If the program is not that important your advice is good. But if other teams rely on it, his manager considers it important (for whatever reason), or it's business critical, then unless there is a good test suite/environment he needs to learn enough of the program to be able to tell if what appears to be a minor change really is. Seems to be pretty weak advice given that the odds of there being a good test suite/environment are likely approaching negative infinity.... Update: Maybe better advice would writing a test suite for the program and then try and refine it to "good" status. Though that would still not help with estimating how much effort is needed to make changes to the code.