CompuServe Messages

Fred, >… you could build ANYTHING out of triangles, so it may be better to work on an engine thet ONLY draws fast triangles. Or maybe also fast squares. From my rudimentary knowledge of integral calculus, the statement about triangles is true. This is a method used to calculate area and volume of odd shapes. As to the use os squares, that doesn't really sound practical to me. I could be wrong though, I frequently am! <G> Seriously, you could use rectangles, and make them sufficiently small, but you would still have an inherant error (actually, +- [(upper riemann sum – lower riemann sum) /2]. > faster than arbitrary-number-of-points polygons? This I guess would depend on the hardware. For instance I know that some SGI models have hardware acceleration for this kind of thing and some do the same thing but emulated in software. overall, i would guess that if would come down to how complex the thing you are trying to represent is. Im thing that with a 'simple' enough model then triangles would be fine, but if the model were 'complex' then any gain you might get in comptation would be offset by the task of calculation and keeping track of so many triangles. Just a thought.