In the beginning was Slackware, and it was good. But that was just the beginning. Ever since then, Linux distros have exploded at dazzling pace. And at the center of much of that creative explosion is Debian. It is simply everywhere. Mepis, Knoppix, Linspire, geez you name it, and if it's not SUSE Novell or Red Hat, there's a good chance that the spirit and code of Debian lives at the heart of that distro. Desktops, enterprise, hackerdom, you name it, you'll probably find numerous Debian fans there.
On Tuesday, August 9, 2005, many of the major Debian-derivative GNU/Linux distros banded together to create the Debian Common Core Alliance. Essentially, the DCCA is a group of Debian heavy-weights who got tired of reinventing each others' efforts, and decided to implement the Linux Standards Base in a set of binary packages that will be
common to their distros. The benefits are obvious. Applications designed to work on one distro will work on another. The Alliance members will save bunches of money on not duplicating each others' work. Customers of the DCCA members will have the option of getting support for DCC code from a broader array of vendors. And perhaps most importantly, it now becomes just that much harder for Steve Ballmer to whine about an absence of a "center of gravity" in GNU/Linux distros.
As former Debian lead Bruce Perens discusses in this article, Debian has long played a central role in helping people all over the world share the fun, frivolity, fascination, and occasional frustration that is the world of free Open Source software. Bruce has been present and intimately involved with much of this work as an author, advocate, and teacher of much of what it means for a project to be an "Open Source" project, so he is in a great position to give us penguinistas some insight as to what the new DCC initiative means for our favorite operating system, and the free Open Source applications that ride on it.
Mad Penguin: Most of our readers probably know who you are, but we have been fortunate enough to get one or two new readers lately, so it is a possibility that a few of our readers actually might not know who you
are.
Bruce Perens: Probably the first thing that people would care about is my work in the film industry. I worked for Pixar for more than 12 years. Actually I worked for Pixar's predecessor, which no won heard about, for an additional 6 years and some months, and so I have spent 19 years total in the film industry. I left the film industry in 1999 to work full time on Linux and Open Source. I spent some time on my own and then became president of Linux Capital Group, a company which creates Linux companies. One of the companies that we helped create (along with its founders), Progeny, is here and Progeny just led the inaugural meeting of the Debian Common Core Alliance.
In 1998 or so, I helped create the term, "Open Source." Before that, we called it free software. About six months before that announcement, I helped design the Debian Social Contract which they still follow today. Part of that was the Debian
Free Software Guidelines. Those became the Open Source Definition, and are today both the Manifesto and the definition of what is acceptable Open Source licensing.
I was a project lead of the Debian software project. I helped build that project from 50 to 200 developers. It is now over 1,000 developers, and is the second largest Linux distribution over all. It's about three times the size of SUSE Novell in terms of users. I ran a project called "No Code International", and we have successfully eliminated Morse code testing as a prerequisite for any class of amateur radio licenses, which is another way for individuals to participate in technology development. That's a pretty common theme in the stuff that I do.
I'm also very active in the politics for Open Source. For example, I've been working on patenting for software, which is a grave threat to Open Source, and I helped push that off in Europe, where we finally had a 418 to 18 vote against their software patenting bill.
MP: So tell us what the Debian Common Core Alliance is, and what gap it fills.
BP: I've been hoping for a long time that we would get something like this done. Years ago, I helped found the Linux Standard Base, which is a very health paper standard for Linux, but I personally didn't want it to be just a paper standard. I proposed at that time for all of the Linux distributions to get together and collaborate on standard set of binary packages that would be the base of all Linux distributions, Red Hat, SUSE Novell. There have been a number of efforts since then, the Linux Core Consortium, which unfortunately didn't work out, and some members eventually left, leaving only the members who are interested in Debian.
So what we're doing now is that we have gotten really all but one now of the leading Debian derivative distributions to gather to collaborate on having a common core which we will certify to the Linux Standard Base 3.0. Debian is very close to that standard now. We will have that core be distributed to application creators, so that they can make sure that their application works on the Debian Common Core, and then they will be able to say without doing any additional work that their application works on Linspire, Xandros, SUSE, Mepis, and a number of other derivative distributions that are named in our press kit. Currently, Ubuntu is not a member, but I believe that if you certify to this DCC platform, it should work on Ubuntu also.
MP: What does today's announcement mean in terms of Debian's install base and in terms of moving closer to the end user experience?
BP: Well, I'm working for Source Labs now, and Source Labs is in the Open Source business, but not the Linux business. But every time I visit a customer, and I visit a good many Fortune 100 companies, there's one thing in common. They all complain the service from a certain very large Linux distribution. One of the things that I'd like to give those folks is a free, standardized Linux system for the enterprise. I think that the DCC will let us do that, and then if you want a greater degree of support, you can go to Xandros or Linspire or another independent service vendor and get that support. I think it's very important for Linux and Open Source to go from the paradigm of a distribution with a single support vendor, which is really what we have now in the case of Novell and Red Hat, each having divergent distributions, and each being the main providers of support for those systems.
In the case of the Debian Common Core, you're going to be able to certify to a Linux distribution, and then there will be multiple support providers who can support that same platform and who differentiate themselves at a higher level up the stack. So they might differentiate with administrative tools, for example, but Systems programming isn't something that we should differentiate our businesses on. That is the piece that is enabling cost center, infrastructure, and we should all be cooperating to make that work and to make it the same so that application providers have a sane situation when they go to Linux.
