I think we decided early on that there was no way to reconcile SGML into
a file format for describing virtual worlds, at least not at the level
necessary to get this to meet our requirements in a practical amount of
time and resources. That's why the "M" was changed from "MarkUp" to
"Modelling".
However, I support the notion of the creation of a language (or the
development of VRML into a direction) that allows someone to essentially
describe the high level structure of a scene ("This is a kitchen with a
table and two chairs") and the browser fills in the details as needed,
depending on how much the user feels they need to see exactly what the
world creator meant (the flip side being, how much can be recreated from
cached objects the browser knows about, like a description of a "chair"),
the bandwidth and time constraints, etc. We're delving into AI territory
here, so let's be pragmatic and start with hierarchical scene description
languages first (which is what VRML is) and work from there.
Brian
--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
[email protected] [email protected] http://www.hotwired.com/Staff/brian/