Re: [Adonthell-general] Anyone there?

[Alexandre Courbot]

> Seriously, I guess we're all quite busy at the moment. But it would
> still be nice to hear at least some sign of life :). An update to your
> dev diary for example. "At least once a week." - ahem! ;)

Uh? Who are you looking like that? ;)

But yes, you are right, I'm really overworked right now. For university
we have to participate to a timetabling challenge, write some mandelbrot
fractals rendering on parallel machines (IBM SP/3 - 4 x 16 Power3, very
nice machine ;)), a Multi Agent system, and so on... It's very a busy
semester, and we have time for nothing. I say "we" because Joel is in
the same shit than me ;) Anyway, don't expect us to be free until the
end of January. There our intership start, so we'll have plenty of time,
but not before I fear!

Still, Adonthell is often in our minds, although we don't have much time
to implement what we have to. But the map engine's physics are still
getting better and better - especially since we started studying
collision detection during our Graphic course! I guess I'll have a
discussion with the teacher about that soon.

Also, since we are making one of our projects (timetabling) in Python,
we are starting to masterize the language little by little and to know
what are the do/do not to get efficient code. And today, I have
discovered a marvellous package for Python: Psyco
(http://psyco.sourceforge.net). It's nothing else than a JIT! On our
timetabling algorithm, we went from 36 seconds of execution time to...
2.6 seconds. To solve the same problem! Maybe we could use that for
Adonthell as well. Of course the schedules aren't what eats the most,
but it could allow us to program more freely. Anyway, it's definitely
worth the try, since it doesn't require any other change to the source
than an import, and it can be optional (you try to import the package
and catch the exception if it doesn't exist - and life goes on). Only
drawback, python 2.2 is more than wished, which would make us upgrade
our requirements. On the other hand, Python 2.2 offers so great language
facilities than I don't think I'll accept to write 1.5 compatible
schedules again ;) Oh, and you *must* read the Psyco page where the
author explains why a Python implementation with Psyco can actually be
faster than a C one! ;)

> Also there are a couple of things that shouldn't require too much
> work, but rather some discussion. Like the v0.4 intro and plot. Any
> chance to get at least that done?

Sure. I'll try to get more involved in the plot discussion. At your
risks! ;)

I'm also wondering: is everybody so busy by real life? It would be cool
to see at least one field progressing, like gfxs (but James & Ben are
busy too AFAIK). Or maybe it's the right time to use every single effort
on finding new people to work with. Our recent searchs haven't been
concluding, maybe we will be more lucky?

Anyway, I'm waiting for news from everyone, too! :)

Alex.
-- 
http://www.gnurou.org