>As far as optimizing goes, it looks as if quite a lot could be
>translated almost verbatim from Csound. To cite one example I happen to
>have noticed - in saolc 'buzz' is implemented by directly calculating
>every partial in a loop, using cos(). Csound uses the closed-form
>formula (and a table), which is ovbiously much more efficient.

Agreed.  Many of the inefficients in saolc are there so that
the code looks like the text of the standard -- the standard
says 'buzz' is a sum of cosines, so the reference software
implements it by summing up cosines.  I think this is an 
advantage to people from non-computer-music backgrounds whose
first exposure to a synth engine is saolc.

>So there is arguably a useful degree of overlap of interests between
>programmers/users of Csound and saolc. For composers working alone, on a
>single workstation, speed and range of facilities are likely to be far
>more important than portability under the terms of MPEG-4, so that I can
>see two semi-independent streams of development arising - the
>'canonical' source under the remit of MP4, which naturally needs to be
>carefully (and presumably centrally) managed, and a much more
>distributed ad-lib development for composers along the lines of Csound.

I agree with this entirely.  And the standard is fixed, so 
everyone will know what is the common denominator for making
portable instruments, and it goes up to each composer to 
decide whether they care about speed more than portability.
Not unlike programming in other languages.


>I am myself interested in working on saolc (partly because it will also
>be relevant to CDP users - especially the imminent enhancements!), on
>optimizing, adding support for WAVE and WAVE_EX, and no doubt some WIN32
>specific things.

I forgot to include WIN32 on my list of platforms before (because
that's where I'm developing it these days, so it doesn't need to
be "ported".)  There's no WIN32-ish features, just command line
interface, but it runs fine.

>Presumably there are dedicated saolc lists, so the question (apart from:
>what are those lists?) is to what extent discussion of saolc is welcome
>on this list, or is it too much of a digression off-topic? 

I don't know.  I'd welcome a judgment by Bath, so at least I know
whether to feel guilty about posting here.  

There's two SAOL lists: people interested in developing new 
compatible tools, compilers, and synths should send a 'subscribe'
message to 'saol-dev-request@media.mit.edu'; people interested
in making orchestras, instruments, and opcodes should send
a 'subscribe' message to 'saol-users-request@media.mit.edu'.

Best,

 -- Eric

+-----------------+
|  Eric Scheirer  |A-7b5 D7b9|G-7 C7|Cb   C-7b5 F7#9|Bb  |B-7 E7|
|eds@media.mit.edu|      < http://sound.media.mit.edu/~eds >
|  617 253 0112   |A A/G# F#-7 F#-/E|Eb-7b5 D7b5|Db|C7b5 B7b5|Bb|
+-----------------+