Re: A full-fledged VRML proposal

Paul Burchard (
Mon, 26 Sep 94 21:42:59 -0600

"Gavin Bell" <> writes:
> On Sep 14, 11:34pm, Paul Burchard wrote:
> > Coordinate-index links enable all sorts of interesting
> > interactive services, and I strongly believe that
> > they should be part of the base level VRML (I leave the
> > format to you).

> There are several pieces of information that might be interesting:

> -- object-space point
> -- object-space normal
> -- object-space texture coordinate (S & T coordinates)

> How about another field of type SFEnum called 'map' that
> can have values of NONE (the default, no information
> tacked onto URL), POINT (object-space point tacked onto
> URL in the format you suggest), NORMAL, or ST_POINT
> (normal or texture coordinate tacked onto URL)?

> Are NORMAL and ST_POINT overkill?

Not at all, that ST_POINT option could be very handy (although
admittedly our viewer won't be able to handle it in the near future).

The "map" field is a really good idea.

> > [I propose adding link relationships, like Embed]
> [Gavin sez the semantics are fuzzy]

Yes, I guess it is essense, these link relationships
are a poor-man's version of some basic things one would want to do
with scripting. On the other hand, we haven't cleared up the
semantics of scripting yet.

(When I say scripting, I mean having a link point to a script instead
of a VRML file, so that the running script will act on the viewer and
possibly feed it some VRML along the way. The semantics would be
further defined by the standard API that scripting languages would
use to control the viewer.)

Paul Burchard <>
``I'm still learning how to count backwards from infinity...''