Here is a transcript of The Linux Show's November 18 Town Hall meeting, which aired immediatley after Darl McBride's keynote speech at CDXPO. Here is how it came to be. A Groklaw reader, Trepalium, did a transcript and posted it. I was concerned we might need permission so I quickly wrote to Jeff Gerhardt, host and producer of The Linux Show, providing him with the transcript and asking him for permission to publish it on Groklaw.
He not only granted permission, he -- in the best open tradition -- built on our work. He corrected the transcript and provided his own commentary on the discussion. There remain a few places where the sound quality made it impossible to make out a word here or there, but otherwise it's the entire show. Enjoy. And Jeff, thanks.
The Open Source Town Hall Meeting
With Editorial Comments added by Jeff Gerhardt
My thanks to the folks at Groklaw both for transcribing the content of the Enterprise IT Week Town Hall Meeting and for giving me the opportunity to add comments to the transcript.
Please see the very end of the transcript for my comments on the emails I have gotten about the "fairness of the panel" and the way it was run.
Keep in mind that I had ONE real goal in participating on the panel and that was to repeat a single message over and over. The message in brief was that we (the Open Source community) had in our hands the definitive data as analytical proof to debunk all of SCO's IP violation claims that we have seen so far and were ready to present it to anyone who asked. I succeeded in that goal. I also wanted to ignore comments by SCO as much as possible, and thus avoid direct confrontation. This would allow me to maintain a perspective from the "high road." Could I have jumped all over some of Sontag's comments? Yes of course. In some cases I did not respond because of my desire to be considered "reasonable." In other cases, I was unable to respond because of the format of panel. There was one case where I wish I had been able to respond. But, my goals were met, and I feel the interests of the Open Source community were as well.
Host: We live in interesting times, along with interesting, interesting folks. I want to introduce our panel briefly. I want to ask them briefly to give us what they thought about it, what they think this is all about. And I've got a million questions, and I'm sure there's a million different answers we can give. Our panel today, starting from my left is David Schatsky, Senior Vice President of Research for Jupiter Research, Steven J. Henry, chair of the IP Transaction Group for the law firm Wolf, Greenfield & Sacks, Jeff Gerhardt, producer of The Linux Show, and Chris Sontag, Senior Vice President and General Manager of SCOsource. So, David, you start. What did you think about what you heard, and what should we be looking for and what should we be worrying about?
DS: [unintelligible] really solid points, that it's kind of chilling in a way because [unintelligible] really solid case for the intellectual property that the company possesses. I think that he, especially if you heard later, towards the end of the presentation, he tends to wrap this whole suit in the flag, as preserving American jobs, standing up for intellectual property rights, and that's obviously beside the point, it's really about the self interest of the company. But I think it all boils down to the sort of confused patrimony of the UNIX source code, and the tree metaphor is a valid one for explaining what the roots of the case really are.
Host: Steve, what did you think?
SJH: Well, I am an IT lawyer, and I believe the system works, or is supposed to work the way Darl explained. I believe it does protect value, does produce incentives to innovate and invest, but there's an old expression that I think comes to mind that "the devil's in the details". I was not persuaded by Darl's speech about SCO's position because I really don't know in that network of something like 140 agreements they have with [unintelligible] back to AT&T what their rights really are and how the codebase really breaks down at this point.
Host: [unintelligible] some number of millions of lines that allegedly are present?
SJH: Millions and millions.
Host: Yeah, interesting. Jeff, what did you think?
JG: Well certainly as part of a community that is advocating Open Source, we feel very strongly about our own intellectual property and the defense of that open source intellectual property. And, so really, at lot of the things Darl said today, on an intellectual basis, I'd have to agree with him, he has every right, the company SCO has every right in the world to defend their intellectual property. But eventually the time comes where they really do have to prove their case. Up to this point, the track record of the hyperbole has been very thin, and as of yet we have not seen, within the Linux community, one ounce of evidence that we've been able to duplicate that SCO has claimed.
