Debian: A Brief Retrospective - page 2
Debian a Decade Ago
People often ask me what I would identify as Debian's most important contribution to the world. I believe it's Debian's decision to adopt a community-based development model. As far as I know, this marks the first time that a project intentionally set out to be developed by the community that used it. This is a central and critical component of the power of the open source movement--after all, if you remove the community from open source software, it's just software. In a lot of ways, Debian showed that "the Linux development model," as we called it then, could work for other projects, and that was a significant step forward.
In addition to Debian's role in the evolution of the nascent community-based development model, three other achievements stand out in my mind:
- Debian's biggest technical achievement was its package system, dpkg. Before Debian, Linux was typically distributed as a series of floppy disk images, which was convenient for the majority of us who didn't have Internet access but which made the system highly coarse-grained and very difficult to upgrade.
The package system concept, borrowed from UNIX, was a major step forward, as packages made it possible to upgrade the system piece by piece and, ultimately, to move the Linux distribution into the Internet era by allowing software updates to be delivered online (e.g., Debian's APT and Red Hat's up2date).
The package-based approach has been the centerpiece of every Linux distribution since, and network-delivered software updates form the basis of most distribution business models today. Ironically, it was the non-commercial Debian that pointed the way toward how Linux would ultimately be commercialized.
- Debian's biggest organizational achievement was its emphasis on project management and infrastructure. It's one thing to have a great idea that generates interest; it's another to have the necessary infrastructure in place so that, when the masses show up to lend a hand, they can contribute.
- Last, and in my mind most importantly, has been Debian's constant reminder that the strength of Linux is the "ecosystem" around it and not the product itself. This has been a crucial factor as Linux has grown from a non-commercial, hobbyist curiosity to a multibillion-dollar industry.
It is often said that Debian's founding goal was to create a non-commercial, "free software only" distribution. That's actually not true; at the time, such a thing would have been redundant--there was no other kind of Linux distribution in 1993!
Rather, I believe that Debian captured the spirit of those early years and brings that spirit forward to the present day. As a result, Debian is in many ways the "core of the Linux community" and a vendor-neutral "arbiter... to make sure everybody plays nice" (these are not my words, but rather the words of IBM and HP executives, respectively). It is absolutely critical to our collective success that we remember our roots, because it is the fertile soil in which those roots took hold that makes Linux so unique and so valuable.
In the early days of Debian, the package system and packaging standards we put together ensured that independently developed pieces came together into a cohesive whole. Later, as the number of Debian developers swelled from dozens to hundreds to more than 1,100, a project management infrastructure took shape to handle this massive task.
This, by the way, is why most open source projects never get beyond the idea phase (just browse through any 20 random projects on SourceForge to see what I mean). No matter how great the idea, there has to be a framework in which to contribute, or critical mass can never be reached. And without critical mass, open development projects are not sustainable.