Zsbán Ambrus wrote: >> I would be happy with a solution to the simpler problem of 2 variables, >> coefficients represented by matrices of fixed size, and multiplication >> in >> the truncated polynomial ring. > > For the latter, could you not use almost the same thing as for one > variables, but replace * in its definition with multiplication of > single variable polynomials, and + with addition of single variable > polynomials? The result would probably be assembled in the right way > by the normal fill rules. I'm not completely sure that would work, > but it's worth a try.
Thanks for the suggestions. I have something along these lines which almost works. I assemble the result using polynomial multiplication of rows, suitably shifted. Selecting the rows should involve Oblique (/.) but I do not have it quite right. > > For multivariate polynomials, I've written a library in perl once that > calculates with a sparse representation of them. It's at > "http://www.perlmonks.com/?node_id=297829";. (My coding style was > rather immature back then.) I took a look at your library. The coding style seems fine to me. I cannot decide whether it is a long way or a short way from Perl to J. Best wishes, John ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
