I think it's too early for this. Don't get me wrong, i really like the simplicity of the OSS API, and i like its multi-platform approach, but it still has some issues about things that work out of the box with Alsa (jacksense, power management) and we must fix these first.
PulseAudio is also gaining popularity and is getting into major distros (ubuntu 8.04 will have it as default audio framework), and that has all the neat features of OSS: mixing, per-app volume control remembered between sessions and supports almost all the sound API's out there, plus network ability with latency compensation. When ran on top of Alsa these features make them both better than OSS feature-wise and noone would like to switch.
We wouldn't have success in this action until OSS has anything that Alsa does, and until Pulseaudio works flawlessly on top of OSS too (on my box it currently fails to run).
IMHO, we can initiate this only when OSS will be a full feature-wise drop-in replacement for Alsa, and even then we couldn't be sure that the "Not invented here" syndrome wouldn't apply once again, like it did a lot in the linux kernel development history.