Hi Ian, If your triangulations are limited to 2-D, you could try using DelaunayFast. It is an imperfect divide-and-conquer triangulation algorithm I wrote, for use with large numbers of points, when speed is more important than precision. It may not be accurate enough for your needs, but I suggest giving it a quick look. Also of interest is the Delaunay.improve() method, which uses edge-flipping to bring an imperfect triangulation closer to the optimal one. -Curtis On Tue, 29 Apr 2003, Ian Graham wrote: >> I _would_ like to understand where the faster algorithms fail, however, >> because this is a very small dataset in my world, and I don't need >> precision. I already make sure I don't have identical x,y coordinates, but >> that doesn't seem to be enough, and I thought only the Clarkson algorithm >> rounds to integers.