This (as a maintainer of one of the most used chunks of Csound code) is
not as simple as it may seem.

The code is really divided into two very separate parts - the common
code and the implementation dependent code.  MIDI, sound I/O stuff, and
user interfaces fall into the latter category, and the "csound" code
which is derived from the old MIT code falls into the former.  In fact,
there is even a line to be drawn between the system (MIDI-SOUND)
interface and the user interface.

Unfortunately, these divisions are at times pretty fuzzy.

We (Mike Berry, Matt Ingalls & I) at Mills DON'T give our code to
people asking for it, although we are planning on doing so at some
point (after all, it was given to us!).  We *do* feedback fixes,
improvements, etc. to the common code to jpff.

For one thing, it is difficult enough keeping up with bugs, requests,
etc. for the user base.  Having a developer base, which is essentially
what will happen if we give out the interface code, would add to
that difficulty.  Then there is the possibility of some later
generations of the code "getting out there" and people getting back to
us with bug reports about things we aren't responsible for.  Also, our
user interface code, and at this point our system interface code, is
almost entirely made from scratch.  There are definite reasons to want
it to be "fully settled" before we put it out there (or at least to
feel done with it, or maybe just sick of it &8^}).

Also of course, if you give out code, you want it to be fully commented,
nicely formatted, meaningful variable names, etc.  (not that ours *isn't*
all these things!).

The common code, and to some degree perhaps the system interface code,
IS public property.  We dutifully report any bugs, fixes, improvements,
etc. we find in it to John Fitch so he can incorporate them into his
more (or less) "vanilla" code base at Bath.  Anyone developer that
doesn't do this is, in my opinion, acting against the interests of the
community and him or her self.

John works very hard to try to keep SOME kind of implementation
independent code base available, for which he deserves the everlasting
gratitude of this community.  I have spent much of the last few weeks
bringing our code (of diverged lineage) as much in line with his as
feasible, and have seen how much he has done first hand. Any developer
that finds a bug in the common code and doesn't send it back to John just
doesn't know which side his bread is buttered on, as far as I'm
concerned.

Dave Madole
Technical Director, Center for Contemporary Music
Listserv Administrator

Mills College
Oakland, CA 94613
510-430-2336

madole@mills.edu