Do you remember reading that Microsoft and Facebook had more contributors than any other organization on GitHub? That stat has limited value because a significant amount of development works happens outside of an employer’s GitHub organization accounts. For example, Red Hat employees represent a significant number of contributors across a wide range of cloud and container-related projects, but may not get recognized for this because it occurs in non-corporate organizations. Furthermore, it is common for a company to create a separate organization for popular projects. Thus, using this methodology Google does not get recognized for its Angular project, but Facebook gets to bask in React’s glow.
Another way to evaluate GitHub organizations is based on their activity. Open Hub data indicates that 61 percent of the most active organizations on GitHub organizations are commercial enterprises. Most of these companies are working on projects where almost all of the contributors are also employees. Non-profit organizations like those supporting Linux and Kubernetes on average have the highest number of commits. Education organizations have the fewest because many of the projects they maintain are just ways to manage syllabi and homework assignments.
Looking forward, we want to evaluate corporate open source software based on the percentage of a developer’s time that goes towards corporate controlled open source software projects, other open source efforts as well as proprietary endeavors. How would you measure the open involvement without using crass vanity metrics?
Check back each Saturday for a new “This Week in Numbers” column from Lawrence Hecht.