This is the kind of thing that would be good for the projected SciPy journal. It might be appropriate for JStatSoft http://www.jstatsoft.org/, although that has traditionally had a bias toward R. fwiw, Alan Isaac