I certainly want to be able to create 0x0 matrices, and would insist that the determinant is 1. So it seems consistent to say that such a matrix is invertible; and then there isn't much choice for the inverse!
John 2009/2/7 Florent Hivert <[email protected]>: > > Dear Martin Albrecht, > > Thanks for your quick answer. > >> > So my question is the following: is there a specific reason why this test >> > has been written ? >> >> I wrote that doctest, because that behaviour is consistent with other >> matrices. >> >> sage: A = random_matrix(GF(127),0,0) >> sage: A^(-1) >> Traceback (most recent call last) >> ... >> ZeroDivisionError: >> >> sage: A = random_matrix(QQ,0,0) >> sage: A^(-1) >> --------------------------------------------------------------------------- >> Traceback (most recent call last) >> ... >> ZeroDivisionError: >> >> >> I have no strong feelings about how things should behave as long as it is >> consistent. > > Neither do I. Actually, As you pointed out, the main problem was more that > this behavior was only tested for Matrix_mod2_dense, and not for the other > type of matrices. Moreover you made me realize something even worse. > For any type of matrices, sage were completely inconsistent since ~a and > a^(-1) did not returns the same result: > sage: A = random_matrix(QQ,0,0) > sage: > sage: A^(-1) > --------------------------------------------------------------------------- > ZeroDivisionError Traceback (most recent call last) > wheras: > sage: ~A > [] > > Basically my patch delegate the choice to ~ to raise an error or not. > >> I suggest, to add doctests to catch these for the various matrix >> types while you're changing the behaviour. > > Ok... I'll try not to forget any kind of matrices. > > I also hope there will be not too much kind where this case is not properly > handled. One community question : suppose that this happen and that I don't > find the correct way to fix it (no one knows about strange data structure :-)) > is this ok to post a patch that test the problem and raise a doc error and to > add a trac ticket for this specific problem ? > > Cheers, > > Florent > > > > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
