Someone noted in ticket 1346 that the fpLLL doctests use random data, and said that we should do tests with fixed data which return a known result.
I don't agree with this. There is no reason why LLL should return the same result from implementation to implementation. fpLLL may well change the lattices returned and this would break any fixed doctests, but would not necessarily constitute a bug in fpLLL. Instead, the tests should generate random matrices of various kinds using the programs provided with fpLLL for this purpose, then is should reduce the lattices using the SAGE wrapping of fpLLL. Then it should test that the lattices really have been reduced, using whatever test you like. The one distributed with fpLLL for this purpose should be sufficient, though one written in SAGE to directly test the Lovasz condition or some such thing would be better. I don't know how the doctests work at the moment, but I don't think the defect as reported really is a defect. Bill. --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com 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://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
