Number of Bugs per product
mattyt-spam at tpg.com.au
Tue Jan 20 06:32:36 UTC 2004
On Tue, 2004-01-20 at 04:49, Christian Robottom Reis wrote:
> That's easy. We lack developers, and most of the developers we have
> are too busy to commit significant time to fix their assigned bugs :-)
Bzzzt. Wrong answer.
> One thing that *would* help would be a mass reassign of bugs we're
> *not working on* to nobody at bugzilla.org. I'd be a lot more comfortable
> rediscussing this list once we stopped listing bugs assigned to default
> owners indiscriminately.
That's certainly true for enhancements, but I don't think cleaning up
the bulk of 850 bugs is out of reach *if divided up more evenly*,
remembering a lot of those are probably enhancements.
While assigning off to nobody is a gain in transparency and perhaps
might help others to see what needs to be done, it's also a continued
cop-out, and that's what I'm trying to avoid.
The bug distribution should be reasonably uniform. Given 850 bugs to 12
developers, that's an average of about 70 bugs. A reasonable range is
therefore 35-140 bugs per developer, we have two developers in that
range, one of whom hasn't done much coding at all in the past few
years. Everyone else is either below it, or way above it with no hope
of ever fixing all their bugs.
I think if the bugs were more evenly distributed across the developers,
we'd each have a more manageable task to work at. We already have NEW
versus ASSIGNED to distinguish in progress reports, let's not reinvent
Assigning bugs to nobody smells to me like a black hole, much like
REMIND and LATER were. Sure, it might help non-core developers find
issues to take, but the bugs are more likely to be fixed by core
developers, and this is just going to take these bugs further off their
radar. At least at the moment, the component owner is at least partly
responsible for their bugs.
I don't believe the problem isn't that we don't have enough time, but
that we're spending too much time writing features and not enough fixing
the bugs with them. We have enough time to fix these bugs, we just need
to *want* to. And I think with gentle reminders we can make people want
to. At the moment it's just "out of sight, out of mind". No one here
doesn't care about these bugs, we just all get dispirited and take it as
too much work. If we were reminded more often of these bugs, I think
we'd get them fixed more often.
I'm not asking for some enormous effort to reduce the bug count down to
zero, I'm asking for ways in which we can turn the upward trend to a
downward trend. Assigning to nobody isn't going to have a significant
impact on that. I don't want to be standing here in three years time
complaining about our 1500 bugs.
Do we need to shame people into fixing their bugs? CVS blame and work
out who inserted the most bugs? Most bugs per LOC? What reviewers let
through the most bugs? I don't really want to go down that road, as I'm
sure everyone wouldn't, I'd rather people just took responsibility for
their fault and/or pride for their areas. However, at the moment it
seems to me that there aren't too many other roads that fix the problem,
and some good statistics could help us understand what's going on.
Matthew Tuck: Software Developer & All-Round Nice Guy
My Short Autobiography: 1985 Grade Bin Monitor 1990 Class Clown Award
1992 Awarded Most Likely To Spontaneously Combust 1996 Crowned Galactic
Emperor 1998 Released From Smith Psychiatric Hospital
More information about the developers