I guess the point of all this is that there is nothing inherent about J that prevents the efficient implementation of depth first search.
470 seconds - simple minded DFS 12 seconds - one single observation about dead-ends 1.5 seconds - reversing selection criteria 0.9 seconds - further refinement 0.003 seconds - Warnsdorff. Real insight! ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
