One thing that moving the audio interface into the web application space affords is the distribution of performance and/or compositional tasks. When you think of developing an electronic instrument, you go through the process of envisioning 3 parts of the resulting instrument: what it will sound like - the audio processor, what effect the user will have on the audio process - the compositional affordances, and how you will control it - the instrument interface.
I like this idea of a distributed composer. Composers borrow (or to quote Stravinski - great composers steal) from other music all the time. There are a finite number of combinations of pitch and rhythm that can be used - even smaller numbers of usable musical forms, gestures, progressions, and morphologies of broader scope. This is not to say that composers can't create unique and individual music, but that the elements that make their music unique and individual are most often references to passages and musical thoughts in other music. (If not found directly in another piece of music.) With the fairly recent addition of remixing audio works to create new pieces, the time has come to explore how this can be used. Copyright issues aside, a distributed interface through web pages could quite easily allow people to compose themes, gestures, harmonic structures, audio processors, sequences, and all sorts of musical thought that could be put into a music collective or database. Once this is established. Composers could essentially collaborate in creating all sorts of music - borrowing whatever portions of musical thought that were necessary to create their aural envisioning.
Now back to the copyright element! Who would be the composer on the created work? Attribution becomes a very difficult thing. Funny enough, we already go through this process as a composer only composes music based on experience - wether it be aural, visual, logical, etc. (which brings up some other cross-disciplinary attribution issues). Now though, we can watch this creative process and have a record of where and how musical thought developed. Being able to directly watch and evaluate musical creativity.
Sociological study anyone?