CS: The first point I'd like to make is that SCO's issues are not about the concept of Open Source. I think you heard Darl when he spoke, mentioned that the ability for the people to share source code for which they have developed or the IP basis is well understood, is something that's reasonable, for which SCO has even participated in. Our issue is when a Open Source model's being used to effectively hijack some of SCO's intellectual property and apply it into a distribution model for which is very damaging to SCO and its business. We have shown source code, we have shown examples of our intellectual property violations, we've talked about the derivative works, over a million lines of code, that IBM and Sequent and others have contributed into Linux that has been identified, and we don't hear much about that. We've also shown some of the examples of literal copyright infringement, and those examples have been either ignored or quickly removed and said, "Oh, there's no more problem". The analogy I like to use in terms of, "oh, just show us the code, and we'll go ahead and remove it, and there'll be no more problem", is the example of a bank robber who robs a bank and now has the posse in hot pursuit behind him. Makes a swing back around past the bank and throws the money bags back into the bank and says, "Hey, no harm, no foul". There is a problem, there has been damages and [unintelligible] occurred, and we intend to get justice.
Host: I have a question. The issue is that in Linux there is lines and lines of the original UNIX code that is owned by SCO. How can Linux work, if it's not based on UNIX? How could you make a Linux that works, if you didn't take, maybe not code, but maybe certain intellectual property from that? What is Linux except a clone of UNIX?
JG: It's in no way a clone of UNIX. Not at all. In terms of historical background, certainly, the kernel owes a historical background to UNIX because the early UNIX developers really developed an art form, a standardized way of developing code, that is being used today, by the modern IT development community, not just Linux but across the industry, and so there is the existing art form. There are those standards that are out there. But is Linux a derivative copy of UNIX? No, that's nonsense. It's not at all.
Host: [unintelligible] finding a bunch of lines that are in the old system and in the new system. Maybe Steven, you can talk about this. If you make something that works just like something, did you just steal something?
SJH: There are two separate issues, and I think Darl correctly tried to separate them here. Firstly, when they filed the suit, there was no copyright claim involved. Just straight misappropriation of trade secrets and breach of contract case. You can read the complaint; it's on-line. When it comes to the copyright side, now we have questions of, "Was protected code copied?" Ordinarily, you don't prove copyright infringement, you don't get the opportunity to prove it, by having direct evidence of copying. You prove it by the fact that the accused party had access to the source code and that the result is substantially similar to the protected source code. It certainly sounds like, from the things we're hearing out of SCO, that it's substantially similar, if not identical, and their burden is going to be to go back and show that those similarities are not due to merely functionality, but due to the fact that somebody saw or had in mind, at the time, the protected copyrighted code when they wrote the Linux code.
Host: So you have to show that someone did it, that there was an opportunity for someone to do it, and the person actually did it, rather than it's just the same because that's the way you'd be to write code for it.
SJH: In fact, that is likely to be the same kind of proof that, or more of less the same kind of proof, they'd have to have on the contract side to show that they didn't just have somebody who independently, without looking at the code that was under wraps, write new code that was the same.
CS: Let me go ahead and make a comment on that because I think that Steve brings up an interesting point. You're speaking actually to the, more of the non-literal kind of versions of copyright infringement, where it's not the direct line-by-lines, which is the majority of copyright cases that actually go to trial end up being these very esoteric, "well, it was the methods, and the concepts and the structures and the sequence of this body of copyrighted work used." We have a substantial body of those non-literals which we have evidence and will present appropriately in court, but we also have literal copyright infringement where the line-by-line code was copied. And we know specifically which version was copied, and the licensee from SCO, who licensed that particular version, it all matches up, the fact points are all there. It's not some old, ancient version of UNIX, it was a particular, significantly newer version of UNIX that was copied. We can tell because the kind of the "DNA analysis" of how each version changes over time. And so, I just want to give that flavor, that there is a strong basis for the position that we've taken, and we look forward to laying that all out, but we're not going to lay out the whole case for the media, or... We have a respect for the law, and we're going to appropriately lay that all out.
