Discovery on Github

March 25, 2016

Github today has sort of a monopoly when it comes to developers open sourcing their work. Be it open sourcing side projects or storing your editor/shell configurations, Github is the go to place for most people. Now this is not a post about how important (and big!) Github has become, this post is about how hard it has become to discover great open source projects on Github.

Based on my experience, the usual discovery process starts out in IRC/Slack/Hacker News/Stack Overflow or during conversations. For a developer spending time committing and pushing code on Github, what would help is a way to discover great open source projects on Github itself. In the absence of a real project discovery mechanism, Github has turned into one unnavigable massive ocean of open source projects. The “trending”/“explore” pages by Github are the closest to what project discovery on Github looks like. Now while the explore and trending pages are helpful, they do not have a social element attached to them. As a result, discovering projects today is tied to the number of its Github stars. The de-facto credibility of a project, irrespective of its maintenance status or project commit activity is tied to its number of stars. The impact of Github stars goes so deep into most developers’ thinking today that, at times the decision of choosing to use a well written new library with fewer Github stars against choosing something that has a larger number of Github stars but a dead community, seems to be a rather unintuitive one .

In my opinion, it would be great if Github could recommended, newer projects or contributors to star/follow based on prior interest shown by a user. This would perhaps allow better project discovery in terms of both usage and contribution.