I'm trying to cruft up some sound code for Linux, and the main worrisome
part is snd_InsertSamplesFromLeadTime().  The key difficulty is that with-
out resorting to a very non-portable DMA technique, once I commit a sample
buffer to the device, I can't fiddle with it any more.  Further, at least
with Linux, there is no signal sent on buffer exhaustion, so I'll need to
trigger a poll with an interval timer.  (I know... BLETCH!)  Consequently,
I need some idea of how often I'll need to schedule the interval timer.
if #startPlayingImmediately: is any hint, I need to merge no closer than
1024 samples.  Is this just a random number?  Might it get smaller?  Could
we set by convention a lower bound for insertion?  This could make the job
of sound hacking much less ambiguous.

Sincerely,

-- 
Christopher Oliver                     Traverse Communications
Systems Coordinator                    223 Grandview Pkwy, Suite 108
oliver at traverse.com                    Traverse City, Michigan, 49684
   (define magic (lambda (f) (lambda (x) x)))
   (define (more-magic n) (lambda (f) (lambda (x) (f ((n f) x)))))