Re: Low Averge Bug Counts
* Anthony Towns <aj@azure.humbug.org.au> [011215 23:26]:
> I'm not sure why everyone's Cc'ing me on this thread.
Sorry, I won't Cc you anymore. Heh.
> On Sat, Dec 15, 2001 at 01:50:01PM -0800, Grant Bowman wrote:
> > > What about people like myself who are stuck with crappy bits of
> > > software like trn which have lots of bugs which are unlikely to be
> > > fixed any time soon? It seems like my efforts are being belittled by
> > > this sort of thing.
> > What I suspect you have an issue with is using the Bug Rate as a
> > statistic to measure effectiveness. So what suggestions do you have for
> > how we should proceed?
>
> Personally, I think measuring developer "effectiveness" is a really bad
> thing to do. It doesn't lead to any desirable outcomes (like focussing
> effort on things that need it) and it's not particularly accurate (it
> doesn't measure contributions that don't result in bugs being closed,
> nor does it take into account packages having different distributions
> of bugs).
>
> But, shrug, whatever.
I should have been more precise. What I should have said was "using the
Bug Rate as a statistic to measure activity." My goal in bringing this
up is to increase participation by rewarding active packagers,
developers and testers. Some people may read into the numbers as I just
did, but that's not truly the intent.
If we also have more QA discussion in the QA group rather than WNPP
packaging (which is also important), I think this will help the Debian
effort and increase the quality of the released software. I hope it
increases participation from non-developer types who can be *shown* how
to file good bugs. It also gives frustrated users an outlet for their
frustration and would simultaneously help the software quality.
That's the idea, anyway. I've had enough commercial QA experience to
know that it's never quite as straight forward as that.
--
-- Grant Bowman <grantbow@svpal.org>
Reply to: