Debian and I
Off the Beat: Bruce Byfield's Blog
Debian is the most influential Linux distribution ever. Of the 305 active distributions listed on Distrowatch, 147 are derived from Debian, and 87 from Ubuntu, Debian's most famous off-shoot. In other words, 77% of the distributions being used today wouldn't exist without Debian. That makes Debian's nineteenth anniversary on August 16 worth a moment's reflection, not just technologically, but socially as well.
For me, Debian and free software are hopelessly intertwined. While I had played about with Linux before, I only went hardcore when I started work on 5 July, 1999 at Stormix Technologies, an ultimately unsuccessful attempt to commercialize Debian. From there, I jumped ship to work at Progeny Linux System, which was founded by Ian Murdock and funded by Bruce Perens' short-lived Linux Capital Group, and very much traded on the reputation of the two Debian leaders behind it.
In those few years, I worked with some of Debian's leading developers, including Branden Robinson, John Goerzen, and Jeff Licquia, and at conferences stumbled across most of the mover and shakers in free and open source software as well. In short, my introduction to free software was also my introduction to Debian.
The technology
Since then, I've always had Debian on one machine. Usually, it's been my main workstation. On other machines, real and virtual, on any given point I have threeor more other distributions running, but I've come to rely on Debian for my main computing.
Partly, my reliance is based on Debian's technical choices. I appreciate the installer, which allows my choices to be as detailed as I need and want. I appreciate, too, the dpkg / apt-get package manager, which is still far quicker than Fedora's yum, and includes a thriving ecosystem of tools to help me to recover from any situation that I might land myself in while installing software.
Nor have I found any repository system that suits me as well as Debian's. The division into Unstable, Testing, and Stable allows me to choose the balance I want between reliability and cutting edge. If I want an utterly dependable system, I can stick with Stable, at the price of using year old software. But if I want something more recent, I can venture into Testing or Unstable, both of which are more reliable than the repositories of many other distros. In practice, I usually venture in Unstable for standalone technologies such as web browsers or desktop environments, while being more conservative about core system libraries and utilities.
The philosophy
However, I am equally attracted by Debian's sustained idealism. I am basically a free software advocate, but, like many people, I sometimes find myself irked by the Free Software Foundation's dogmatic approach. By contrast, in Debian I find a more respectful approach.
Anyone who has ever dropped by the debian-legal mailing list can be in no doubt that Debian developers are as devoted to the idea of free software as anyone. Nor has any other major distribution bothered to develop its own definition of free software, and worked so hard to apply it. Yet, despite the occasional general resolution to the contrary, so far the project has resisted the temptation to impose its ideas of software freedom on its users.
For example, Debian installs with only the main section -- the one containing only free-licensed software -- enabled. However, you can easily enable the contrib (free but based on non-free software) and the non-free sections if you choose.
The same goes for kernel firmware. If you want, you can install non-free firmware that often works better than any alternative. But you also have the alternative of installing a completely free kernel.
In other words, Debian encourages you to install a free system, but whether you do so is left for you to decide. Personally, I always choose a free system, but I appreciate that Debian lets me be the one who decides.
This faith in individual extends to project governance as well. What other major distribution has general resolutions to allow major policy issues to be decided by active contributors? Or not only allows contributors to vote on the project leader, but determines results by the Condorcet method, which many believe is the most democratic way of determining results?
Being a consumer rather than a contributor, I don't participate in either Debian's general resolutions or leadership elections, of course. But I take both as a sign of good faith efforts to live up to the project's ideals. While I'm sure that in Debian, as in any organization, behind-the-scenes discussions settle matters more often than votes do, I have to respect any group that puts itself to the inconvenience of trying to do the right thing.
And many more
All in all, I have only two reservations about Debian. First, it has been slow until recently to recognize the contributions of non-developers, such as artists and technical writers.
Second, in the past it has been a hostile place for women, who remain greatly under-represented among Debian developers, especially when compared to other distros such as Fedora. The recent general resolution to accept a diversity statement may be a sign of change to come, but in these areas, Debian still has a ways to travel.
In other aspects, though, Debian has managed to survive while maintaining a high leve of idealism. Possibly, this idealism has cost it market share compared to the more commercially-oriented Ubuntu, yet even that is not certain: the last 18 months, 19 Debian-derivatives have been started compared to 13 Ubuntu-derivatives.
Moreover, distributions such as Linux Mint and CrunchBang, which started by using Ubuntu, have answered user demands and are now available in Debian versions. If Debian's influence is less obvious than it was a decade ago, indirectly it seems as strong as ever.
I wish Debian decades more -- partly for the selfish reason of personal preference, but partly because I think that free software in general could use a lot more of Debian's anarcho-syndicalism.
comments powered by Disqus