The Catmull-Clark surface subdivision algorithm is a fun little algorithm: http://rosettacode.org/wiki/Catmull%E2%80%93Clark_subdivision_surface http://rosettacode.org/wiki/Catmull%E2%80%93Clark_subdivision_surface#J
But there are other approaches to this task, and some of them require very different sorts of abstractions. But testing the result - especially when the result is not exactly equivalent but only visually similar - requires a way of visualizing the results. Basically, you need a decent rendering system (one that gives lighting effects to represent the orientation of surfaces). Meanwhile, I've not been spending much time with all of the possibilities we have here. What are some of the current best 3d rendering systems we have in J? (Bonus question: what does the data structure for a medium-grey or green cube look like, for each such system?) Thanks, -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
