Re: Minutes of SoC meeting

[Federico Mena Quintero <federico ximian com>]

El mar, 06-03-2007 a las 23:01 +0100, Vincent Untz escribi�> > >  + performance analysis projects are interesting but might not produce
> > >    concrete results in the end
> > 
> > Interesting.  How did you come to this conclusion?
> 
> Hrm, I don't have the logs of the meeting, so maybe someone will be able
> to jump in and help with the reply.
> 
> But, IIRC, what was told is that performance analysis is something hard,
> and that it might require a lot of effort to get (code-wise) a small
> result. The result can still have a big impact on user experience, but
> it's hard to know this before the performance analysis is done (ie,
> before the end of the project).

I don't know why the Nautilus project didn't succeed, but it sounds like
it was due to lack of guidance.  At the time very specific problems were
known to happen in Nautilus; they just needed to be tracked down.
Instead, the result was some rather general profiling of parts of the
program that had already been fixed, or that were not relevant to the
known problems.

Writing code is easier than reading it.  I think SoC is a great
opportunity to focus a lot of energy into teaching people how to use
tools and their brains to understand foreign code.  Performance
optimizations are just a good vehicle to do that :)

  Federico