JG: I'm a little bit surprised to see you're not going to lay it out for the media, because you seem to have been battling this case in the media quite a bit. There's also some subtleties about copyright and intellectual property. Let's take a case in point of one of your licensees, Sequent, as a case in point. Yes, indeed, if there was a derivative work developed by Sequent, and there are derivative works developed by Sequent, you do in fact own a copyright in the source code, but you do not own intellectual property to the trade secrets and methods by which they developed that software. If then, the people from Sequent come along and develop very similar code, it's their right to because they own those trade secrets and they own those methods. And so, there's going to be a lot of these situations for you that are going to be next to impossible to prove out. I've got to be honest with you, the Open Source Software Initiative sat down and they did a many-month-long, detailed analysis comparing the various Linux kernels to System V system, and other than basic noise from the standards of which software is developed and an awful lot of preexisting open source code, such as the BSD code that you and I have talked about before, there's not a whole heck of a lot that we have found in a complete multi-month-long analysis. I'd like to talk to the people -- you don't have to show me the code -- just let me talk to the people, who did your analysis so we can compare notes. Maybe we're wrong.
Host: This is why they have law and lawsuits.. Lawyers and lawsuits.
Host: David, you had a question.
DS: I had a question for Chris. In the cases you were just describing, "crystal clear examples of copying lines of code", what motive can you imagine there would be for legitimate companies to engage in those practices when the infringement would be so clear cut.
CS: You know, we don't know what the motivation was. It could've been accidental, a process by which several developers worked on something, they were testing something and didn't realize that code got into a production stream, but the copyright notices were removed and [unintelligible] we even have leaders in the Linux community that have basically agreed that yes, that code should not, that development process was flawed by that particular vendor, and that code should not have been included, and that all future versions of Linux will have that code pulled out. So in terms of showing the code, yeah, we've shown an example, and the Linux community is trying to pull it out, but that doesn't completely solve the problem, and we've many other examples that have to be rectified as well. Some of which, I think, are substantial. Derivative works, we can argue about - when a derivative work is a completely substantially different body of work. We also have a huge amount of case law regarding what is derivative works in lots of different areas and trades. Copyrighted works, the derivative of those copyrighted works, when they're using the methods and concepts of that original work. There is afforded protection by copyright.
Host: Chris, I want to raise our eyes up here a little bit. What we've been talking a lot about here is this specific lawsuit, this specific case, and you want your money that was lost because this happened, and the other guy says it didn't happen, and it all goes to court, and some lawyer and some, god help us, judge decides which is right in this case. And at the end of the day, either you get paid some settlement money, or you charge everyone for the rights, and so on. Okay, fine. I never watch CourtTV because listening to the discussions on how the trial is gonna go drive me nuts. But, here I'm in business, I'm an enterprise IT guy, I'm trying to figure out what the heck to do in this area. Of course, I can say, "Oh, my God! Linux, too complicated" or "Too risky" and I have to go someplace else. Short of that, what do you, when you come a-callin' someday, assuming you win your case, what are you gonna want from me? What's it gonna cost me to use the Linux I've already got sitting inside?
CS: Simply, we want to come up with a fair and reasonable licensing policy for the appropriate IP that is SCO IP that's in any product. And we've done lots of licensing over the years. Lots of people have licensed our intellectual property, including IBM, including HP, including Sun and many others. And we just want to get reasonable licensing for that [unintelligible].
Host: And what is your regular retail price for the regular retail products? Today's cost? What does it cost to get . . .
CS: Anywhere from $800 to several thousand dollars for the different versions of our operating system.
Host: So, I hate to say it this way, but, worst case, I'm gonna have to pay you some money.
CS: Certainly, sure.
Host: Best case, it all goes to hell. Or my HP, for example, indemnifies me against having to pay this extra money. I mean, when we get to a court case what we're really talking about is money at some place. You're saying, "Alright, it'll cost me this much to run my multi-billion dollar data center to make this problem go away when it becomes a problem for me."
CS: And if I'm an IT shop, and I want to have a solid basis for the technologies I'm running in my organization, I think ultimately, companies want to pay for the IT, you know, for the IP that they're using in their company. Every company that I've talked to when I've sat down with them have said "We want to pay for, you know, we want to make sure we're do things correctly." I don't know of any Chief Technology Officer, CIO, that does want to be right with the law.
