Stuart Cohen - interview

From LXF Wiki

INTERVIEW STUART COHEN "So Microsoft could join OSDL quite happily?" ABSOLUTELY Is the Open Source Development Labs the centre of gravity for Linux? Is open source licensing getting out of control? LXF met Stuart Cohen, the OSDL CEO, to find out...

Linux has to reach out to users as diverse as large multinational companies and home desktop noodlers ­ which means it has dozens of groups of people trying to pull it in dozens of directions. One of the ways we can ensure that no company has ultimate control over the Linux kernel is having its creator, Linus Torvalds, employed by a non-profit organisation jointly funded by the big vendors.

The Open Source Development Labs (OSDL) is that organisation: it employs Torvalds and several other big-name developers to ensure that no one vendor can hold undue sway over Linux development. It also acts as a centre point for discussion about standards, patent reform, licensing and other topics that all free software projects care about.

Paul Hudson spoke to Stuart Cohen, the CEO of OSDL since 2003, for his views on how well the OSDL is working with both vendors and the community, and how he'd like to see patent and licence reform proceed...

Linux Format: I think a lot of people might say that the OSDL favours large organisations. Would you say that's fair?

Stuart Cohen: We certainly have a lot of large vendors that are members of OSDL. And we have a lot of large user companies that participate in our customer advisory councils. We work with some large government agencies around the world, like the Ministry of Information Industry in China, the EU in Europe, the US government agencies in the United States.

So we work with big business if you will, but it's not just business ­ it's universities, it's government agencies as well. And we do have a number of startup companies and individuals that participate in the OSDL's activities, not only in the working groups but in the customer advisory councils ­ the LUACs as we call them, Linux User Advisory Councils ­ and we've had individuals contribute to the legal defence fund.

Although we're primarily focused on enterprise computing, which lends itself to larger businesses, and governments and universities, I think as Linux has success there it will deploy more into the small and medium business. That will be an important role for us as well.

LXF: But OSDL pitched itself as "the centre of gravity" for Linux.

SC: Centre of gravity is an interesting term. When I first got the job, actually a couple of weeks before I came on board, [Microsoft's] Steve Ballmer wrote a speech where he said we'll never be successful because there is no centre of gravity. We took that as a tongue in cheek phrase to kind of say, "We will be that centre of gravity."

Our mission is really to be the place where the vendors, developers and users can come together. It's really about accelerating the use of open source software in the enterprise market. `Centre of gravity' was really just kind of poking fun at a comment that Ballmer made poking fun at Linux.

LXF: I've spoken to one of the lead developers on the Linux kernel who said, "They're not the centre of gravity of Linux, they're the black hole of Linux." They had the feeling ­ at least among their colleagues ­ that although a lot of money comes in from a lot of big industry bodies, it disappears into the OSDL vortex and not much comes out. Would you say that's true?

SC: I think if you asked Linus or Andrew Morton or Andrew Tridgell or some of the kernel developers who we support or are employed by us, they are certainly producing a lot of good things that are coming out of OSDL from that standpoint.

We have a fairly small engineering group that works on those problems themselves, so when you think about the masses that are involved in the open source projects, and the masses that are involved in the kernel, a very, very small percentage of them work for us. But we always want to ensure we're doing our fair share ­ both on the leadership side and the code contribution side.

We have some subsystem maintainers and kernel developers that work for us and play key roles and they would certainly feel like it wasn't a black hole.

LXF: So you have Linus, Andrew, Tridgell, is it Chris Wright who still works for OSDL?

SC: Yes, and Steve Hemminger.

LXF: You have a limited number, maybe ten or so. So does the amount of work coming out of OSDL justify the money coming in?

SC: Well once again, let's talk about what we do from a business standpoint, what we do from a legal standpoint. Let's talk about what we do from a market standpoint, let's talk about what we do from an industry education standpoint as well as a technical standpoint.

LXF: Where does the biggest chunk of your money go to?

SC: Probably more of our money goes into engineering than anything else; about half of our budget.

LXF: What is your annual budget these days?

SC: It's about $10 million.

LXF: So $5 million a year is spent on engineering?

SC: Approximately. And it is made up of work that's done by kernel developers, it's done by work that's done in our testing group, it's work that we do around binary regression testing, and the OSDL Working Set ­ the open source project we have on our website that looks at the programs and libraries that people run on top of the kernel. It also goes to fund the working groups and the various groups around carrier-grade or desktop or data centre. It also goes toward the initiatives around networking, storage, clustering and security. It also goes to maintaining the equipment that the open source community uses to to run approximately 50­70 concurrent projects that are running at any one time in our data centre.

