On Sun, 28 Mar 1999 12:46:55 +1000, Lance Birch <lan### [at] usanet> wrote:
>Well, it could be done... all you have to do is specify the bounds of an
>object and an XYZ threshold.  If there is an object intersection routine in
>POV-Ray (which there has to be somewhere!) then you could make a patch of
>POV-Ray to allow keyword access to such a function.  Then all you have to do
>is an ordered test from the top of the bounding box to the bottom calling
>the routine.  If it returns 1 then you know that the point lies on the
>surface of the object, so you can join the previous row of triangles to the
>current and so on...  And there you have it, a MESH!  WOO HOO!!!

I don't believe this will work as well as you might think, because it's not
always obvious even to a human how to connect the dots to make a surface, let
alone to a computer, but there are other ways of creating a mesh from an 
implicit surface.  Once you do create a mesh, you'll want to have a mesh that 
you can use reliably in CSG as well.

The superpatch does have keyword access to the object intersection routines,
but I think it would be unacceptably slow.  Better to implement something 
like this in core code, especially because it's trivial to generate meshes 
for some primitives.

Post a reply to this message