All of these are generalizations with a kernel of truth. In real life, I imagine the frequency of occurrence of various hiring techniques falls into something

All of these are generalizations with a kernel of truth. In real life, I imagine the frequency of occurrence of various hiring techniques falls into something like a normal distribution. Similarly with hiring results. My thoughts: 1. It's perceived by some to be a maximally objective and efficacious measure of skill and experience. This might be true in principle, but in practice, reading code carefully takes a tremendous amount of time and making judgements about someone's work based on a commit log probably doesn't happen much. My intuition is that this gets more lip service than anything. I imagine employers are looking for lots of green boxes and lots of commits, if anything. 2. All the time. The virtues of open source software often outweigh the perceived loss of revenue and control. 3. The interviewers just don't have the time. They are too busy going to design meetings. Or doing dozens of interviews with a minimum of prep or decision time.