Host: Okay, so that's part one of this discussion. That whatever happens is just money. Turn on the mike for Amy.
Amy Wohl: So, first of all, several folks have gone out, surveyed the large IT shops about whether they're concerned about this issue, and 85% of them have come back and said it isn't an issue. That is, they have not stopped installing Linux and they have not signed up for licenses. So I would suggest to you they're a lot less concerned about this than you might imagine. Second of all, as I understand this, having spoken to several of the lawyers directly involved in the case. If you have a license already from a company that you bought Linux software from, you have in fact bought a license and SCO can try to say anything they like, but, they're going to have a lot of fun trying to get you to, litigate and be forced to sign another license for something you already own a license to. The GPL is, in fact, a license.
Host: It's interesting...
AW: It is an intellectual property protection, in spite what Darl is saying.
Host: Okay, I want to take it to the next step with this. It's an interesting figure -- 85% say it's not a problem because they've got money and lawyers. And if we have, if we draw the rough parallel between SCO coming after people for infringing its intellectual property, and the RIAA music people coming after people, when they sue 12-year-old Brianna LaHara and take $2,000 from her, that can make a lot of people in that area, in that world, scared. When a company, a 60 million dollar company, comes and tries to sue General Motors because of a license, it becomes just a lawsuit and just becomes a question of dollars and cents. That's one level that is a reasonably understandable problem. But I want to knock it up a notch, because I want to say, okay, suppose I'm running Linux, suppose I wait to see whatever's gonna happen with this thing. But suppose I'm building a product on top of Linux, I'm building software on top of Linux, I'm building Linux-based products. What do I, in the way things work today, am I creating new problems for myself, after this first problem, by building something on top of an open-source product, or is what's going to happen, is my new product I create from here on in, is that gonna get stuck with this same problem or not?
JG: It depends on the tools you use to develop the technology, and the license you use to license the product. There's plenty of Open Source GPL products that are also available with commercial licenses. Apache is one. MySQL, a tremendously successful company, that has the same software available with two different licenses. So if you're a software development company, sure you can develop commercial software based on the Linux operating system. You just need to be aware of the tools you're using and not use GPL tools, and you have to look at the way you're manipulating the operating system. And if you do modify the operating system to run the software, then yeah, you do have certain licensing issues you have to address.
DS: But you're asking if you build an application on top of an unmodified operating system, are you enhancing your liability?
Host: I'm asking where does my liability start in this next step.
DS: The problem is the suit has injected a massive and unwelcome dose of fear, uncertainty and doubt in an already embattled IT environment, and to the extent you're using UNIX of any flavor, you now have a giant headache. And to an extent you're building applications that run on it, those, while they may not infringe, they're increasing your dependence on this shaky foundation that's just been shaken up.
Host: And so what should I do to protect myself?
SJH: You have a basic problem here, and the question a couple moments ago highlights it. And it underlays everything that Darl said, and that is, and I'll spare you the Latin, but the basic premise of the law is, you can't give what you don't own. I mean, the license isn't worth the paper it's written on, if the licensor doesn't have the right to give you. It's just like going out and buying a house. You better check the title. And the problem in the open source world is that developers have been extremely lax in worrying about title. They don't, in the simplest case, if you have a small company and it's got some stringers that do a little bit of code here and there. I mean, they don't even have written agreements to assign the rights into the company, so they give an open source license GPL or otherwise to what? To what they have, but that consultant or stringer can come along at any time, and say, "What a second, you, Mister Licensee, are now an infringer."
Host: So, we need a much more explicit understanding of who owns what.
CS: And that's probably a limitation that exists with the... I mean, that's a very big problem with the development process in Linux right now, is that there isn't good checks and balances there.
JG: I would disagree with that. I would completely disagree with that. There is a tremendous amount of peer review that goes on within the open source marketplace that is extremely detailed, and I would venture a guess that Microsoft and SCO, not the present SCO but the historical SCO, have been guilty of far more code confiscation than the open source community.
Host: But you can't prove it because nobody tracks it.
JG: Oh, we certainly can prove it because they both, AT&T and Microsoft, infringed on the BSD license.
