Reproducible research in computational science is only possible if scientists distribute the computer codes they used to generate their published results or archive them in such a way that other researchers can later examine them. The author describes some difficulties in achieving this goal, along with a set of python tools for facilitating reproducible research on finite volume methods.
