Re: [Evolution] How our minds work (was: Msg Selection)

[Matthew Barnes <mbarnes redhat com>]

On Wed, 2010-06-16 at 21:39 -0430, Patrick O'Callaghan wrote:
> A few questions:
>
> 1) Is the bug queue growing or shrinking? Perhaps there isn't a simple
> answer to that, but nearly 5000 outstanding bugs for 4 or 5 full-time
> devels is a lot. No doubt there's some overlap between them, and not all
> are equally urgent, but still.

We have bug masters finding dupes and closing out obsolete or already
fixed bugs, and developers trying to keep up with new and active bugs.
Overall, seems to be a wash.


> 2) Do you have an rough estimate of what proportion of fixed issues
> correspond to enhancement requests? How about bugs or ERs related to
> Exchange?

Pretty low ratio.  ERs-to-bugs is already ~25% and a lot of bugs are
just quick fixes.  ERs tend to take awhile.  No idea about Exchange.
Two of our devs work on Exchange: one full-time, one half-time.


> 3) How are bugs and ERs prioritized?

As it gets close to release time we try to identify important bugs to
focus on, but otherwise I think we're mostly autonomous.  Partly because
we're scattered across time zones (India / Europe / North America) and
have only a small overlapping window each day (for me, early a.m.), and
partly because we're all just focused on different parts of Evolution.


> You seem to imply a LIFO ordering, so what is a reporter supposed to
> do when a report is old but still not dealt with? Perhaps some people
> mess with the priority/severity fields because they don't know what to
> do (though IIRC I've never done this myself).

I don't really have a definitive answer other than what I already
suggested.  Patches almost always helps, constructive comments in bugs
sometimes helps, discussing it on the mailing list sometimes helps,
hopping on the IRC channel or even the monthly IRC meetings [1] and
engaging us in real time sometimes helps.  Other times we're just too
busy to address it that day or that week or even that month.


> Also, if the priority/severity indicators are meaningless, why are
> they there?

It's a standard form for all projects using GNOME Bugzilla.  Small,
tightly-knit projects might even use them effectively.  For us, there's
just too many bugs.


> I hope you don't think I was being snide. I've made a number of ERs over
> the years and seen one or two being accepted. The fact that most of the
> time one gets little or no feedback about these things makes it hard to
> have a mental model of how they're dealt with.

No, that was more about comments with the maturity level of "WTF, this
bug has been opened for 4 years!  What have you people been doing!?  No
wonder Evolution still sucks!"  I'm not exaggerating.

Incoming enhancement requests tend to get marked as such, stamped with
appropriate topic keywords [2] and thrown on the pile.  Not because we
don't care, but because a lot of them require a large time investment,
possibly having to rethink deep or long-standing design decisions.  But
as long as they sound relatively sane I never close them no matter how
old they are.  Bugzilla is a gold mine of good ideas and every so often
I like to go in there with a pickaxe.


[1] http://projects.gnome.org/evolution/meetings.shtml
[2] http://live.gnome.org/Evolution/BugzillaTopics