Well, I missed this whole thing! Bummer.
I'd say I'm with William on the whole sample thing... at least, in general. I think it's true that the vast majority of VSL use is intended to emulate the "live" orchestra. However, I also think there's a whole world of interesting work to be done in the area of using these samples as individual sounds, with their own inherent musical qualitites. This is potentially a very interesting line of investigation. I'm trying to do a little of it, and am hoping to explore it in depth in future. The difficulty is that I think that it demands quite a different way of thinking about musical form, because once you've abstracted these samples into pure sounds, then suddenly any notion of arranging them based on the parameters of instrumental performance is essentially arbitrary. Not good, not bad. But simply one method of organization among an infinitude.
If one makes this step (which is extremely difficult to make) then I would argue that s/he is indeed in the territory of the early electronic music composers. I say this because to make this step would require a complete dismantling of one's understanding of what these sounds signify, and how they relate to one another... which may, in fact, be impossible to do.
It's even arguable that this is a *more* difficult step to make than the pioneers of electronic music made, because it implies a process of tearing each sound away from a 300 year history of associations. Seriously. To find a new context for a new sound is not such a big feat. But to find a new context for a sound with as much "baggage" as a flute, or violin... now *that's* tough.
Anyway, that's just my two cents.
Here's something I've done that mucks around with VSL a little. Not at all what I'm talking about, but at least it's clear that this is not intended for "live" instrumental performance:
http://rubato-music.com/Media/WFT_orch_morph.mp3(just an excerpt from a contemporary dance piece I'm working on)