Re: Idea for maintaining packages up for adoption
Le Fri, Jul 15, 2005 at 03:25:44PM +0200, Jeroen van Wolffelaar:
>
> As far as I can see though, there are not really technical problems
> with the low maintainance of orphaned packages, but more the reason
> that they are still orphaned -- nobody has a really big interest in
> them.
When you say that, or you're wrong, or "nobody" is to be understood as
"nobody in the Debian developpers" in which case you're right. But the
fact that a package is of no interest to a Debian developper does not
prove the package to be useless or deprecated.
Get one step further. There are some packages that will never be
orphaned, because it would mean the end of Debian: packages like dpkg or
debhelp. Would you claim that debhelp is the most useful package in all
Debian for all users? That sounds false. It's clearly one of the most
useful packages for Debian developper, and so they all have interest in
it being maintained and improved, so it will never get orphaned, but it's
much less usefull for Debian users, which are the real target of the
distribution.
The orphaned packages are quite the opposite: they are of no interest
for the Debian developpers, so there is no maintainer any more, but they
might be of interest for users, or for developpers in other technologies
than the ones used for Debian. Of course, some of those are deprecated,
or of bad quality, but this is not always true.
> Also, bear in mind that if nobody really is interested in a given
> package, it is sometimes better to just move onwards to get the package
> removed, instead of having it fixed by someone just for the sake of it
> (and not because one uses the package). Natural selection on Debian
> packages (and in general on open source projects) is a good thing to
> have, and just fixing orphaned packages reduces the natural selection
> efficiency.
I don't agree on that point. The natural conclusion would be that the
users of a package are natural candidates to be maintainers. Which mean
that a package *must* have developpers in its users. Which can lead to
the fact that there are mostly packages for developper-related topics.
This is probably one of the reasons why there are not enough packages in
certain areas (nothing to help establishing a good scolar edition of a
text, for an example, while there are a numerous free software to do so),
but I don't think it's a good thing.
The fact that no developper cares about a package just prove that the
package is not interesting for developpers, not that the package is a
bad one. It can be seen as a clue, but not as a proof by itself.
Regards,
Benjamin.
Reply to: