[freesci-develop] Re: Event-driven sound server

[Matt]

On Tue, 11 Sep 2001, Christoph Reichenbach wrote:

> > 1) how it's started (ie fork(), new thread, etc)
> > 2) how the game loop signals the sound server
> > 3) how the soundserver passes cues back to the game
> > 4) the mechanism we use to sleep until the next note is fired.
>
> OK. This is all fine on UNIX, Win32, VMS, DOS, and other systems where
> this is possible.
> However, PalmOS and MacOS classic provide MIDI output as an OS service-
> OTOH, they do _not_ provide v?fork() or MT (AFAIK).

MacOS 8.x and newer might be worth investigating -- I really think PalmOS
is a waste of time. It is simply too restrictive of a platform and
requires too much hacking around it's memory/display model for it to be
worth the time/effort, in my opinion.


> Now I'm aware that we're pretty far from supporting these anyway-
> FreeSCI's nowhere near the portability Sarien has achieved- but at least
> MacOS classic might be a worthwhile port, and we should at least consider
> its perceived deficiencies when designing a new sound architecture.

I agree -- MacOS 8.x and MacOS X would definitely be worth shooting for.
(I say 8.x, because it did have 68k support in addition to PPC support.)

--
http://www.clock.org/~matt