A few weeks ago I came across some Academia.SE posts that concerned me. One was a surprised computer scientist who didn’t expect a reviewer to want to see the code their paper was addressing, and the other was related to a bug in closed-source scientific software, producing incorrect results.
In both cases it seemed frightening to me that somehow academic code is often able to escape the radar when it comes to peer review. Jenni A on twitter elegantly summed it up:
It's like submitting a maths paper with no proofs and just saying "yeah, trust me, it's definitely true"— Jenni A (@jensylvia) May 2, 2017
It’s clear that scientific code needs to be open - to the peer reviewers at a very minimum, but open source would be better. There’s a reason Open Source’s mantra is “Linus’s Law” states “given enough eyeballs, all bugs are shallow” (note: the article makes it clear that code review is the important part here).
So, what can we do about it? I’d like to start by finding out some of the motivations for keeping source code “closed” and figure out who we need to inspire to create a culture shift to promote code as an examinable, peer-reviewable part of science. One of the ways I would like to start is by gathering thoughts from academic and open source peers, and running a workshop at MozFest if they accept my proposal.
You can read more about my proposal plans in this GitHub repo. If you’re interested in Open Source or Open Science, I’d love to hear from you if you have ideas, thoughts, or maybe want to join my proposal / help deliver the workshop. To contribute, please tweet me, email me (firstname.lastname@example.org) or open an issue on the github repo.