Re: Debian-edu/Skolelinux and Edubuntu cooperation
On Mon, 2005-05-09 at 08:21 +0200, Markus Gamenius wrote:
[...]
> > Debian is right now one of the longest standing distros _and_ going
> > stronger by the day.
>
> Many people feels that Debian is geting weeker, not stronger. Partly
> because of developers runing to Ubuntu.
[...]
Seen it all before... remember Progeny? or any other "exciting" Debian
based distro's? The appear on the scene every two years, just as Debian
stable is getting really stale, and garner a lot of hype because they
are more up to date. Then Debian releases, they loose their edge, and
they fail to maintain the momentum to make another release.
Meanwhile, Debian absorbs all their good work, and keeps chugging along.
> > If all we need is newer stuff, then Debian will eventually have it.
>
> But when? Sarge is new to Debian standards, and it will be "new" when it
> is stable, but not for long. when will Sid be released? In 3 years?
I think sarge will be the last ever Debian to be "released", at least
under the current model. Debian has grown so much, that it is not
possible to get that many packages into a single coherent release. It is
the n^2 interactions problem.
Either Debian will have to reduce the coupling problem by breaking it up
into components (base, gnome, etc) that can be independently released
(Progeny is doing something like this I think...), and/or it will never
release, and be a constantly evolving pool of packages with things like
"testing" being the only attempt to make sure they all work together.
Note that the big argument against "testing" was security support. There
is now a security team for testing set up to prepare for the sarge
release. I hope this team stays there after sarge finally releases,
because testing is probably the future of Debian.
> It is wise not to use all the new stuff, but it is not wise to depend on
> to old stuf either.
[...]
Agree.
The reality is, software is a constantly moving target, and you need to
keep up. "Stable" releases trick you into thinking you have something
stable to hang off, but in reality it just means you are falling behind.
Perhaps it's a good idea to dispense with the illusion of stability, and
instead accept that it is dynamic.
Once you accept that, you realise that "not too new, but not too old" is
also constantly moving, just a bit behind the leading edge. The trick is
to have some mechanism for filtering those "just right" packages into a
distribution. The current "testing" is a very good attempt, but I'm sure
there are ways it could be improved.
--
Donovan Baarda <abo@minkirri.apana.org.au>
http://minkirri.apana.org.au/~abo/
Reply to: