I would tend to be less in favor of an attribute-based system, and
more in favor of devising algorithms which operate on arbitrary
geometric objects, or groups of such objects. Attributes tend to
require a designer to forsee all of the uses that an object will be
put to. We want a system that is as open-ended to recombination as
possible.
It seems like many such algorithms are quite obvious, and extend
directly from computer graphics modelling and database technology
iteself. Stretching, shrinking, object substitution, inheritance,
joining, splitting, coloration, materials change, motional path, etc.
So it's probably not worth discussing the possible algorithms a user
could employ - that's more in the realm of technical practice.
What may be worth more discussion, though, is how the _environment_
could change and influence the process of the user's symbolic
construction. Starting from a black void is one possible approach,
but it is only _one_ approach. There is no particular reason to
consider this to be "better" than any other approach, although it
might profit us to think about what sort of user-user interactions an
initial black void will generate.
For instance, I have occasionally played a game of "free association"
with people at a local coffee house. Each person says a word, one
person after the other. There are no particular rules for what word
will be said. I find that when you get a random mix of personalities
together to play this game, you tend to get a "least common
denominator" string of word associations. People invariably get stuck
on talking about sex. In the limited time they are alotted to come up
with a word, they most often cannot come up with anything else.
The environment, as both a physical machine and a conceptual machine,
heavily influences the kinds of dialogs that will take place within it.
Beyond the basic recipe for Surrealist experience that we have
established so far, it is worth investigating how the VR environment
can manipulate the users' experiences.
I think we need to be careful that we don't introduce our own
interpretation (however correct it may be), to the purpose of objects.
It should be the user who decides what an object means to them...?
No matter what interpretation we may have when creating an object, a
viewer will always generate their own. So there really is no risk, in
this regard - just a battle of whether the artist or the audience's
opinion is more "important" to someone.
Cheers,
Brandon