Re: Debian IS for the enterprise
Quite a long thread :)
> On my main server ... I have 1155 packages installed. On my workstation
> at home 447...
I do have 1194 packages at home. How many match yours ?
> And removing all that crap ... will make releases smoother and go quicker...
I don't think that releases are slowed by optional packages. It seems to me
that the *critical* packages are the ones who generates the more work.
Why ? Because other packages depends on them, so they need more fine tuning,
more work and more attention than optional packages. And they break the
system more easily.
Releases are slow because we do it (partially) the wrong way, and also because
it is not that easy. And also because debian packagers tends to spend more
time on flame war than fixing bugs :)
> I should say that SERVER is the keyword _I_ use when talking about
> 'Enterprise'... If you want a workstation, install SuSE, RH, Whatever.
Hey ... for me SuSE or RH are close to non free :) Debian is my server and my
desktop. I don't want to use f**ing rpms for installing openoffice or
crack-attack, when I spend my days working with dpkg and apt. How many Debian
developpers do want a server only system ? Not me, and I'm not alone.
> But since we're (more than?) 600 developers, maintaining 8000 packages
> is difficult. But if those 600 developers only had say 3000 (just a number
> out the head), then more bugs could be fixed ...
Are you sure of this ? Those 600 developers choose the packages they wanted to
work on. Nothing prevents them to help on *critical* packages. But they WANT
to work on their game package, or lg package. I have a server package, and
I'm preparing a game package. If you remove this one, I won't spend more time
on other packages. I do this one because I want to. This is my choice. Most
game packages are, I suppose, prepared outside duty time by debian
developpers.
8000 packages have been selected by 600 developers. Remove them the right to
choose what they want to do, and they'll do it elsewhere. These reflects the
choices we made.
> Those that are interested in making the best server distribution the
> 'Net have ever seen .... can continue. Those that aren't, ...
> can go work for RH. They make a sucky distribution so that should fit
> them just right.
You think there's no place in the project for people who like packaging games?
This is YOUR interest in Debian. What makes debian a succes is that each of us
find in it what he's searching for. And that's what we're working for. And we
put in it what we want to. No package did it into debian without a debian
developer willing to.
> Oki, fine. In any case, we need to rethink our 'target audience' and make
> sure that audience get the BEST of the best...
Debian is a free software project. Audience is for marketing and reality
shows. We do think in term of technical achievment and service. I do. This is
YOUR definition of the best, not mine. I used to be a redhat user, quite a
long time ago. I spent my time compiling programs, because redhat was quite
tiny at that moment. It bothered me. It was fun for some month. But today, I
want my 1194 applications to be up to date without an effort. Debian is doing
this for me, because 599 other guys are doing a part of the work.
You want debian to be better server suited ? Give some more help to server
packages. A sub project is a fine idea. Find some way of motivating people in
doing the work. Don't frustrate them in taking their toys from their hands.
That's not the way kids work, and we are all kids here.
Releases are slow, but thing are changing, very, very slowly. QA has improved,
a lot. There must be a way to do it without pain. I don't know how, you
don't. Somebody will find out, sure ...
Raphael
Reply to: