Hi Bjarke! On 28 Feb., 17:18, Bjarke Hammersholt Roune <bjarke.ro...@gmail.com> wrote: > Is it possible to write a test that runs over all polynomial ring over > a field in Sage? This is as opposed to a test that works in a specific > ring that the test constructs. Obviously there are infinitely many > possible rings that can be constructed, so the test should run over > some representative finite subset.
A test requires some computation that has an expected output: The computation is done, and it is checked whether the output really is as expected. I see several possibilities for what you describe: 1. Choose some representative polynomial rings, ideals,..., and compute the expected result. Then, in your tests, the input is fixed, the expected output is known, so, it is easily tested. 2. Construct a ring/ideal/... randomly. Then, of course, you don't know the output a priori. (a) Mark the test as "random". Then, the output of the random computation is not checked. (b) Do consistency tests. E.g., if you know that two different computations on the same input will always have the same result, then test the equality. Example: sage: R = PolynomialRing(QQ,'x',randint(1,5)) sage: I = R*[t for t in R.gens() if randint(0,1)] sage: len((I^2).gens())==len(I.gens())^2 True So, you have a somewhat random example, and test a property that should always hold. The disadvantage of method 1. is that you need to fix the input data. The advantage is that you have full control on the output. The disadvantage of 2.(a) is that this only tests whether there is an error. The disadvantage of 2.(b) is that it is more difficult to read for a user -- mind that the test examples in the sage library are part of the *documentation*. 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
