On Dec 6, 11:15 am, Robert Bradshaw <rober...@math.washington.edu> wrote: > On Mon, Dec 6, 2010 at 8:01 AM, David Kirkby <david.kir...@onetel.net> wrote: > > On 4 December 2010 05:32, William Stein <wst...@gmail.com> wrote: > >> On Thu, Dec 2, 2010 at 6:40 PM, David Kirkby <david.kir...@onetel.net> > >> wrote: >[*snip*] > > It's fairly clear in the past that the "Expected" result from a test > > is what someone happened to get on their computer, and they did not > > appear to be aware that the same would not be true of other > > processors. > > Most of the time that's due to floating point irregularities, and then
http://docs.python.org/release/2.6.6/tutorial/floatingpoint.html#representation-error If the "numerical noise" issue in sage testing has been in controversy for so long, why not replace all such failing doctests with a warning (if triggered) promising to convert it to a sensible test (not dependent on floating point order of operations in hardware or said base conversion); and dispense with all the vitriol? (Not referring to any particular persons' vitriol -- I'm an equal opportunity observer of circumlocution.) > there's an even smaller percentage of the time that it's due to an > actual bug that didn't show up in the formerly-used environments. In How does this help your side of the argument? > both of these cases the test, as written, wasn't (IMHO) wrong. Not Why? If the test is non-deterministic, then you can have false- positives and false-negatives. What's a good argument for that if you can avoid it? If there are counter-examples (test case scenarios) that prove you must take a statistical approach, then that would be an entirely different testing framework. > I agree, people of all backgrounds can make significant contributions. http://trac.sagemath.org/sage_trac/ticket/8336 Robert and I had a long off-list discussion on this round() bug. The problem IMHO, is not sticking to an interface; the requested invariant (that the same precision type be returned) is not possible in some cases. In other words, the interface/invariants are wrong, not the test. Speaking in maths terms, if the relation fails the vertical line test (and is therefore not a function); why on earth would you call it a function? -- 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