Host: But I think that Steven pointed out...
CS: What's your point about?
JG: My point is that there is peer review, there is checks and balances. Do checks and balances need to be improved? They always do. Review always needs to be done as well as it possibly can because everybody needs to protect everybody else's IP.
CS: Well, unfortunately, I think it's kind of the fox guarding the hen house because many of the core people that are involved in that peer review of the IP going into Linux have a very high degree of disdain for IP in general. There's emails that have gone back and forth between many of the principles involved with Linux talking about their disdain for patents and the, that whole process and that they ignore it. I mean, that's a documented fact. That's out there. So, we may have peer review, but it may not be a very effective peer review, because like I said, it's like the fox guarding the hen house.
Host: I mean, the more important point is that you have to prove the provenance of software you're hoping to protect. You are needing to show an audit trail, the right signoff up and down the line that what I'm giving you, I own to give you.
SJH: And, in fact, my understanding is that a lot of the contributors to Linux have UNIX employment in their background with companies that were under these various restrictive licenses, and where they gained access to the source code, and even if they didn't have it on paper in front of them when they were sitting down to do their Linux coding, and there may be infringements there too, but that doesn't mean that Linux doesn't infringe.
Host: We have a question back here.
Male from Audience: Going back historically... Let's roll the camera back, right, even before SCO got a hold of System V, the issue around System V code being in Open Source is historical, and what I point back to is BSD, for instance, was basically in violation, and then it came out with BSD Lite. 4.4BSD came out of the University and was basically stripped of all the System V things, that was settled and that was done. And I'd like to hear. . . CS: I'd actually give you one correction. Not all the System V code was stripped out. There a certain amount of code that was joint copyright attributions had to be applied for Berkeley and USL and AT&T. So there was a certain set of files for which the over, there was this splitting, of Solomon splitting the baby, that some assets that had to be held in common. But you're right there is a certain settlement that was made for correcting a BSD-based solution.
Male in Audience: I'm seeing that correction seemed to go a lot more smoothly than this one. And I'd like your opinion on what we can possibly learn from that historical event and what separates that from this.
Host: Well, there wasn't, was there a bunch of stockholders hoping to cash in? Listening to Darl's story from the beginning, it seems like, it's a very interesting and compelling story. A guy shows up, there's a lot of money that he sees laying on the ground, and goes and tries to do it. I mean God bless America. For us on the other side of this, it's a question of where's our, or what's our exposure and what do we have to do? And truly, and what should we do from now on, all the time, when we make software so this kind of thing never happens to us in a small, in a microcosm case?
DS: I think this casts, as I said before, creates a lot of doubt, and it probably undermines certainty that any given technology purchase, that might have some kind of open-source influence, can be ever completely safe. I think it creates the impression in the industry that somebody is trying to patent the carbon atom, something that is ubiquitous and permeates of everything. Someone went back and found out they happen to own it, and now we're all liable.
Host: Interesting, interesting. Any other questions? Amy?
AW: I think it's interesting to note that while some of the members of this panel seem to think that Linux developers or customers are avoiding making open-source decisions because of the SCO activities. In fact that is not borne out by what's going on in the market place. The Linux market share has increased during this period, to the detriment of UNIX, not to the detriment of Microsoft operating systems, which is just the continuation of a very long-range trend and that is in spite of the fact we're paying more for Linux as time goes on. It's actually correct to note that it is not a free operating system. Most people choose to buy a paid version and to buy maintenance and support at the same time, rather than than simply getting the software for free. And also there are more and more developers moving their applications to the Linux platform, or developing first port, the first version for Linux, and then porting off of Linux, instead of porting to Linux after they've finished the application development. There's absolutely no indication at all as far as I can tell that that's going on.
SJH: The same could be said for illegal music file sharing online. Until lawsuits started dropping like rain, it was undiminished, even when they were threatening them. But now that hundreds of people have been sued, I think there is some evidence, that it is in decline. But, on the other hand, that's also accompanied by reasonable legal alternatives to using the music, and that's what we need to see here, too.
