It'll be language - hopefully a language that is modular enough and
extensible enough so that applications like MOOs, MUDs, Internet Talk,
anything could be written to use it. But being pragmatic, all I think we
should aim for now is a document language that can be downloaded and viewed
in a static environment, with hooks and obvious thought put into allowing
real-time behaviors in future versions. For VRML 1.0, because we want to get
this out soon and particularly get clients available across ALL platforms,
all we want is the ability to download a museum a room at a time, and walk
through it using a mouse and the keyboard. For VRML 2.0, an API should be
standardized so that external programs can control variables in the code
which modify the orientation and look of objects in a scene. VRML 1.0
should be open enough that adding that API later is no problem.
Brian