Re: quality assurance for games
On Fri, Dec 14, 2012 at 10:03:54AM +0800, Paul Wise wrote:
> Every time this topic comes up, it makes me want to quit the team and
> to some extent, quit Debian altogether. Removing stuff instead of
> fixing it is very demotivating, which is probably part of the reason
> there hasn't been much feedback on this topic.
I feel pretty much the exact opposite. IMHO, having bad quality packages
rotting in the repositories has a negative effect on the perceived quality of
Debian as a whole, and if something is not being maintained (or is not
maintainable) to a high standard it should go for that reason. It can always be
reintroduced if someone steps up to do the job; there's little motivation for
people to bother if the packages linger regardless, the risk of them being
removed from a release is sufficient for people to consider their priorities,
or for users to step up and become contributors. The packaging work need not be
lost, that can be preserved regardless of whether the binaries make it into a
release (or sid.)
I am demotivated whenever I install a package to try out and find that either
the program itself is poorly written, or it's poorly maintained and bugs are
ignored¹. This happens time and time again. That's demotivation as a user as
well as a developer.
Reassigning stuff to -QA should not be a long term solution, and neither
should dumping it in a team including the games team.
There's simply too much to fix, relative to the amount of human resource
available. That applies to teams as well as Debian as a whole.
From my POV Markus is taking on a difficult and uninspiring but necessary task
and I applaud him for it, especially for trying to be open and transparent
despite the negative (and sometimes passive aggressive) responses that this
inevitably results in.
In a way, this reminds me of the two polar opposite philosophies on
Wikipedia.
To just return to your first point
> Every time this topic comes up, it makes me want to quit the team and
> to some extent, quit Debian altogether.
I sometimes feel this way about things (not least what I've described above),
and it generally means I'm too emotionally dragged into Debian and need to
cool off a bit. You work *really* hard across the board on Debian and that's
fantastic, there are only a few people like you (and have only been), but you
can't fix all the problems by yourself, and indeed, many of your pet peeves,
and mine, will *never* be fixed, we've just got to live with it. Or move on.
¹ I'll pick on Kartik Mistry for a moment, as the last example
of those such bugs for me were #679234, #679246 and #679247,
all ignored, whilst I've seen Kartik post several ITPs in the
meantime. I'm considering working on rationalising the media-tagging
package ecosystem post wheezy, but that will be a big and hostile
job I fear. E.g. tagtool uses the old id3lib3.8.3 library that does
not support id3 v2.4, is abandoned upstream so far as I can see,
whilst libid3tag from MAD exists and has none of those faults. The
former should be got rid off and rdeps updated or removed too IMHO.
Reply to: