On Sat, Sep 22, 2007 at 09:58:33AM -0600, Charles R Harris wrote: > Umm... that doesn't look quite right. Shouldn't it be something like
Puzzling. My implementation works, as far as I could test it, which I did as much as I could. Maybe the two are equivalent. > Algorithm L can be chunked pretty easily also: Yes, I think this would scale even better for really big sets, as the size of the chunk can be more easily controlled. The good think is that with chunks like this the calculation is fully "local", so if the result of the calculation does not depend on the number of points (eg its a sum of the calculation for eache point) then the calculation can go on for ever, there is no limit to the number of points other than the patience of the operator. So if the algorythm is right, they can leave it running for two days, and get great results for a paper. Thanks for your input, Gaƫl _______________________________________________ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion