How does the set of "unintersected edges" computed by your algorithm differ from a Delaunay triangulation?

That's a good question. I'd never heard of Delaunay triangulation until your post, so I'm going to need to do some reading. If the answer is that it doesn't or maybe if it's even a better way to show what I'm really after (the idea of remoteness of a set of coordinates... vague enough? ;-) I may just go this route.

I admit to being somewhat of a perl biggot, however; speed in this case is really more important than my love affair with perl. This code will be triggered by a web app submittal of a file with the coords in it. Infact, speed is probally more important than finite accuracy of the stats I'm trying to produce. The end result will be used to relativally diffrenciate sets of coordinates. Or I should say it is my hope that the people who will use these number see them that way. :)

Thanks for the links and the example. I've grabbed qhull and installed it so I can play around in the next few days. I'll update when I learn some more so the next guy that does a search.