I am currently writing something comparable for my master-thesis. The
thesis itself is about collaborative applications (using
multitouch-tech.), their design and change of paradigms in current
interface-design principles. As example/proof of some concepts, I am
writing a sequencer that's geared towards live-sequencing. It
incorporates a lot of interesting concepts for music
notation/sequencing; in a way it is like ableton live insofar that it
supports outlining and playing with musical ideas independent from a
main-arrangement - just like live does with the triggering clips in
slots in the grid-based view. It is more abstract and general/powerful,
though. The sequencer itself outputs OSC and primarily drives
supercollider but you can use anything else on the synthesizer-side of
course (like processing to sequence visuals (according to the music)).
The application will be ready around december (deadline for the thesis).
The code base is fairly abstract, so more common devices for
input/controling like mice/keyboard/midi-controlls or any other
TUIO-stuff can be supported. I am planning to add some functionalities
that increase the workflow on the usual (single-user input)
computer-setup afterwards.
I've been producing music commercially for some years now mainly using
ableton live and a sequencer with a similar workflow is what kept me
from using linux or F/OSS as main production tools; so I expect the
final application to be basic, but quite usable :)
As a side note, keep an eye on the awesome Renoise, since they are
adding more and more live-features (midi-control/learning for
device-control (fx/synth) etc is already quite good); patterns can be
triggered live but currently not via midi.