Re: RFC: Changing the NM system
Martin Schulze <firstname.lastname@example.org> wrote:
> Colin Watson wrote:
>> Perhaps we need a more organized system of sponsorship, so that people
>> who are stuck waiting in the NM queue can do QA work with some degree of
>> ease. At the moment it seems to be largely a matter of whether you're
>> lucky enough to find somebody who'll quickly and consistently sponsor
>> your uploads.
> They already can by looking at the large amount of bug reports and by
> fixing and classifying them. This doesn't require an account and a
> direct way of uploading packages.
Getting QA work done (instead of just working on it) is harder than
that, even if you are a developer (or have a very responsive sponsor
like I do). Everyone can work on the problems, but unless the work
is being done, it is just a waste of energy.
The real problem for the quality of Debian, IMO, does not lie in the
quality of orphaned packages, i.e., those in the good hands of the QA
team. If nobody want to adopt that package, then it is likely that
people have already stopped using it. The real problems are those
whose maintainers were MIA. As they are (supposedly) being taken care
of, nobody worries about them, and as the maintainers were MIA, the
bugs are not being addressed. So what should somebody who would like
to work on QA do?
Well, you have to make sure that the maintainer is really MIA, because
they might take it personal if you jumped the gun and fixed their
packages for them (and to avoid duplicate of work). But that takes a
long time, and you can never be sure. And there do not seem to be a
standard procedure on what to do once you are fairly sure that the
maintainer is indeed MIA (that may be defined somewhere, but I seldom,
if ever, see people doing that). Of course, you can work on the bugs
without touching the package, but what would be the point for doing
that, as the poor users would percept no changes or whatsoever?
Note: I have been trying to reach the package maintainer for tin, who
disappeared after orphaning fetchmail (but never mentioned tin), and
the maintainer for sslwrap, who seems to be MIA, in both cases
unsuccessfully. So I speak from my experience.
-- Chuan-kai Lin