Future releases of Debian
some of you might still know me, some time ago I was a Debian developer.
>From the point of view of a user there are two big problems of Debian:
1. stable releases  are infrequent
2. stable releases are unpredictable
Point 1. affects many people - the software in Debian 3.0 is currently
roughly one and a half years old. It's not only a problem if you want to
use the latest KDE or GNOME - it's often non-trivial to find a new
graphics card that is supported by XFree86 or a new scanner that is
supported by SANE if you want to use Debian 3.0 . It's not needed to
release as often as other distributions, but one release a year is badly
If you need more recent software there are currently the following
1. unstable (the name says everything)
2. testing (lack of security fixes, changes every day)
3. stable with one or more backports, often of varying quality and
with additional problems if you mix different backport sources
4. RedHat, SuSE, Gentoo,...
In production environments you usually have to choose between 3. and 4.
The other big problem is that stable releases are unpredictable - at
least for a non-developer like me it's currently not predictable whether
Debian 3.1 will be released in 2004, 2005 or 2006.
Yes, "It's ready when it's ready.", but a big system like Debian doesn't
become stable without a plan how to get it stable. Other free software
projects like gcc and Mozilla have roadmaps and relase cycles that are
working. Yes, they sometimes change their roadmap and rarely release at
the estimated release dates, but you can at any time roughly estimate
the date of the next stable release.
Some time ago I had some (perhaps now slightly outdated) ideas about how
to organize a release cycle  and as long as my time permits I'm
willing to help in QA and/or release management if this results in a new
stable release within clearly less than a year.
It's not important for me whether some ideas I have/had are used or
whether everything works completely different.
The only thing that really matters for me is to get a distribution that
is reliable usable in production environments.
 To prevent people from nitpicking, when I talk about "stable
releases" I'm talking about major releases like 2.2, 3.0 or 3.1, not
about bugfix releases like 3.0r2.
 This shouldn't start a "Why wasn't XFree86 4.2 included in
Debian 3.0?" discussion - XFree86 4.2 would have lowered the
problems for some time, but within half a year the problems would be
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed