4 years ago Robert Kern defined guide lines or check list for the review of the functions in scipy.stats http://projects.scipy.org/scipy/wiki/StatisticsReview This is a useful checklist for evaluating legacy functions,... in scipy.stats. And I think I (implicitly) followed this most of the time in my stats cleanup. But the criteria should not apply to only existing functions (i.e. that have entered trunk), but also to new code. The only point I want to strengthen is number "1. The function works. Sometimes, you just have to state the obvious." to "1. The function works and produces correct result." "works" sounds too much like "it doesn't raise an exception" "correct" is also a vague term, but it captures more the spirit. The checklist could be reviewed or rephrased, but I would like to have guide lines spelled out more explicitly, so we know what the rules of the game are. (even if they are guidelines) Josef