Host: The lawsuit in question here, is the lawsuit against IBM, and kind of famously, they didn't pay you to go away or buy you, as everybody says. Is IBM the wrong target for this, or -- they've got 7000 Linux developers, I've read -- I mean, in terms of people in the world, did you start with IBM because they have the deepest pockets, or they were most likely to settle? I mean, why IBM first?
CS: We filed suit against IBM because we had a significant problem in terms of the major contract violation that was very damaging to SCO. And it was very obvious and apparent to us. A lot of the other issues, in terms of IT issues we've identified, took a lot longer for us to be able research and uncover. It's not . . . in a way you're talking about 7 million lines of code or more, and going through all that is like looking for needles in a haystack in some cases.
Host: How do you prove this in a court of law, I mean no offense Steve, but when I talk to lawyers it doesn't seem like they have these bigger pictures involved, and you guys have picked David Boies, who is one of the top lawyers in the world. But, how do you go about convincing (even though he lost his case with Al Gore), how do you go about convincing and proving to people who are not computer people, people who are not coders that this [??] exists. I wonder if this is difficult or. . .
SJH: Two words: expert witnesses. And if they are believable, then the judge or jury are going to accept it from one or the other.
Host: So you pay money to an expert to come in and say, "I say this." And the other guys, the other.
SJH: We have lay judges and lay juries, and you hit them with sufficiently complex cases, and if you want to be pragmatic and cynical about it, they'll do their best to understand it, but at some point they start to have psychological reactions because they're human beings, and they believe that somebody is feeding them a cleaner story than somebody else, and they end up going that way. Some of what we're hearing is that it's going to be really a black-and-white situation, you put the code up side-by-side. You don't have to be technical to know it's the same. In the old days, in the copyright world, there were things like map maker's tricks. Map makers would salt their maps with false towns, railroad crossings that didn't exist, things of that sort because it was an open-and-shut case when you found the copy. There was no way for those things to be there as artifacts, unless there had been copying, and SCO presumably will be able to put the code side by side, and say "there's just no reason that everything is going to line up this way if someone did it innocently -- couldn't happen." JG: Again, I'd have to disagree with you. Although I agree from the standpoint that expert witnesses is going to be what decides this, putting the code side-by-side is helpful, but is it ironclad? No. Without the genealogy of that code along with it, it really doesn't tell the judges a whole heck of a lot. And a lot of this code that is absolutely identical has its heritage from BSD. We have seen that when we have done side-by-side code analysis, and I wish we had evidence to the difference, but...
CS: That's the same burden you have, even in a music copyright, you have to separate protectable from the unprotectable. And I'll agree with Jeff that genealogy also has to be factored in, but I'm not sure where he's looking at his genealogy because some of what he's attributing to BSD is not appropriate, and even some of BSD has ended up in Linux without appropriate attribution, and all of that is a problem.
Host: Interesting. Chris, what's the date of this lawsuit? When is this all going to come to a head, or start to come to a head?
CS: April 2005
Host: April 2005
CS: 17 months out.
Host: Maybe I'll actually watch CourtTV and start worrying about it then. Thank you very much, and please give me a big round of applause for our panel. Go forth and be open.
The following section is 100% the opinion of Jeff Gerhardt.
First let's do the math.......
So, was it Biased?
In my opinion I can only find fault on a couple of trivial issues. Sontag was given more time and more frequent questions and responses than others on the panel. But I was in no way prevented from delivering my message or representing my agenda. It would have been nice for the host to have taken the previous 45 minutes being dedicated to Darl's Keynote, as a factor in his allotment of time. But it is really unfair to find fault in this. The panel needed to stand on its own.
In general I feel that Jack did the best job one could expect under the conditions. Lets look to the positive, that someone from the community got to sit on a stage next to Chris Sontag and three times call his bluff on the source code issue.
Although SCO had already begun to hint at possible BSD inquiries, I find it interesting that the following day SCO made public a significant change in strategy.
The comment by the audience member was one of the most inciteful comments of the evening, in my opinion. Why did SCO have no trouble resolving the issue when BSD code got included in UNIX? The "bank robber" drove back to the "bank" and seemed to have no trouble getting the money back to the bank in that case.