Pulse and ALSA - ability to merge?

[mrguitarmann]

Please look at this link on UbuntuForums? - it's regarding the state of audio. http://ubuntuforums.org/showthread.php?t=988142

Would the PulseAudio devs ever consider merging ALSA and Pulse into one entity?

[coling]

Personally I don't see the point in merging.

The merge would only be a name anyway, e.g. it would be political or advertising motivated, nothing more.

The ALSA project is really two-tier already. There is the kernel driver structure which is very well defined and is pretty good from what I can tell from general opinion. The second part is the alsa library/plugin system (basically user-space stuff). This is less well defined and the alsa API itself is pretty complex and can be abused very easily if you don't know what you're doing.

So what you currently call "ALSA" is really two projects anyway, they just have the same label.

So if pulse was to "merge" with ALSA, what would we call it. "ALSA Server" woot! Nothing changes.

I don't mean to be flippant (well maybe a bit :p), but the only real problem with linux audio is the fact that due ot the open source nature of things many apps do not use the ALSA API correctly which causes problems with Jack and PulseAudio etc. when used via the plugin system.

If apps were coded better (well not better but "safer") then there wouldn't be any problem.

Ubuntu made a big mistake with it's LTS version. They integrated PulseAudio very badly. It's really up to the distro to make a good job of it. I know, I spent a very long time trying to get things right when we introduced it in Mandriva. My work paid off and I'm very happy with the end result, as were the majority of Mandriva users. Ubuntu have realised their mistake and did a much better job of the pulse integration last time round thanks to TheMuso?.

IMO, merging solves nothing. Better distro support and integration is key, and more pressure on application developers that use alsa lib to follow the safe alsa subset guidelines that Lennart wrote is a much more productive idea IMO.

Just my €0.02 ;)

[coling]

Also another solution may be the libsydney project. It's a long time coming, but it may be the ultimately correct solution.

See an old article about it: http://0pointer.de/blog/projects/foms-lca-recap.html

Lennart will be able to give more up-to-date info when he returns from swanning about the place having one of these strange things call a "holiday"... I wouldn't know what that's all about... I'm chained to my desk :p

[coling]

Even though you closed this ticket in a way I'd agree with, it should be noted that I'm not a PA dev per-se... I've hacked about a good bit and I know the project fairly well, but the only really authoritative voice is Lennart. That said, I suspect that (from the above linked blog about libsydney and our previous chats) he'd probably concur. I wouldn't like to state that for certain tho' :)

Cheers.