Re: replacing esd with gstreamer

[Bill Haneman <Bill Haneman Sun COM>]

One big concern with our sound frameworks in general is how mixing and
multiple/concurrent sound requests are handled.  These are of special
interest to accessibility, so any move to revise/revamp the gnome sound
interfaces needs to consider that issue and ensure that, at least for
the maximum reasonable subset of hardware,  concurrent sound is
possible.  Furthermore, some way of _preventing_ concurrent sounds (i.e.
a policy that queues them sequentially instead) probably is necessary in
order to support text-to-speech users.  Another key characteristic and
question is how such support will coexist with console
applications/sessions which make use of sound, since many screen reader
users also use console-mode text-to-speech systems when doing
command-line and virtual terminal work.

I don't know the answers to these questions, or even know to what degree
they can be addressed in gstreamer, but I do know that the answers are
very important to some users.  This is currently an area of active
discussion which is slated to lead to standardization activity in the
Free Standards Group in the future.  Can those "in the know" clarify the
relationship between the esd-compatibility wrappers, gstreamer, and the
use cases/requirements I mention above with regard to software or
hardware mixing and concurrent/sequential sound?

Thanks in advance,

Bill