I did some searching around, but couldn’t find an answer. Sorry if this is in the wrong area.
I was looking at the roaming ralph sample to help me understand collision detection, and on the manual I saw
Caution: this program uses an extremely inefficient method for detecting the height of the terrain. There is a much better way to do it, but we have not had time to correct the sample program.
I was curious, what’s the better way to do it as said in the caution? Sorry if the answer is obvious.
Actually, that comment might be dated. Roaming Ralph used to detect collisions against visible geometry for the terrain, which was horrible, but a couple of years ago Josh reworked it to use collision geometry instead.
I suspect that comment is a leftover warning from the old collide-with-terrain approach, and should be removed now.