[Sugar-devel] Bug tracking Vs Patch review

[Bernie Innocenti]

(I had overlooked this message)


El Sun, 01-08-2010 a las 16:47 +0200, Tomeu Vizoso escribió:
> - several (most?) FOSS projects track patches in a bug tracker and
> seem to thrive in contributions. Maybe they are wrong and we are
> smarter, but we shouldn't assume they are all stupid.

I thought it was a tiny minority, actually... A complicated review
workflow is a common "best practice" in Java/RUP shops. The only
high-profile OSS project that was using a bug tracker for reviews was
Xorg, but they switched away from it because it was not working very
well (I know this because I interviewed several Xorg hackers on
#xorg-devel).


> - this whole bugtracker vs. mailing list issue may be a red herring
> that we have come with to not have to deal with the real, harder
> issues.

I think it's one of many cases in which Sugar development is way too
bureaucratic and procedural. Alienating the contributors who code mostly
for fun is a big loss for a FLOSS project.

In order to get any non-trivial patch in, one has to go through several
stages of approval, each one managed by different people with
contrasting opinions: design process, feature process, review process
(one per affected module). As a result, lot of good patches have got
stuck along the way, bitrotting in trac for 6 to 9 months. Often, the
contributor gives up in exasperation before reaching approval. Even when
a patch gets approved, time has been spent 10% coding and 90%
interacting with web forms. Not fun.

Let's face it: Sugar development is damn boring! I hate to say something
so negative, but Sugar is probably the *most boring* FLOSS project I've
ever participated to.

One would think that this is a necessary price we have to pay in the
name of code quality, but how long does it take for a bug to be fixed in
Sugar? Months. Sometimes years. The quality of our codebase is indeed
very, very low and not improving very fast.

I'd rather take inspiration from the Linux kernel, which managed to
attract the best hackers by making patch submission fast, easy and
rewarding right from day 1. The book "Just For Fun" makes a good reading
on how to bring a project to success with a $0 initial investment. One
needs to engineer the process around social and psychological issues,
the technical aspects come second.

-- 
   // Bernie Innocenti - http://codewiz.org/
 \X/  Sugar Labs       - http://sugarlabs.org/