On 7/13/06, John Randall <[EMAIL PROTECTED]> wrote:
Maybe this is solving the problem in too much generality to be efficient.
Postscript gets by with just two operations for this purpose: swap (which interchanges the top two elements on the stack) and roll (which does a cyclic shift of the top n elements). As you know, a transposition and a cycle of length n will generate the permutation group on n letters.
Yes but wouldn't that just be putting the accounting back in the users hands? If you could just say C. has special code and be done, with it then it is a much more usable facility, and much easier to explicate. Certainly Roger can decompose the problem as you suggest. But better him once, than every user in the future. i am doing simulations on a (possibly high dimensional) game board and i need to systematically test a variety of neighborhood swap operations. my code would be a lot uglier if i had to accommodate a partial implementation of C. special coding. ~greg ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
