Bruce Perens - interview

From LXF Wiki

Bruce forthright

Bruce Perens isn't as big as he looks in this photo, but he certainly casts a long shadow. We spoke to the former Debian Project Leader about freedom, Ubuntu and *ahem* Red Hat...

Some people sit on the free software side of the fence, and others sit on the open source side. But only a few people ­ people such as Bruce Perens ­ have been on both sides of the fence, have learnt that we could actually get along much better if the fence didn't exist at all, and work to bring the two philosophies together. You might have come across Perens in any number of different places: he led the Debian project for a while, he founded the Open Source Initiative and he even has a whole series of technology books released under his name.

We met up with Bruce in San Francisco, and asked him about his Linux distro of choice, the embarrassment of licences and the future of UserLinux...

Linux Format: Just to introduce you to our readers, can you give us a very quick autobiography?

Bruce Perens: I'm probably best known as the creator of The Open Source Definition. This is the manifesto of open source, and sets the rules for open source licensing. It's still followed today. In this I'm standing on the shoulders of giants, Richard Stallman the largest among them. Richard did a tremendous amount of the work toward what we have today and will never be appreciated for that.

And I have a book series ­ there are 17 books now, I do it with Prentice Hall. All of the books are under open source licences; indeed, the day you buy them you can stick them in a copier and sell the copies. No one does; however, I make unencrypted PDFs available for download. Some people use those without buying the book but I don't feel we're losing significant sales. People just like paper.

I teach at two schools, I am a professor II candidate at Agder University College in Kristiansand, Norway, and I am a senior research scientist for George Washington University in Washington DC. I get to GWU about once a year and I get to Norway much more often.

LXF: Where are you based?

BP: Berkeley, California. And I'm vice president for Sourcelabs, which is a venture capital-funded startup that's supporting open source software in the enterprise. Their most important product right now is a Java stack containing most of the open source components that large businesses latch on to; things like Hibernate etc.

LXF: You also spent some time as DPL [Debian Project Leader] and wrote the Free Software Guidelines for Debian.

BPs: Well, actually the Debian Free Software Guidelines became the Open Source Definition. But Debian I came on to when there were about 60 people, and Ian [Murdock] was a college student and was running it. I felt that Debian was important. My main motivation at the time was that I wanted to make a Linux system for ham radio operators, combining the two most nerdy pursuits ­ someone else has done that, actually, there's a version of Knoppix for radio hams [amateur radio operators] now.

But what happened is that I got busier with Debian than amateur radio and became the project leader. I'm still involved in it today, so I think I've been with the project for almost all of its existence, and I'm on the board of its corporation today.

LXF: It seems that Debian has forked an awful lot ­ others have taken Debian as a base.

BP: Debian GNU/Linux is meant to have derivatives.

LXF: And yet you seem to have stuck very strongly with Debian, you haven't moved on to Red Hat or Morphix or Knoppix...

BP: You know, I got pissed off at the Debian folks once and tried to run Red Hat on my systems, and I came back to Debian pretty quickly.

LXF: Have you tried it recently?

BP: I have a laptop with Debian on it. Have I tried Red Hat recently? Well, once in a while I have to sit down at a Red Hat system in some work capacity, and I don't think there's anything wrong with Red Hat, it just can't compare with Debian's breadth of applications or the number of developers that Debian has.

LXF: Since you were really active in Debian ­ you said you're still part of the project, but I don't think you're quite near your previous activity levels ­

BP: I'm not the benign dictator any more...

LXF: How have things changed?

BP: Well, I mean, it's huge. I was well regarded for having built Debian from 60 to 200 people, and it is over 1,000 people now. Quite literally there are few landmasses without a Debian developer, and I think that it's been and will remain a very important part of the Linux scene. I think it will be commercially more important as time goes on we've seen some unsuccessful attempts to commercialise it, including some that I have started. I hired Ian to run Progeny, which was originally going to be a Debian company and that didn't work out, the market wasn't ready. User Linux may still be successful but the Debian Common Core Consortium actually embodies a lot of the User Linux principles. So whether the User Linux project comes out or not, I think a lot of what I've thought of for it is going on.

LXF: The Debian project seems to have really slowed down; the Sarge thing...

BP: Well, Debian took an awful long time to make the Sarge release and one of the main reasons for that is that they changed the installation system. That was a really, really big job, and they actually held back from doing it during the previous release, and decided "now we have to bite the bullet; there's not really any going on without fixing this problem".

Debian is now focusing on 18-month releases and I would have actually liked to get it down to a year, but maybe that's asking too much. I think everyone in the Debian project is aware that if they don't make a release in 18 months no one is going to take them seriously.

LXF: When you thought of the Open Source Initiative as a concept, did you envisage so many open source licences?

BP: No. Had I envisioned there would be this many open source licences I would have written the rules differently.

However, it's almost never a good idea to plan on an embarrassment of riches, and I think that when we started we actually wanted to have a licence from a large number of different large companies, because we could say "there's an IBM licence and a Sun licence", and that was taken as a degree of endorsement by the companies connected with that. Now, from a standpoint of making things workable for the community that was a terrible, terrible idea and I understand that Open Source Initiative is working on the issue of licence proliferation. They have a committee about it, but actually I was rejected for that committee ­ I applied to be on it, and got almost a form letter that said "sorry, you weren't chosen".

LXF: So perhaps there are still some hard feelings somewhere.

BP: Well, I think there are a couple of people on the board of OSI who...maybe haven't grown as much as they should have.

LXF: Let's talk about UserLinux. You announced it, what? A year, two years ago?

BP: We've been working on this for almost two years. The website still stands today but the project is pretty much a ghost town. I believe it can be revived; I believe that now I am working with Sourcelabs I can get some other things off my plate and spend more time on UserLinux again.

The problem for me with UserLinux was that when I started it I was an independent entrepreneur; I had a one-person consulting company. So I had to go and make enough sales so that I made a living every month, I had to be an open source leader, and I had to run a number of projects. There simply wasn't room for all of that, and at some point... no one was paying me to work on UserLinux, and supporting my family came first.

Now I have an enterprise that's paying me to work on whatever I want for open source with half of my time, so I am able to do this sort of work more. There are some things that I am getting off my plate: I just did a security model for Ruby On Rails, and with some of the core Ruby On Rails developers I'm working on making a release that normal people can run right now you've got to be a guru. Then I have some book work that I have to get out of the way, and those are really the two main projects that I would have to clear up so that I could spend more time on UserLinux.

LXF: Since UserLinux started up, Ubuntu came along and it seems to play the same sort of role. I think Jeff Waugh actually made a post saying "Hey, UserLinux guys, you can work with us on Ubuntu."

BP: I actually considered going to work with Canonical when Mark [Shuttleworth] was starting it, and there were a couple of problems with that. I think that Mark is eventually interested in having a successful and profitable company, and I don't believe that Linux distributions are a natural fit for for-profit enterprise. Indeed, if you go on my website I have a very long paper on the economics of open source [1] (http://perens.com/Articles/Economic.html), and one of the things that you can derive from that is the fact that open source works almost worst for a for-profit Linux distribution.

And thus I feel that over the long term more of the work of maintaining Linux distributions will go back into the customer-driven space, where customers who are interested in adding features put their own people on it, pay for them to work on it etc. Debian has always followed that model and we have seen sort of half-hearted efforts to put what is good about Debian into the other distributions; Fedora and Novell are going to do something like that. The problem with those efforts is control. You know, Michael Tiemann has been trying to make Fedora independent but it will never be to the degree that Debian is.

As far as Canonical is concerned, one thing that struck me about Mark is that he really insists on control. For example, when I considered being an employee one of the things standing in the way was the fact that Mark doesn't give his employees stock in his companies. If I'm going to work for someone I'm going to be a little entrepreneurial about it, so I felt that although Ubuntu and Canonical could do a great deal for Debian and be excellent community members, they were never going to be the core, and we could actually get closer to the core by following what I have set up for UserLinux.

LXF: How do think what you have said applies to Red Hat and Novell in the long term? Red Hat depends basically on making a very good distro and getting support maintenance out of it.

BP: [sighs] I see a lot of customers, because Sourcelabs has customers. [This is] not an official statement of Sourcelabs, because Sourcelabs has no problem with Red Hat: every customer I talk to hates Red Hat. I think it's the service mostly.

LXF: Do you think their business model is somewhat endangered?

BP: I think that a for-profit enterprise is not the right way to create a Linux distribution. It is potentially the right way to service one, but only if there is a service market rather than a service monopoly.

LXF: Let's talk about Sourcelabs. You've already mentioned what the company is about, you said they pay half your time to do what you want...

BP: They let me send my son to private kindergarten, which is one of the major reasons why I took a job, and work on what I love 100% of the time, and half of that time is on my own agenda, so if I want to hack on Ruby On Rails or go to Europe and protest on software patents these are equally acceptable.

