closing out old bugs of unmaintained releases

[David Woodhouse]

On Tue, 2008-01-08 at 20:06 +0000, Paul W. Frields wrote:
> > This seems a little reminiscent of the old school "go read the man page" 
> > response that helps and challenges some people, while turning off a 
> > greater number.  We can't grow a bigger, more vibrant community if we 
> > dis the people trying to get involved and help.
> 
> +1!  Fedora has come a *very* long way respecting the community of
> developers and packagers.  Now we need to get even better at welcoming a
> broader community of users, not the least of which are those who take
> the time to do that which we passionately implore of them -- "File a
> bug! File a bug! File a bug!"

I think there's more to it than that -- it's not just the PR angle. I
think that if our packagers aren't willing and able to take an active
rôle in ensuring the quality of the software we ship, we have a serious
problem.

I welcome the non-programmers who volunteer to package software, and
their work is very valuable -- but it really isn't going to turn out
well for Fedora if we prioritise quantity over quality to the extent
that we end up shipping packages which we can't _properly_ maintain.

The quote I posted earlier in this thread ("this is a bug in the
software, not the packaging") is taken verbatim from bugzilla, and I
find it extremely shocking. We have to do better than this, really.
Either that or go back to having a clear 'second class' repository
again, which I don't think we really want.

Perhaps one option would be for non-programmer packagers to team up with
a programmer/sponsor to take on the task of package maintenance? Where
currently a package has one 'owner', there would instead be two rôles --
a 'packager' and potentially a separate 'hacker'. Those wanting to
package software which they can't fully maintain for themselves would
have to recruit a programmer-type to sign up for the job with them.

I would certainly welcome being part of such a partnership -- I lack the
wit and the attention to detail for actually packaging software, and I
would like someone clever to do that bit, leaving me with the
monkey-work of poking at the damn thing in gdb when it's _really_
misbehaving, which I happen to find enjoyable and which my deranged
brain can actually handle.

-- 
dwmw2