Stop Writing Bad Science Code

Between them, the authors have logged many years as developers, collaborators, and supervisors on scientific software projects. Over that span, we've seen a lot of bad science code (some of which we perpetrated ourselves, much to our chagrin). Nobody expects scientists to be star software developers, but the fact is, these days you simply can't do science without being at least a part time software developer. Even a simple data analysis task involves (or should involve!) writing a program to reproduce the analysis on demand. Basic familiarity with some best practices for developing scientific software will make you dramatically more productive as a scientist. In this talk we share some tips for avoiding the main pitfalls that lead to bad science code.