The other half of the time I've defined pretty much for myself at the moment, which is that Sourcelabs is a start-up company and I've been helping them associate with Fortune 100 companies. We're not announcing the customer list but I can say that there is a world-class city, there are a number of very, very large companies that everyone knows their name.

LXF: The thing is, to me, Sourcelabs seems another of these very new companies coming along seeking to recreate the dotcom boom. Lots of base capital, lots of grand ideas, open source mixed in there somewhere, and I still don't know what they do.

BP: There is a lot of confusion, and there are a lot of companies that tend to say similar things. And you can tell Sourcelabs apart from them ­ and this was very important to me ­ because Sourcelabs does not have a lock-in. When you see that Sourcelabs has released a new software stack, that software stack is available on their website. Right now they want your name and address because they want to send you junk mail, but that's not asking you very much for a download, and of course since it's all open source-licensed you can go mirror that site and not ask anyone for their name and address.

They have put into the versions of open source that they release a great deal of testing, where open source right now has a culture of unit testing that we've taken on over the past several years, but we've not done scalability testing, whole systems testing etc ­ things that you need to say, "You can base your business on this and it's not going to leave you in the lurch."

So Sourcelabs has essentially created a support market, and attempted to be the first citizen in that support market. There is nothing that says another company cannot support the same software. We think that's very much to our customers' advantage and we hope that our customers will select us because of that; also because we've got a good many experts.

There are only three people who aren't engineers in the whole company at this point, and when you call for support to Sourcelabs you do not get a script reader in India, you get a software developer who is working on the software that you are having a problem with and may well be one of the core committers. As the company gets larger we expect to able to cover more projects that way than the stacks that we have currently announced, and we are looking towards the future. For example, I have developed an expertise on Ruby On Rails, and we think that's a very promising platform. We don't officially support it yet, and there is other work in that area going on where people are developing expertise that the company may eventually be able to market.

LXF: So with others like SpikeSource aimed at productising open source...

BP: We're not entirely sure what SpikeSource's business plan is yet; we're not sure they're sure either. I like Kim [Polese], she's a nice person. They're well capitalised and they've hired a great many people, and in the open source world you don't diss a competitor who's doing useful things for open source ­ that's not the way we work. So I hope that SpikeSource and Sourcelabs can be as HP and IBM are: both doing good things. I can't say very much about what they do yet, and I think we'll be more clear on that in a little while.

LXF: What is your view on industry bodies that help open source ­ OSDL, for example?

BP: Well, some of the industry bodies try to help open source. OSDL is actually handicapped in one very important way, which is that the majority of OSDL's membership have a conflict of interest where the agenda of open source is concerned.

Most of those folks are in the pro-software patenting camp. Software patents are the number one weapon that can be used to hurt open source, and we must fight against them as a community. Unfortunately OSDL can't do very much about that. I've discussed that with Stu [Cohen, OSDL's CEO] and I think he understands my point perfectly.

I'm very happy that OSDL has put some money into Eben Moglen's enterprise, which is sort of a legal aid for open source where you can get competent legal counselling, which is not available on Slashdot! I think that it's admirable when companies stick their neck out a bit regarding politics. So we see for example MySQL funding NoSoftwarePatents.org, which has been an extremely successful campaign in Europe and has put off the mfight for a number of years it's not as if we've won it permanently. It's nice that Sourcelabs is sponsoring me to say things about that without governing what I say.

LXF: Do you expect to be able to fight for US patent reform?

BP: I think that that might be possible now; that we are getting to the point where most people in the industry acknowledge that the system is broken. We had a patent bill this year [2005], which doesn't really do much to solve the problem ­ it is mostly on behalf of the large patent-holders.

LXF: A while ago people were saying that figureheads like yourself or Richard Stallman are no longer needed because open source promotes itself. How do you feel about that?

BP: I always look at myself as a person who puts a voice to the concerns of the community rather than as a leader, and that's actually borne out by the email I get. When I write an article or talk to the press I very often get emails that say, "thank you for putting this into words. I felt very strongly about it and I didn't really know how to explain it to the people around me. Now I can point them to your article." It's funny, I was a communications arts major, not an engineering major, so maybe I'm better at that piece of the equation.

I think that people like myself and Richard Stallman are much more necessary now, because if we were not there the vacuum would be filled by various corporate spokespeople who would be speaking for their companies' agenda rather than that of open source. I assure you that we are not the same. LXF