I find that the major advantage of dealing with anything about
computers in terms of "theory," is that one can manipulate theory a
LOT faster than one can manipulate implementations. It lends itself
well to breadth of exploration, and the risk of such explorations is
minimized.
There comes a time when you have enough theories about a given task,
and you have to get down to implementation to learn anything further
about that task. However, I'm quite certain that within this
particular mailing list, we are nowhere near that point. So I
advocate sticking to "theory as usual," and worry about the actual VR
state of the art when we've got a body of theories we're ready to try
out. Then it will probably be time to move over to the "vr-art" list,
which has a more pragmatic focus.
Cheers,
Brandon