> Well, yes, it's CLM that spurred me to make my comment about how Csound
> doesn't truly *compile* the orchestra code. It seems like CLM, however,
> does in fact do this, int the sense that it outputs a C source file
> which codes the CLM instrument

yes. CLM takes the Lisp instrument definition and translates it to C
source code.

> So in a sense, unlike C-sound, it is really a compiler

No. it writes a C source file, then calls the C compiler. It's not a 
compiler itself (although, historically, it used to compile to 65000
DSP code, but I think that's now completely abandoned as host 
processors outperformed DSP's).

> The not-so-cool thing about CLM, of course, is that there is only a
> handfull of unit generators, compared to the immense richness of them
> that Csound has. 

*WHAT*?  Csound is a pathetic inflexible dinosaur when it comes to 
unit gens.  Yes, there are about 50, but face it, stuff like out, outa,
outb,... etc are really only one.  CLM on the other hand has myriads of
functionally different unit generators as you can see its home page.

The funny thing about this debate is, though, that I personally quit
using CLM exactly because it does compile, inspite the fact that it
offers a vastly superior orchestra language and expressiveness.  I 
find myself re-writing and adjusting instruments all the time and can't
wait every time for the instrument to compile and re-link.  It is 
so annoying that I was willing to put up with Csound's pityful
stoneaged assembler language and meager expressiveness.  Just think
about it: no useable if or even switch statements; horrible "goto" 
instead; you can convert sound to "spectral" data types but can't get 
it back; etc. Csound is horrible.  The only real advantage it had to 
me was that it did NOT compile!  :)  Running Csound with Cecelia does
actually make a quite useful combination, IMO.

-Tobias

______________________________________________________________________

Tobias Kunze                     tkunze@ccrma.stanford.edu