Hi Martin! As you have pointed out in the wrong thread, having a smaller ring *has* advantages.
But the more I think about it, the more I find it stupid that I let any element of an infinite polynomial "sparse" ring have its own underlying finite polynomial ring. It should be better to have the following model: * The infinite polynomial ring, R, has an underlying polynomial ring R._P, that may change during computations. It has to contain all variables that were created during computations, but not (much) more than that. So, if a generator x of R creates a new variable x[n], then R._P will be updated, and if a permutation acts on an element, R._P will be updated as well. * Any element, t, of R has an underlying polynomial, t._p. At the time when t is created, t._p belongs to R._P. * Later, the underlying ring of R may change. As soon as t is involved in an arithmetic operation, t._p well be updated so that it belongs to the new R._P. Probably that will always be better than the current sparse implementation, and (when doing overallocations) might even replace the dense implementation. Cheers, Simon -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
