Some writing about sound

[Tom Tromey <tromey cygnus com>]

Ville> I am _not_ trying to start a serious flame war in here, but our
Ville> (Adam and me ) gnome sound proposal follows (Adam correct me if
Ville> I am wrong):

Ville> 	-Base API to NAS which has been here for years. And also it
Ville> has sopport in some app's (like mpg123)
Ville> 	-NAS is network transparent which is added plus.

[ The followups list many reasons not to use NAS: hard to port, works
  poorly on Linux, etc ]


My take on this whole subject is that, generally, the thread has been
oriented along the wrong axis.  There's been a lot of discussion about
the poor quality of the NAS daemon.  Ok, I can understand that: it
sucks, and we have to replace it.  No problem -- but that's just the
particular implementation.

The deeper questions for me are about protocols, compatibility,
usefulness, and the future:

* Protocols.  Is the NAS wire protocol inherently bad?  Ignore the
  fact that the existing NAS daemon sucks -- assume we've already
  decided to completely replace it.
  If the protocol is a good one, then it doesn't make sense to get rid
  of it.  If it is only mediocre, or if it is genuinely bad, then it
  does make sense.
  I don't know enough about audio, and I'm not really interested
  enough to learn (right now), to make this evaluation.

* Compatibility.  How many apps already use NAS?
  If many do, and the protocol doesn't completely suck, then adopting
  it makes sense just for compatibility's sake.  (For me this is the
  biggest point.  YMMV.)
  If no apps use it, then it hardly matters.
  What does KDE use?  (This isn't critical if we can easily write
  a protocol translator -- but this does depend on the underlying
  daemon having good feature coverage.)

* Usefulness.  Can NAS do everything we need?  Can ESD?

* The Future.  Is NAS and/or ESD tied closely to a particular platform?
  It's ok if the implementation has not yet been ported, provided
  that nonportabilities don't leak up into the protocol layer.


Anyway, those are the things we should evaluate.  I don't actually
have a preference for NAS -vs- ESD.  Ideally I'd like to see a person
knowledgeable about both (and audio in general, perhaps) re-examine
the various daemons from this perspective.

Oh, one very important point I left out:

* Practicality.  If one solution has ready volunteers, and the other
  doesn't, then that should be a big plus for the former.


All the above is IMNSHO, of course.

Tom