Lex,

> First, doesn't OSS require adding kernel modules for most systems?  This
> would prevent many users from being allowed to install OSS--and thus,
> OSS isn't as portable a target as the web page makes it seems.

I don't know (yet).  One of today's projects is to find out, and then
build VMs for Alpha and Solaris that include sound support if
possible.

(Besides, kernel hacking is good for the soul.  Maybe the promise of
Squeak sound would incite people to go and indulge themselves in a bit
of Zen Driver Installation? ;-)

> Wouldn't it be worthwhile to get at least Linux support
> working?

Sorry if it wasn't clear from my original message: sound is working
perfectly on Linux using OSS...

> Right now Unix Squeak doesn't support sound at all, on any platform,
> unless you install outside patches.

....it just isn't released yet.  (OSS is built into the kernel on
Linux, so there's no "customisation" problem due to kernel modules or
whatever.)

> Wouldn't Linux support be better than nothing?

It'll be out later today, if I can avoid "real" work for a couple of
hours.  ;-)

> Why don't we stick *some* version of *someone's* code in the standard
> release, and then have all the rest bang on it?

Idem.

I wasn't aware of your OSS support.  I'll take a look today.  It's
entirely possible that you've done a better implementation than me.

> a single, good, bug-free sound implementation, even if it does only
> work on Linux.

Unix > Linux, and always will be.  As long as there is a
platform-independent route to feature X then it is implicitly the
correct solution - even if it's a nightmare compared to going for
support on a single platform.

Ian