Why does science work so well? I claim (following Ian Hacking), it's because of the experimentalist tradition of ever finer and better control of matter. The three pillars of that tradition are tacit knowledge, a conceptual core, and a concrete triad of methods, instruments, and materials. Software craftsmanship seems weak on that triad, so I recommend (1) an open source tool that will carry the conceptual underpinnings with it, and (2) the use of GitHub as a resource akin to the Materials and Methods section of a scientific paper.
Although this is Part 2 of a series, the first part isn't a prerequisite.