Re: Base distro + Requirements definition

[Mikael Hallendal <micke codefactory se>]

sön 2002-06-23 klockan 18.41 skrev Jeff Waugh:
> <quote who="Mikael Hallendal">
> 
> > RPM's is imho better. It's a nicer package system to work with, and
> > since I don't think we will be able to use the packages out-of-the-box
> > anyway I think that we should choose with respect with what will be the
> > easiest way to manipulate the packages/add new ones.
> 
> I don't think that's entirely true. A few examples of packages that we (as
> the GNOME OS distribution thing) should have little to no interest in
> maintaining:
> 
>   - X
>   - PAM
>   - openldap client libraries
>   - most of the basic GNU stuff that should always be there
>   - most application packages, say evolution, etc.
>   - glibc, basic system stuff
> 
> This is all stuff that other people can, and should do. We don't need to
> have people who either know about this stuff, or people who attempt to
> fumble through it. Building from, and working with, an established
> distribution provides network effects that we shouldn't just waste away. We
> have better things to do.

Sure, we get this for free with any of the distributions named.

> [ Yes, some stuff we'd want to do ourselves, and some packages we'd
> customise to a certain extent. That doesn't mean we can't have a good,
> thorough platform to work from. ]
> 
> > >   - infrastructural elegance
> > 
> > Which elegance? :)
> 
> /etc/network/interfaces, packages that 'just work' (restarting services,
> good init scripts, configuration file handling, etc), Debian has a lot of
> very well thought out infrastructure, and some that isn't all that well
> finished off, but it would be a senseless waste of time to reimpliment
> these.

Same here.

> > The way we would have to go would be further if we were using Debian
> > than Red Hat.

> Only in some areas. The usability issues we'd want to address both
> ways, but I think Debian has the infrastructural stuff mostly worked
> out. RH gives us some short term positives, with a lot of mess, whilst
> Debian gives us tidy, long term benefits. Plus, contributing our
> usability/sexiness work back to Debian will give them more reason to
> be interested and helpful with our effort.

I agree that using RH would give us more short term positives. Though
that is the same for choosing Debian in front of Bagheera too. Since all
the packages we need works well in all three discussed distributions I'd
say that any of them would serve our purpose.

Looking at how well we'd be able to integrate with other distributions
using the same packages I'd say that:

Red Hat) Pretty much none. Red Hat is a company that has a product which
         works very well, they will not make changes to it to make it
         easier for us.

Debian)  We would be a small distribution always having to prove
         ourselves and why our idea is better then the current. The 
         Debian model also makes it hard to get things in. And we'd be 
         forced to interact with _LOTS_ of different developers which is
         very time consuming (many of them that probably don't have the
         time/interest to help us).

Bagheera) We would be one of the distributions helping out designing the
          meta distribution. That will put us in a good place for 
          getting our ideas in mainstream. Using Gentoo gives about the 
          same drawbacks as Debian though it's a younger distribution 
          where they haven't locked down everything yet (which is both
          good and bad for us).

> > Using Bagheera would be even further _but_ we get the freedom of having a
> > system where we can do pretty much what we want (as long as it's
> > reasonable).
> 
> That's not freedom - that's a metric fuckload of work.

When the fuckload of developing the package system is done (which we
won't have to do) we have a much better system that will serve our
interests much better.

> > >   - people working on stuff that may not be in our direct interest (fixing
> > >     bugs in X, etc)
> > 
> > We get this stuff anyway, without having the people on our team doing
> > the actual fixing.
> 
> Sure, at the code level, but not at the distribution / integration level. If
> we're doing all that work ourselves, we're wasting time on solved issues
> when we should be looking at the ones that *haven't* been solved. Let's
> concentrate on our areas of interest, not the basic building blocks of a
> distribution.

same here, if we don't maintain the package ourselves someone else will
make the package include the fixes from other distributions.

None of these points sets Debian before Gentoo. Yes, Gentoo has less
developers _but_ each developers does more work than each of Debians.

The only thing that I can see that speaks for Debian is Debconf (which I
understand doesn't really serve our purpose well enough anyway).

The number of packages is totally uninteresting.


Regards,
  Mikael Hallendal

-- 
Mikael Hallendal                micke codefactory se
CodeFactory AB                  http://www.codefactory.se/
Office: +46 (0)8 587 583 05     Cell: +46 (0)709 718 918