So there's a wide variety of things that we're doing, from an engineering standpoint. Then we have a small IT group that not only supports our activities internally, supports the projects with the two data centres ­the one in Tokyo and the one in Beaverton ­ but also contributes to a variety of open source projects.

LXF: But the large chunk, about half of it, is engineering?

SC: Correct.

LXF: How many engineers, if any, were laid off recently?

SC: There were a couple of engineers when we laid a few people off [in summer 2005] to basically make room from an expense standpoint for some areas we wanted to focus on ­ global expansion, IP activities. The spread was across marketing, business development, IT, finance ­

LXF: What do you mean by "IP activities"? That's a big term.

SC: For instance, there are various activities going on on the legal side today. There are issues around copyrights, obviously, with the SCO lawsuit, and a few years ago we announced a legal defence fund, in support of any engineers sued by SCO. We raised about $3 million within less than a week to help support those two end users, and that worked very well from the standpoint that SCO no longer sued any other end users and came out to say so publicly.

We've also done a lot of work around licensing, and the issues around licence proliferation. As you know, there's probably over 60 open licences out there today between FSF and OSI, and we think there's probably half a dozen licences that could do the majority of the work needed. So we're very involved in that.

We're involved in trademark issues, specifically around the Linux trademark and the Linux Mark Institute that Linus put in place years ago. I am on the board of the Linux Mark Institute along with Larry Augustin to look at putting together a sub-licence around the Linux trademark to make sure the trademark is solid, is upheld and becomes a valuable portion of what goes on and so the Linux mark is protected.

LXF: Didn't you launch an IP Support Fund?

SC: Yes, we raised money for two functions ­ one around some patent and IT issues that we're taking on internally, and one to help fund the Software Freedom Law Centre which Eben Moglen started; we made a contribution to that. You might have seen that we contributed $4.2 million over a few years to Eben for his work around providing legal services for open source projects that we thought were critically important. As these projects take on more and more software running in big business, big universities, big government agencies, we want to make sure they have the right kind of legal services in place. And we in turn want to be involved in that activity; plus one of the key things that he's going to take on is the revisioning of the GPL. So as he revisions [sic] from GPL 2.0 to 3.0 through the work of the Free Software Foundation, we want to make sure the Software Freedom Law Centre is in place to do that.

So that's just on the IP side alone. There's a significant focus on, "What are the inhibitors to the acceleration of Linux?" But while some of the developers may tell you it's all about code, there's actually a lot of things that are well beyond code, because right now Linux is technically good enough for mass deployment across the world. The issues are more business issues around things like open source software working together on top of Linux and the distributions, as it relates to things like standardisation across the distributions; making sure there's compatibility and consistency across the distributions, both in the US and globally.

LXF: Do you not think, though, that perhaps the issue of standardisation is best left to the LSB [Linux Standard Base], the issue of integration is best left to the vendors, and perhaps OSDL is best really dealing with the areas that it was really intended for?

SC: Not in the least. We've got a board of directors, we have a business plan, we have an agreed programme for what it is we're involved with. So we think the LSB is a great first step but needs to be a lot stronger and more powerful for it to work effectively across the distributions around the world. When you look at the activities with Asianux, with Mandriva and Debian, you look at the reference architecture coming out of China, and the Open Source Symposium projects that are coming out of China, Japan and Korea. We think it's very important that there is very strong compatibility between, you know, Red Hat and SUSE and these major distributions from around the world. That's what makes the value proposition strong and keeps customers' confidence so high. We do a lot more than just writing code.

LXF: So the engineers had to go to make room for the funds.

SC: Once again, there was just a couple of engineers, OK?

LXF: How many engineers do you hire in total?

SC: In total we have over 50 employees. We let a few people go that took us into the 40s, and as I said that was across all our functions.

LXF: Let's talk about patents. This is an issue where OSDL stands on both sides of the fence.

SC: How's that?

LXF: Well, a lot of your money comes from companies that hold quite a few patents. And quite a few of them claim, "We don't like patents very much, we'd love to abolish patents, but we keep registering them anyway." I mean, IBM is registering 3,000 a year or something crazy like that. You don't think that OSDL has some conflict of interest there perhaps?

SC: None. I think there is a reality in this whole conversation. When you look at patent offices around the world, whether it's China, Japan, the EU, the United States, they are all looking at patent reform.

We think patent reform is a good idea, we think there are way too many software patents out there. Most of them are invalid, most of them shouldn't have been granted in the first place. But the reality is, it will transition from where we are today to probably no software patents over a long period of time. And in that transition period, what we want to make sure of is that there is a number of different activities going on to make it viable for the developers to continue to work, continue to innovate and develop code without any issues regarding patents.

Whether you see the patent pledge that Red Hat has or whether it's the indemnification announcement that HP makes, or the patent pledge that IBM makes, we think that's very, very good. If you look at Sun, it pledged a number of patents but for Solaris users only. It would be nice if it was a broad, open source pledge if you will.

LXF: Nokia's patent pledge was for the kernel only. This is Open Source Development Labs, not Kernel Development Labs. Is that really enough for you?

SC: No, but once again it's a good first step. You can't expect overnight for these companies to take all of their patents and give them all away. But I think these companies are trying to work with the business models that are in place today. The governments are trying to reform the patent policies. It's not going to be done very quickly.

At the same time these companies are trying to reassure the developers around the world that their patents will not be used against them in any way, shape or form and that they have access to these patents as it relates to code development and innovation.

Now, is it complete elimination of all patents tomorrow? No. But I don't know anywhere in the world where that's really going to take place, other than in a few people's minds.

LXF: You did say you thought most of the patents that have been granted so far are invalid. You didn't say all of the patents. So am I right in thinking that your stance is that software patents should be allowed, just under closer scrutiny?

SC: You're saying two different things. I do believe that over time patents will reform to the extent that there will be no software patents.

LXF: And that's what you want?

SC: Yes. We think that's what's going to happen. But it's not going to happen now. So that's forward looking, we think patent reform is the right way to go. It is a long, long period of time before the governments collectively around the world get to that point.

Now, looking back, if you look at all the software patents that are out there, most of them were probably invalid to begin with. People talk about a third, two-thirds, 80% ­ you hear different numbers from different people speculating on the number of patents that are invalid. The point I was making is that if you look back a variety of people will tell you most of the patents are invalid and probably never should have been granted.

LXF: So just to confirm, OSDL's standpoint really is that software patents shouldn't be allowed.

SC: Our... Just to be completely clear: software patents are going away, and they're going away over time. I'm not going to stand here and tell you we should eliminate all software patents tomorrow, and then have the rest of the world tell me it's going to take 25 years, and then have you say, "You see OSDL was wrong."

LXF: I'm not after that kind of answer. I'm just saying, you believe at some point in the future ­ it doesn't have to be tomorrow, but maybe in 25, 30 years ­ you want to see patents eliminated.

SC: I think that's where the world's going, and we will support that.

LXF: Have you ever considered some sort of sanctions against members who completely ignore that stance?

SC: We're a non-profit, vendor-neutral organisation. So it's kind of a hard ­

LXF: But that doesn't mean you would let anybody in, does it?

SC: Well it actually does, by law.

LXF: The law says you have to allow anybody to join?

SC: Yes. You can't run a non-profit, vendor-neutral organisation and be selective over who you let in.

LXF: Surely you can?

SC: No, you can't.

LXF: If you ran the charity Save The Children, and a certain company joined that was also part of the We Hate Children organisation, you wouldn't have to let them in. You could say, "Listen, you completely disagree with what we're saying."

SC: The point of it is, there are laws that govern the way non-profits work. And obviously we wouldn't want to allow anyone to become a member of OSDL that wasn't in agreement with our mission, strategy and direction.

LXF: Right. So you can block membership of certain people?

SC: It's not a question of blocking, it's a question of, "These are our mission, goals and objectives ­ are you in line with and supporting of these activities?" And if they say yes, then we'll accept them as a member.

LXF: But if they're not?

SC: Then they won't join. But we can't stop them. We can't rewrite the way the laws work.

LXF: So Microsoft could join OSDL quite happily.

SC: Absolutely. We have another 70 companies who are OSDL members, probably growing to close to 100.

LXF: And how many of them do you think conflict with your patent point of view?

SC: Probably the vast majority of them are working on patent reform. I don't think there's any company that's not working on software patent reform to some extent. It's more about pace and degree, but in general I think every company that is a member of OSDL has said patents need to be reformed.

LXF: I think patent reform has a different meaning in the US. I believe there's a kind of patent reform going through right now where they're trying to basically lessen the dangers of litigation. Patents will still be as easy to get, but it's making them harder to sue over. And that's great if IBM or whoever that will be using that, but it doesn't really mean very much in terms of eradication of software patents in the long term.

SC: Well once again, I think it's a long process. Just take Brazil, China, Japan, the EU and the United States. Let's not cover every patent office in the world. If you just take some major patent offices, they are all at various positions on patent reform.

Is the United States government probably behind? Yes. Are they not moving as rapidly as some of the other patent offices are? That would be true. But I think it's all moving in the right general direction. LXF