Coordinate Systems in VRML (check this out)

Jan Hardenbergh ([email protected])
Thu, 11 May 95 11:28:00 E

In VRML, geometry is defined in `object coordinate space.' You can use
matrix operations to position objects in a scene. These matrices are
sometimes referred to as modeling matrices, as they allow for the
creation of a model. They transform geometry from object space, where
they are defined, to `world coordinate space', where different pieces of
the scene are placed together in one global coordinate space. Since
nothing in VRML is defined in world coordinates, this is just a
conceptual crutch for us.

In the second step of the geometry pipeline, the geometry within the
'view volume' is transformed by the camera (also implemented as matrices
internally) into `window coordinates' or whatever the target display
surface is. Cameras allow the user to look at the model from any angle.

Cameras must precede the objects to be viewed. (Inventor == VRML?)

Render culling takes place when an object is not `in camera' after being
transformed to view space. This is tested by transforming the bounding
box of the object.

All coordinate systems in VRML are right handed.

===

I my original review of the draft spec, I said it would be nice to have a
little
blurb on coordinate systems. Gavin said "good idea, when can we have it?"
That was only 6 months ago.

I am not an Inventor guru, but this is my understanding. Is it helpful?

How much of this gets in is up to Gavin, Tony & Mark. But if it ain't
ready...

-Jan