Re: Why I changed my mind about Separator/Property changes

J Gwinner/VisNet, I (75162.514@compuserve.com)
27 Nov 95 12:03:14 EST


Gavin:

>>Hmm... anybody want to take a shot at figuring out where Lights, WWWAnchors,
WWWInlines, prototyped user interface components, etc fit in?<<

What I'm doing is basically to 'ignore' most of these fields :-( as I've
recieved absolutely no comments on 'CompuServe' specific URL's. This will
probably change, as I'll just 'make something up' that will irritate everone
(tough, I've waited for comments and not gotten much feedback)).

However, overall, I'm generally going with 'object as a file'. That is, most of
my avatars and forums are listed as seperate files, of which the geometry is
loaded (and LOD's, textures, etc). Cameras are ignored. (controlled by the
larger application).

WWWAnchors is probably an object by itself, or a property of an object. I guess
it could be a NULL pointer in the object pointer hierarchy.

Inlines, should probably be allowed as 'any' of the pointer fields, i.e. their
use would be pretty much transparent to the syntax under discussion, as is now.

Lights and cameras could be tough. I basically see lights as being an different
type of object; most browsers must (internally) keep a seperate track of lights
anyway. I would also, however, like to associate geometry and other nodes with
a light, so that a spotlight could be a physical object tied to (in a behaviour
sense) the object. So how about defining (in the pointer list) a camera
pointer? (I don't like this either, the list of pointers is becomming large).

As a general comment, can we just consider the syntax of VRML 'as is' adequate,
and instsad assume that when parsed, objects will have (internal to browsers)
the pointers to properties/nodes as listed? i.e. you could still code
'standard' VRMLl, but when building behaviours, assume that each object has it's
own pointer to a texture2 node, etc as you've listed. I esalize that this is
specifying the internal operation somewhat, but has the advantage that the VRML
language doesn't change. I think that we are going to have to make _some_
assumptions about how browsers operate internally when we discuss behaviours.
This may not be clear, should I exlain this better?

Forgive me if this has been discussed before, I'm still catching up.

== John D. Gwinner ==
== CyberForum/Visual CompuServe
== VisNet, Inc
== "Making CyberSpace Real" (TM)


  • Next message: Jan Hardenbergh: "RE: PHIL: Government Involvement"
  • Previous message: J Gwinner/VisNet, I: "Re: Why I changed my mind about Separator/Property changes"
  • Maybe in reply to: Bernie Roehl: "Why I changed my mind about Separator/Property changes"