What are bugs? Are bugs really work items?

Christian Robottom Reis kiko at async.com.br
Tue Oct 19 11:26:51 UTC 2004

On Thu, Oct 14, 2004 at 09:45:42AM +0100, Gervase Markham wrote:
> Jason Pyeron wrote:
> >I think we fight to many battles for the sake of perfection (this causes
> >bit rot). Or even more trivial are the situations where a patch is given
> >as step 1 of 2, and step 1 is never committed since step 2 was not done
> >too. If there cannot be an agreement or an absolute "RIGHT" decided for
> >step 2 then don't r- if step one is good on its own merit.
> If step 1 is good on its own merit, then don't present it as "step 1 of 
> 2" :-)

Well, Jason may have a point. The point is that too often we are
reluctant to allow people to file follow-up bugs, which is something
that would relieve them significantly of a lengthy, disencouraging
review cycle.

The problem that derives from that is having people file follow-up bugs
and disappear. But I don't think we can go on living with that
mentality, because IMO it tends to intensify the issue which brings
people to disappear in the first place.

> >Sometimes changing a 'function name' or other detail for 'cosmetics' is 
> >just not worth it.
> I'd disagree. Coding standards are important for current and new 
> developers to know what's going on. Changing a function name is usually 
> so easily that it's hardly ever not worth it.

Sure, unless you take two months to review and then say "ah, can we
please rename X to Y and Z to W and split A into B and C?". These are
all valid requests in themselves, but after a lengthy review cycle you
may have to leave them to follow-up bugs, or give them up if you want to
keep the contributor, or take it up yourself and fix it. Not doing any
of those means we just bitrot everything people do and then we're
sitting here in a circle wondering why nobody stays around.

Being allowed to mark something as FIXED is a major incentive to
participating for me -- I would actually be a lot more productive if I
was allowed to mark things that are `almost perfect' FIXED and move on
to fallout or follow-up there towards `perfection'. I think we err on
the side of caution too often and this brings us to stalemates all

Take care,
Christian Robottom Reis | http://async.com.br/~kiko/ | [+55 16] 3361 2331

More information about the developers mailing list