From: Richard Bown <bownie@bo...> - 2002-06-26 07:22:56
An interesting strategy meeting which in my opinion is long overdue
for our kind of stuff.
I shall probably attend and will report back with The Plan. Despite
ALSA/JACK having (IMHO) the edge at the moment for what we want to
do it'd be good to still be a part of where aRts is going - as long
as they don't completely break us in the short term.
---------- Forwarded Message ----------
Subject: KDE/aRts strategy meeting
Date: Tue, 25 Jun 2002 11:24:36 +0200
From: Stefan Westerfeld <stefan@...>
Currently, what is summarized as "KDE multimedia development" works a lot
better (at least that is my opinion) than before the introduction of aRts.
We have a working sound server (maybe even the best sound server out there),
we have centralized PlayObjects and effects, we do a lot of code sharing, we
have a very modular media player on top of that, and so on.
However, I've been thinking quite a bit how to go to the next level ;).
It seems that aRts is doing a lot of things these days, mainly
- sound server
- media framework
- synthesis/music creation
and each of these tasks faces different challanges right now. For instance,
while for the sound server, APIs, speed, backward compatibility,
universality (usable outside KDE) and more than stereo support are key issues,
for the media framework we will need to care about threading, video,
modularization of monolithic PlayObjects, and for synthesis there are issues
like persistence, gui generation, latency and completeness of the module set,
and so on.
I was wondering if not a good strategy would be trying to seperate the tasks
from each other, so that we would have smaller but more consistent,
understandable and flexible units covering approximately the same scope. I.e.
we might use CSL to make KDE work with any sound server, redesign our media
player in a way that it uses a common codec API (to be written) but inprocess
decoding, and go for a new synthesis/music framework that contains effects
and the music stuff applications like Brahms/Rosegarden and artsbuilder
currently use. this is a rather longterm decision, that affects quite a
bit of code, I'd like to have as much discussion as required between those
developers that are interested and affected by this (possibly we decide that
going on more or less like before is the better way). I suggest starting
with an IRC meeting
- Server : irc.kde.org
- Date : Friday, 28th of June
- Time : 21 o' clock, Europe/Berlin time
- Channel : probably #kde-multimedia, in doubt ask me via /msg
(my nick will be stw)
-* Stefan Westerfeld, stefan@... (PGP!), Hamburg/Germany
KDE Developer, project infos at http://space.twc.de/~stefan/kde *-
kde-multimedia mailing list