> One way that a browser could handle this, in order to save network
> time, would be to only download objects that are not currently on
> file. User prefered defaults, ones on file on the local host, could
> be used in all places called for and new ones downloaded from the
> URL selected.
Actually I'm working on something like a MOO with graphics using PEX,
but I dont think that VRML is intended to do anything like this.
(Actually i got quite schorched when I first joind this list for
assuming that ;)
which once obtained may be interacted with but that is a browser
issue. Objects may perhaps be given behaviors, but this should not be
necessary, and I thought the survey results indicated that this should
be thought of, but not implemented for V1. There is not a concept of a
"VRML server", and such things as the moving a chess piece and other
users seeing the results is not a VRML issue.
Also, The desire to use iteration to define a stack of objects is a
nice idea, but not necessary, as that is a moder issue.
Finally, coordinate space is also a modeller issue. an object should
be defined in whatever local coordinate space, and when the designer
is building the scene, the designer will preforce know the orientation
of the object he is adding to the scene! if it is not oriented
correctly, translate it! The end results will still be an object with
a translation matrix defining its postion in the scene. A standard
local coordinate space would be nice, but if it is defined and
enforced, it removes all objects not designed for VRML, and the
concept of using an object in DXF format with the proper translations
That being said, I'll go back to working on my PEX stuff and let you
all ignore me ;)