| > We looked at: > The number of active repositories, a useful proxy for the projects that people are working on right now. This is better than looking at all of them like many of these stats do because it eliminates has-been languages. > The total number of pushes per language as well as the average number of pushes per repository. These metrics are indicators of the rate of innovation occurring with projects being written in a particular language. This is skewed towards developers who push small changes often, and even includes code shufflers, and ignores projects where the developer commits large samples at a time. > New forks and open issues per repository, which also show active use and innovation. Well written code has few issues raised. > New watchers per repository, an indicator of developer interest. This can be cheated by language promoters who run campaigns on twitter to double the Github watchers within a week. This stuff happens. |