Today we’ve gone much further in terms of utilizing the internet then just sending emails. In fact academics are so swamped by emails that it’s not often the best way to reach them (Melissa Wilson Sayres excepted), because the noise level in their inbox is so overwhelming. That’s why I was confused and a bit disturbed by this sentence from the supplements of the paper I blogged below, “All source code for analyses is available on correspondence with authors.” The sentence is on page 7 if you care. Two things. First, with services like GitHub you should be able to put together a repository of all your code and data pretty easily. Even if it’s a messy tarball, that’s better than nothing. Likely it would save time in terms of not having to respond to disparate emails (and finding the files, etc.). Yes, the one time investment to comment the code and clean it up a bit might be non-trivial, but you only have to do it once. Second, this is a Nature paper. In fact, it was a thorough Nature paper. Why this specific oversight? Perhaps the authors didn’t think it mattered. I happen to think it does.
Not making distribution of code in public repositories upon publication habitual and reflexive slows down science, and keeps it bogged down in the past. In fact, I’d argue that for non-computational papers there should be ways of uploading unfiltered data too. We can do better, and demand better.
Comments are closed.