Re: vertex ordering for primitive shapes

Anthropohedron ([email protected])
Fri, 20 Oct 1995 20:16:59 -0400 (EDT)


}
} Chris Laurel writes:
} > In this example, the vertexOrdering is set to CLOCKWISE before a
} > Cylinder node is used . . . this is done because the intial camera
} > position is inside tme cylinder. Is the assumption that a spec
} > compliant browser will generate primitives with counterclockwise
} > vertex ordering valid?

Okay, this was a quick hack to get the texture to map on the inside of the
Cylinder in Webspace 1.x on SGI. I have not tried it on any other platform
or with any other browser. If there is a better, more dependable way to
tell a browser to texture-map on the inside of a primitive, please let me
know. In the meantime, this works.

} No.
}
} Cylinders don't *have* vertices, therefore the vertex ordering does not
} apply to them.
}
} In fact, there's no guarantee that all renderers will tesselate the
} cylinder at all; current ones do, but that's strictly a function of
} the rendering technology being used.

This is, indeed, my impression. In actuality, due to Webspace's
annoying/convenient (take your pick of adjectives) leniency in supporting
Inventor nodes I tried the texture-mapping of the QTVR pan on several
levels of cylinder tesselation using the Complexity Inventor node. I found
that the default tesselation does, in fact, affect how the the pan looks.
The pan got some funny bends at low tesselation, but high tesselation
(Complexity of 5) brought a 2-processor Onyx with a VTX board to a crawl
(the pan looked beautiful, though). No tesselation at all (can we have
splines? please, pretty please?) would give a much better pan.

I'll mention that I really think that, at a minimum, primitives (Cone,
Cylinder, and I think Generalized Cylinder was being added) should support
curves. Not knowing a gesat deal about how renderers work, however, I do
not know how hard this would be to add. Really, I would like to see some
kind of curve support in VRML in general, but that may have to wait for
improvements in hardware.

} --
} Bernie Roehl

--Greg


  • Next message: Tom Gaskins: "Re: vertex ordering for primitive shapes"
  • Previous message: Art Yerkes: "Re: Anyone working on Collision Detection yet?"
  • In reply to: Bernie Roehl: "vertex ordering for primitive shapes"