VRML spec

Colin Dooley (cdooley@ibm.net)
Mon, 2 Oct 1995 22:36:38 GMT


First, some more questions about VRML 1.0:

1) Why is it compulsory to index the vertices of a face set, but not
compulsory to index anything else? Surely it should be all or none...

Second, somebody wanted comments on the version 2.0 proposals.

1) In the "Changes" section the author wanted to know which order to do the
translate/rotate/scale. I say Scale-Rotate-Translate. Translate should
definitely be last. Rotating something, followed by a non-orthogonal scale
doesn't make a lot of sense....

2) If we are going to complain about speling mistakes in the 1.0 spec, I
think we ought to correct "tessilate" which appears 4 times in the "Sphere"
section :)

3) Curves are useful, and can be used for many things other than motion
paths and interpolations. However, I think we need to spell out very
literally the mathematics of the curves used (maybe a separate appendix) -
even to the extent of providing some C code so that we are all doing exactly
the same thing.

BTW. What exactly is a "3DSTUDIO" curve? What advantages does it have over a
b-spline? Is the fact that the author is from Autodesk influencing anything
here?

4) In materials, what exactly do the "ambientFactor", "diffuseFactor" and
"specularFactor" do? It seems that they scale the respective colour
properties, but as these are floating point numbers it doesn't make much
sense to include an extra number to multiply the colours by.... :?

5) I don't like the "hither" and "yon" clip planes associated with cameras.
"Hither" is acceptable, but "yon" can make screen displays very messy, with
objects being visibly cut in half etc. I esalise that some limit needs to be
set because of the numerical limitations of z-buffers, and I can't think of
an easy (and cheap) solution. Let's just put it on record that I don't like
it much...

6) Will the cone node disappear?


  • Next message: Avi Singh: "re: Loading .wrl file problem..."
  • Previous message: Len Bullard: "PHIL: MAXT <chunk2>"