the utter failure of bugzilla (and us?)

[Shaun Reich]

On Wed, May 25, 2011 at 2:06 PM, Matt Williams <lists at milliams.com> wrote:
> I've had experience with The Bug Genie and when recently analysing the
> various open source bug trackers available, it easily came out on top
> (particularly in flexibility and usability). What's more, its lead
> developer is a KDE user :)

Yes, from what I've seen of this it really looks like it fits so many
needs. I especially love the JSON api, the SCM integration, etc...
(yeah, I noticed...they even use KDE icons. they've got my money ;-p )

I also 100% agree with Beat Wolf:

In my mind, these issues need to be tackled in various paths. As
mentioned(?), there is not One Single Issue that is creating all of
what we know now. We should really come up with a list of what needs
to be tackled, and I believe that getting the best performance would
be achieved through all of those points.

(just to reiterate so we have some form of organization in this thread
and don't get lost)

1.) Better bug tracker (honestly, I avoid bugzilla like the plague. I
*want* to be there, I really do, but I just hate it too damn
much...perhaps too much for my own well being ;)

1a.) we should investigate if it fits many of our needs (Bug Genie
appears to be the best -- speak up on important points it may be
missing). Then we'd just have a test run, exactly like the icesrum
installation. Bug Genie has some really sweet looking features, and
even a JSON interface, along with SCM integration and such. I really
think that one could be our golden egg.

2.) Hunt for and assign maintainers for various Plasma components.
Even a "bounty board" blog post would probably help, I imagine.

2a.) even if the code is not innately known by that person, at least
there is a dedicated individual which can (ideally) be counted on.
This would be a far greater alternative to the now defunct workflow of
having "random people randomly trying to cover every single base there
is".

We all know that it's best to have specialized individuals, as they
would have innate knowledge of dups, up/downstream relationships, and
of the app itself.

3.) Some kind of karma system, as mentioned, which could hopefully
give us a sort of "filtering" mechanism to prevent morons from
creating ranty bug reports. This would also help prevent a huge
build-up, as would point #2

4.) Removal of bug voting. Unless there are objections, this seems
pointless. How many developers actually abide by it/care? And how many
users do you see running into the nearest IRC channel/Internet hole
asking for someone to vote up their "pet" bug report. Really silly if
you ask me -- it has no bearing on the development status, like High
Priority, Critical, etc. would.


-- 
Shaun Reich,
KDE Software Developer (kde.org)