ABRT frustrating for users and developers

[Christoph Wickert]

Am Montag, den 18.01.2010, 21:58 -0800 schrieb Adam Williamson:
> On Sun, 2010-01-17 at 15:12 +0100, Christoph Wickert wrote:
> 
> > I doubt this very much. Many people don't report the bugs when the app
> > crashes but later, many reports in a row. Most of my reports read "I
> > have no idea what I was doing when foo crashed", even if they
> > submitted
> > it straight after the crash. Only 2 out of ~40 contained the
> > information
> > I needed to reproduce the crash reliably (as a site note: both are
> > fixed, so the number of crashes fixed it 4 but not 3 as I wrote in my
> > initial mail. 4/40 is still a bad percentage)
> 
> 'Bad' in what way? it's probably 4 - almost certainly 2 or 3 - more than
> you would have fixed if abrt didn't exist.

4 is an absolute number while I was talking about a percentage. 10% of
useful bug reports means I spent a lot of time on the other 90%. If I
just got the 4 bug reports from active testers who are willing to
provide all the necessary information, I would have been able to fix the
bugs too, but without wasting time.

But it's getting better. ABRT is getting better and I was just able to
(hopefully) close another two bugs. :)

> I used to work in a supermarket, and noticed that it'd be much easier to
> run a supermarket smoothly if there were no customers. They do insist on
> coming in and messing up the shelves and dirtying up the floors and
> asking stupid questions.
> 
> In much the same way, it'd be ever so much easier to run a distribution
> really *efficiently* if no-one ever used it...:)

In that analogy, most of the customers return something that they bought
before in your supermarket. I guess you don't have a problem giving them
their money back, but you expect them to at least tell you what's wrong
with your goods instead of just throwing them at you. ;)

Regards,
Christoph