Github for Science

Blog

My friend and former labmate Marcio von Muhlen recently wrote a thought-provoking piece on why we need a Github of science. My take on his central argument: our centuries-old system of for-profit academic journals and peer review could be vastly improved if it included aspects of modern Open Source software publishing tools like Github. For example, instead of relying on the opinion of two or three anonymous (and possibly unqualified) referees to determine whether my research belongs in a high-impact journal, I could post my paper on the “Github of science” and the entire community of my peers could weigh in on its strengths and weaknesses. Like quality hits in a Google search, well-regarded research rises to the top and is rewarded by additional visibility, and weaker research sinks to the bottom [..].

I’ve been thinking about a backwards approach [..] How horribly broken would Github become if we recast it in the image of Reed-Elsevier, Springer, John Wiley and Sons, and the rest of the for-profit scientific publishing companies?

First, we would need to put some locks on the doors. Access to our new Github is only available to subscribers, and it isn’t cheap. If your school or employer doesn’t pay for access, you’re probably out of luck. If you’re an unemployed coder using Github to learn a trendy new programming language, or a hobbyist coder interested in contributing to a project on Github, you’re not welcome anymore.

Actually, it’s a little more complicated than that. There’s one subscription for Github Ruby, another subscription for Github Python, another subscription for Github Java, and so on [these are all different comp languages, it’s stupid to differentiate between them]. If you have a Github C subscription but you’d like to browse Github C++, get your credit card out.

–

The mere suggestion of it makes me shudder.