[LAD] alsa and OSS (again?)

[Ivica Ico Bukvic]

OK, so here's a question that may seem inflammatory but let me assure you it
is by no means my intention to add fuel to the fire. Rather I would like to
clarify some of the salient bits for myself (and hopefully others). So,
please do enlighten me...

If, dmix/snoop is seen as not real-time safe and as such cannot be used
concurrently with JACK (I know that one would probably not want to use it
that way anyhow since in pro-audio last thing you need is "you got mail"
ding in the middle of your recording session, but I don't think that
dmix/snoop were designed with this in mind but were rather quick fixes to
the lack of hardware mixing for average Joe user who only needs dings and
beeps, which all in all still makes it only a half-solution to the problem,
no?) and we have pulseaudio to in part solve this headache while still
allowing JACK to access hw practically directly with no added latency
overhead or loss of real-time safe operation (at least that is what my
current understanding of the pulseaudio project is)

and

If ALSA sequencer is subpar to Midishare (something I cannot attest to as
I've not had an opportunity to mess with it, so I am making an assertion on
what I've learned from the existing discussion, again please correct me if I
am wrong)

then

Why bother with either dmix/snoop and/or ALSA sequencer when they appear not
to be complete solutions?

Wouldn't it be better for ALSA devs to have more time to focus on making
sure that drivers and supporting operations are working so that we don't
have to worry about Freebob vs. ALSA vs. OSS vs. whatever (I know, this is a
bad analogy to draw given the different scope of the Freebob or what is now
known as ffado project but hopefully you will understand what I am alluding
to here), and have pulseaudio and Midishare take over the dmix/snoop and
sequencer respectively? (again, please note that I have no idea whether
Midishare can do this without the ALSA sequencer, so I am here speculating a
bit).

In other words, provided that my conjectures are true, I could see ALSA
becoming simply an API to access the soundcard and everything else being
moved out of the ALSA which IMHO could help spur some of the ostensibly
bogged down aspects of Linux audio experience.

Any enlightenment here is appreciated.

Best wishes,

Ico