Hello list, Just my thoughts on this problem: 1. pulseaudio is now the prefered audiooutput system in most distros. It brings great new features which are useful at all. I like the feature mooving streams from one audiodevice to another. This makes using of headset comfortable and is realy useful. 2. many applications are now writing native pulse drivers. They will stop writing alsa drivers at a time we don't know. This will happen if pulseaudio is more stable. I have two use case where pulseaudio makes thing not better: 1. Accessibility You all know of the outstanding latency problems (which are now fixed using the simple pulse api) and adjusting the pa_buffer_att.*. A speechserver shouldn't run as user because some times it's useful to read some boot messages. Or are you turning on your monitor after login :-)? Unfortunately you are always writing that running pa as systemservice will not supported at all and should be avoided. The distributions are integrating pa this way and don't care about console users. 2. When I decide to run some daemons like asterisk or mpd etc with other uids this is a problem for pulseaudio. The Problem can be summarized in one sentence: Pulseaudio currently breaks multiuser systems and is only useful for one-user-desktop. Running more pulse servers at a time can work but is difficult for setup. Regards Halim @Lennart DMIX is enabled in alsa since 2.6.16 and works user independend. I can run my speechserver with uid speechd and run mplayer on my halim useraccount at the same time. an alsa application which uses directly the hw0,0 device will block dmix. When apps use alsa:default all works fine over dmix.
