OSNews

Well... seriously now. Here is my take on Debian in the year 2007.

Up until a few years ago, Debian development was mainly important only to Debian and Debian users. Certainly, there were Debian based distros, but they were few and far between and definitely second string players. (In popularity; Not necessarily in quality.)

In recent years, *strong* Debian based distros have arisen. This gave the Debian devs quite a shock. (Funny how shocking "success" can be sometimes. Be careful what you wish for... you might get it.) They had never had to worry about competition in their own backyard. They had never had to struggle with the challenge of retaining .deb repo compatibility in a world that they did not control. But that's really the subject for another post. I think that issue has not actually turned out to be that much of a problem.

What I do want to say is that the effective function has changed. And I do not believe that it is due to Debian redefining itself. Debian's role has been redefined by forces outside of its control.

IMO, the most important thing that the Debian project does today, by far, is their development of Debian Testing.

Testing seems to be what most bona fide Debian users seem to use anyway. But, more importantly, the strong Debian based distros, which (IMO) have a lot more users than Debian proper, depend upon that solid core produced by the Debian devs in Testing. If Testing went away, Ubuntu, Mepis, et. al. would suffer greatly and quite possibly die.

Debian has been redefined by the larger community as being the solid core... the trunk... from which a whole solid family of distros spring. This may or may not be what the Debian devs wanted or intended... but that is what we've got.

Personally, I don't see that there is any shame in that.

For a while, I felt that perhaps Debian should shift to developing Testing, ad infinitum, without actually making any releases. (PeeWee Herman fans might think of this as a cable knit sweater that someone keeps on knitting and Knitting, and KNITTING!) ;-)

But after more consideration, I decided that was probably a recipe for allowing Testing to float off into some ivory tower somewhere. There really does seem, to me, to be a need for an actual release. That gives the devs focus. Some realworld goal to achieve to keep them solidly connected with reality. By working towards an actual release themselves, they stay in better touch with the needs of their off-spring distros, which are also trying to make a release that works well in the real world.

So my perception of Debian's role today is as the solid core upon which other distros are built. And I think of their actual stable releases, like Etch and the future Lenny, as being "reference implementations" which are good enough that a lot of people actually use them.

Notably, these strong Debian based distros sprang up during the longest release cycle in Debian history. And during this last release cycle, I heard some complaints that Testing was moving too fast and making it harder to build a distro from. (Didn't Mepis decide to start basing their distro off Ubuntu?)

In the context that I am presenting, this seems suboptimal. Perhaps it would be best if Debian aimed for longer release cycles than last time.

But this would require that Debian devs accept their primary role as a core for other distros.

They would need to accept the idea of Debian being a proto-distro more than a distro proper. It would mean that Debian would need to work closely *with* its children, rather than intending its reference implementation to compete with them.

At any rate, it's hard to deny that, far from being dead, Debian has gained in strength and popularity in a huge way in the last few years. Even if a lot of that strength and popularity is not due to the spread of Debian proper.

Debian would do well to listen closely to one of my favorite RedHat mottos:

"A rising tide lifts all boats."

This is as true for Debian as it is for Linux in general.

Edited 2007-04-15 19:53