After giving this some thought, I realized that traditional permutation generators are going to have a problem here. Consider the following four permutations:
Those can be different permutations. The second could have the first and second ones swapped, the third could have the last fourth and fifth zeroes swapped, etc. Even though, to us, they look the same, the permutators that I've played with don't see that (and how would they calculate it?).
And it was an interesting coincidence to see the Knuth thread on /. this morning. I'm checking it out now.
Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.