Hi All,

I had sort of random idea just now for a collaborative music making
program that might be an interesting test of the API.  The idea is
that:

-a host program on one's computer could network with other hosts
-each user runs the host and has their own directory sandbox that
hold's a CSD and any relevant files
-you could connect to rooms of people to perform with, when you do,
everyone gets a copy of your sandbox directory, including csd and any
relevant files within it
-for every user in the sandbox, a Csound instance would be created in
the local host, each running all CSD's in the sandbox in realtime, the
host collecting the audio data and mixing together
-a performance interface would generate OSC data, sent to the host;
your own data would be routed to your instance of csound running on
your own CSD, and your performance data would also be broadcast to
other hosts and routed to their csound instance running your csd
-users could hot swap in new CSD's to be updated live (would require
stopping and restarting the specific CSound instance associated with
that user once that CSD is received)
-All CSD's in current room could be view by everyone else (useful for
learning new techniques from everyone)

Usages:
-collaborative performance
-general collaborative music making
-Csound study sessions
-Remote Csound teaching

For example, let's say for an online meeting group you could get
together and experiment with a synthesis technique, or perhaps work on
sample manipulation ideas.  Maybe a room sandbox could be a part of it
too so that all users in the room could share samples.  So maybe one
person says "Hey, I just got a great sound, check this out" and
everyone could hear it right then, as well as see the user's CSD.

Just a random thought I guess (I like the idea of casual performance
of art music and general music making with friends). It's perhaps not
very efficient to have that many instances of Csound running if it's a
big online room, but since each instance is isolated from talking to
another instance, the instances could be pushed onto separate node
computers which would communicate to the local host and mix audio with
it too.  I imagine would help stress csound's abilities a bit and
could expose some places where Csound might need work too.

Doing network code is fairly easy with both Python and Java, and OSC
libraries exist for both of those platforms, so those might be good
candidates for trying something like this out.

Any thoughts?
steven

p.s. - Anyone with experience using Jabber know if it'd be possible to
set up rooms as well as do such a thing as communicating non-chat
messages over it easily (for the OSC messaging)?