MP: So this effort is a centripetal force, rather than a centrifugal force. How far could this go?
BP: We're trying to have an implosion, rather than an explosion, yes. And as how far it could go, the logical conclusion would be for SUSE Novell to join, because all of the packages that we're working on in the DCC are packages that SUSE Novell maintains single-handedly for their own distribution, and that does not make economic sense. That drives their costs up without doing anything about raising their profits. Having had SUSE Novell join, which I think is an achievable goal and not a distant future goal, the one remaining hold-out is Red Hat, and frankly I don't think that Red Hat has to hold out. I think that what Red Hat does is at a higher level than this. Red Hat has always been a little reluctant about the Linux Standard Base, and I think that they'd like to be monolithic, so my hopes aren't quite as high for SUSE Novell as a collaborator in the short term.
MP: Well, the timing seems to be fantastic, with the release of OpenSUSE coming at the same time. Do you see that as being fortuitous?
BP: Actually, I smile a lot about Fedora, and OpenSUSE is going to bring the same smile, because these are commercial distributions that are trying to get closer to the Debian way of doing things, and why not do it the other way? Why not start with Debian and add commercial support and services to it. I still believe that Debian is the superior technical solution. Wherever I visit companies, what I hear from those companies is "All of our developers run Debian on their workstations, but we deploy on Red Hat or SUSE Novell." What do those developers know that makes them run Debian on their workstations? And why can't they take that final step of actually deploying on Debian? I think that the DCC will help to fix this.
MP: Some asked during the press conference just now if this was a shot across SUSE Novell's bow or Red Hat's bow, and you said no. Could you elaborate on that?
BP: I really don't want to diss an Open Source contributor just because they're a different business. I think that Red Hat makes great contributions to Open Source, as do SUSE Novell. However, I do have a problem with some of the business methods, and I think that Red Hat Enterprise Linux has not been the best thing to happen to the customer in terms of getting excellent support at excellent prices. I think that anything we can do to make the market more competitive will both help the customer, because they have more choices, and will also help Red Hat fix their problems.
MP: I'm not an Open Source historian. Tell me a bit more about who was involved in the process of coining the term "Open Source."
BP: The Open Source Definition was originally proposed as the Debian Free Software Guidelines, and is part of the Debian Social Contract, which is up on the Debian website. It was essentially a contract that we were making with our users regarding what we would do as a distribution to be fair to the users. I helped propose it in 1997 as a result of a conversation with Red Hat and what was Red Hat's social contract with the community. Ean Schuessler was actually the person to raise that question, and there were some other people raising the question of what software belongs in Debian. I think Erik Andersen who works on BusyBox today, and was a Debian developer at that time, asked the question, "What is free software and what isn't." I helped combine all of those questions and helped to write both the Social Contract of Debian and what Debian felt was permissible, which became the Debian Free Software Guidelines. That was what was permissible in licensing.
Six months later, there was a meeting of a number of people concerned with marketing Open Source, and Christine Peterson, who at that time was the wife of Eric Drexler (the nano technology guy) suggested that we call it Open Source instead of free software. The feeling was that free software was a bad way to talk about it to business because free had other connotations that were not necessarily desirable to business people. So Eric approached me the day after this meeting, and said, "We're thinking of promoting free software, we're going to call it "Open Source." And I said, "Great idea. I'll take the Debian Free Software Guidelines and retitle them as the Open Source Definition and tie them to this name, so that we have an actual brand, which is all of the licensing that fits the Open Source Definition." That brand name was incredibly well accepted. Essentially everyone in Open Source follows those Guidelines to this day.
Six months after the Open Source Definition was announced, a reporter asked then Vice President of Microsoft, Steve Ballmer, whether they would Open Source Windows, and he started to explain to the reporter, "Well, Open Source isn't just source code, it's a particular kind of licensing." I read this in the press and realized that here I was sitting in a cubicle at Pixar, and Steve Ballmer had read our manifesto and was explaining it to the press. That really meant that the world was ready for Open Source for some reason, I think because they were not satisfied with the quality of software that they had at the time. Customers felt that the main vendor, Microsoft, was almost contemptuous of its customers. I think simply the existence of Linux and the fact that Microsoft now has a viable competititor, something it didn't have for a decade before that time, has changed the quality of software for everyone.
MP: And now we see that the concept of "Open Source" is spreading to other fields, such as hardware. Take Power.org, for example. Do you think its work meets the Open Source Definition of Open Source?
BP: I haven't checked out Power.org, but I think that you should look at OpenCores.org. It has Open Source integrated circuit designs. They have seven or eight CPU designs that are already Open Source. I think that this is what Power.org intends, although they may be a little tighter with the intellectual property, I'd have to look. So it is possible to do real hardware designs in Open Source, and I think we'll see more of them. There's also been the Wikipedia, which has been astonishingly successful. There's been my own book series. I've published with Prentice-Hall 17 books that are under Open Source licenses that are under all manner of licenses, including John Terpstra, one of the Samba core developers; Julian Smart, the guy who helped create wxWidgets has just finished a book on that; and all of these books are Open Source. You can take them and shove them in a copier and sell the copy the day you get them, it's perfectly legal. And all of that documentation is going to be with the Open Source software forever. And you know what? We didn't lose a cent on this, because the books sell just as well as any other technical book.
