Talk about fishing for expert help! Thanks Guido.
It's a pruning algorithm where I strip way pieces that don't meet up at (x,y,z) bridge points. Lots of symmetry about the origin so just pure distance won't work (not unique enough). I think might still get away with a tuple-indexed dict (tallying for pairs, pruning the rest) if I boil my floating points down to fixed decimals -- a chance to play with the new decimal type perhaps. I appreciate the quick lecture on intransitive nature of fuzzy equality and __hash__ vs. __eq__. Helps me think. Kirby > You never said what you wanted to do with these points -- are you > looking to find which points are in fact close enough to be considered > the same point? I suppose you could sort them by the vector length > (sqrt(x**2 + y**2 + z**2), that sort of thing) and then you only have > to compare points that are close together in the sorted list (as long > as the abs values are closer than your epsilon). Or you could divide > 3-space up into cubes with sides equal epsilon, and then you'd only > have to look for points that end up in the same or in adjacent cubes. > > --Guido _______________________________________________ Edu-sig mailing list Edu-sig@python.org http://mail.python.org/mailman/listinfo/edu-sig