Everybody in academia receives email spam, much of it from China, inviting us to conferences that, if they exist, we probably don’t want to go to. One ignores it, and hits “delete”. Late last year I was about to do that when I realised that on this occasion, it was a genuine email from a Chinese PhD student who was using some code that I had published, and asking for help with it. Great!
I couldn’t immediately understand why the code wasn’t working for this student, so I asked him to send his files to me… and I found a bug in my code. With a growing sense of dread, I realised that this bug would have affected results that were already published. I notified my co-authors on the paper and in haste, on evenings and weekends (because this was not my full-time job at the time), I corrected the code, reran a hydrodyanamic model, redid the analysis… and found to my relief that the overall conclusions of the paper were not affected and so there would be no need for a retraction. It did, however, feel as though the record should be corrected, so I got in touch with the editor, and… long story short, a “corrigendum” to the original paper appeared online today. For those without a subscription to Ocean & Coastal Management, I’m also hosting it here for now.
I was already an advocate for good software development practices in science, in order to reduce the liklihood of exactly this sort of thing, and now my feeling on this matter is strengthened…
If you have been using my code for inserting tidal turbines into Delft3D you should make sure that you are using the current version, which had this error fixed in December 2017.