> should be doing all of the processing to interpret the VRML and actually
> render the images and respond to user inputs. The server would have to
> provide the VRML along with object and world data to the client. This
> initial transfer of geometry info could be quite time consuming however.
>
Maybe a multi level VRML such as a High, MED, Low so that
different computers/user can access what they need/want for reasonable
speed/resolution.
High == Full realtime raytracing (is this too much)
Maybe just light source 3d simple objects
MED == Geometric / Symbolic shapes color filled
Low == Geometric shape no fill / TEXT even or ICON only
And each level up could also support the lower object pictures. IE a HIGH
could also have MED and LOW integrated Like ICONS on all 3 levels etc etc.
> To have any kind of interaction with the Virtual World, other than flying around
> it, we need to attach some sort of behavioural characteristics to objects.
I don't see the ability to have things really flying around My
conceptualization for a first step would be object placement and
interaction first.
> This is a current research interest of mine. What we need is a language
> that can describe what objects do when certain events happen to them, such
> as, begin selected, moving towards them, walking into them, animations,
> objects colliding with each other, etc...
I think it would be good to develop a higharchy of information,
then choose a languages or even modify a langugae (with macro/scripting)
capabilities to suit
> The use of PLG files would be a useful way of describing objects. However,
> I'd tend to favour 3D Studio .ASC files at the moment as then can be
> easily edited. The VR/REND386 .WLD files are quite good, but I'm not keen
> on the state-machine approach to animation and object interaction.
>
Have not seen theses I will have to check them out on my friends machine.
later