enterprises if it were just because of cheap hardware. It's got to be a robust operating system.
McBride: Understood. And the point is if you look where Unixware was in the 90's, it was way ahead of Linux by any stretch of the imagination. If you look at Unixware today, it still runs on more boxes from an SMP standpoint or from a NUMA standpoint than Linux does, but Linux is closing the gap fast. And from a certification standpoint, obviously Linux is ahead, because all the hardware vendors have jumped ship with us and gone over and now are supporting Linux. The value proposition is Unix for free, which is obviously a good deal if you can pull it off. But if you're going to do Unix for free, you've got to make sure it's also IP-free.
DF: Well, we'll have to leave it at that Darl. I want to thank you for joining us today. Darl McBride, who is CEO and president of SCO.
I'd also like to thank my colleague, who is Charlie Cooper from News.com.
I'm Dan Farber. This is Face to Face.