This is an interesting thread...
In *my* original conception of a shared-space/place setup, the
imagery would be user-definable. This is not the way VRML is
heading (exactly, wait lemme think about that... no, you could
do this in VRML)
Using an example of a broad representation of the present Internet,
you could swim around using the shapes defined by the server, or
you could ignore the shape-definitions, and supply your own.
.com domains are blue pyramids, .edus are fluffy pink blobs,
sparcs are green nuggets, Alphas are silver diamonds...
A 'browser' could choose to accept non-local object definitions,
or override them with locally defined ones. Thus, the pencil with
the eraser would become a whirling mass of tearing blades...
In my world.
I am of the opinion that 'reality' is probably pretty individual.
(Gibson's "consentual hallucination") (yeah, we could argue
whether my "red" is the same as your "magenta" for hours)
Why not design Cyberspace(tm) with this preconception in
mind? (who's mind?)
I wrote to Mark to ask whether the current Lab browser supported
stuff like server-side includes/parsing. This could go a long
way towards making a user-selectable environment representation
doable in VRML, as it stands. I put a user-selectable background
pattern intot a web site I'm doing. Ubique is using user-selectable
faces in their Internet chat software (see: http://ubique.com) -
I think that this is a good avenue to pursue.
[ d*b ]