Thanks for your help. I will take a closer look at the email archive.
For your information, I am planning to create a general-purpose,
user-extensible software synthesizer of the academic type, a successor to
Csound and Cmix, that will be usable via a number of standard interfaces,
including JavaSound, VST version 2, and Buzz. This synthesizer will be a
shared library written in C or C++. SAOL would be usable in this context
only if it were in some sense a runtime compiler. It would be just barely
conceivable to use your SAOL to C compiler as follows:
Modify Sfront to generate Java source code.
Embed Sfront in the synthesizer.
Run Sfront and general .java files.
Shell out to run the java compiler on the generated .java files.
Embed the Java virtual machine to execute the Java bytecodes of compiled
SAOL.
Other options that I am exploring are writing a new synthesizer for Csound
orchestras and scores, and writing an all Java software synthesiser from
scratch that also will shell out to compile Java source at runtime. I have
several parts of such a Java synthesizer written and working well in other
contexts (soundfile reader/writer, wavetable oscillator, envelope
generator).
In this multiple-use, multiple-platform context it would actually be more
useful to have a SAOL compiler that produced machine-independent, executable
pcode, something like what Csound does. Of course Csound is almost a
"little" language and that is why this approach is practical for Csound.
-----Original Message-----
From: John Lazzaro <lazzaro@CS.Berkeley.EDU>
To: gogins@nyc.pipeline.com <gogins@nyc.pipeline.com>
Date: Thursday, July 01, 1999 7:14 PM
Subject: Re: SAOL Compiler
>
>Hi Michael,
>
> Thanks for the note, and for checking out sfront --
>it's a moving target right now, especially in the documentation
>department, we're hoping to unveil a new website in a few
>weeks that will make it much easier to get the tools running
>and to write code.
>
> If you're trying to get it to work with a PC running
>Linux, you should be able to get sfront up and running quickly;
>if you're running Win98 or WinNT, we've had success using the
>UWIN environment (available for free at:
>
>http://www.xraylith.wisc.edu/~khan/software/gnu-win32/
>
> with the Microsoft compilers, we actually have gotten
>to the point of real-time interaction demos with that combination
>(although the Direct Sound driver currently with sfront is
>very preliminary, best case it will hang up your machine after
>around 5 minutes ... we're still debugging it).
>
> There's been a lot of postings on the mailing list
>about Java work in progress (if you haven't been on the list
>until recently, check out:
>
>http://sound.media.mit.edu/~eds/mpeg4-old/saol-dev-archive
>
> and search for Java, the postings start out negative
>but as it goes on a few people start indicating they might
>start working on it ...
>
> Realistically, these folks are probably your best
>shot for a SAOL implementation in the near term that runs
>on Java,.
>
> However, since Sfront at its heart spits out ASCII code
>that is quite readable, and C is relatively close to Java
>syntactically in a lot of ways, its not impossible to
>imagine someone starting with Sfront and turning it into
>Jfront. This will be much easier to attempt once we get
>the documentation done on how sfront works inside ...
>
> And since its licensed under the GPL, there shouldn't
>be any issues there ...
>
> Hope this helps!
>
>-------------------------------------------------------------------------
>John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
>lazzaro [at] cs [dot] berkeley [dot] edu www.cs.berkeley.edu/~lazzaro
>-------------------------------------------------------------------------
This archive was generated by hypermail 2b29 : Wed May 10 2000 - 12:15:24 EDT