About thanking reviewers

One often complains about bad quality reviews, and the fact that there is not a lot one can do to avoid these. Today, I’d like to mention the other problem: there is no good way to thank reviewers.

Of course it is a bit strange when you want to thank the judge: they just do their job, and it looks suspicious to try to thank them. But reviewers are not judges: two reviews with the same decision can be very different, one just saying “yes, ok, interesting” and the other highlighting missing pieces of the story, giving meaningful pointers to references (without saying “please cite my papers”), listing typos, suggesting simpler proofs, etc. And sometimes a negative review is worth much more than a positive one. See this tweet.

One could argue that since we work on both sides, as authors and reviewers, it is kind of fair: we don’t receive “thank you”, but we don’t send any, so it’s balanced. I don’t think so. I think any piece of work should be acknowledge, because it feels good, and it is important to feel good to do a good job. I think part of the reason why there are bad quality reviews is that there is no feedback loop. Even a positive-only feedback loop is efficient, I believe. Without it, it is difficult to not drift slowly in the bad direction.

So, how to thank reviewers? You can mention them in the acknowledgment paragraph, but this is so standard that it does not mean much and most probably the reviewer is never going to read these lines. Sometimes for journal reviews the editor is sending some kind words, which is nice. Sometimes the PC member for whom you do the subreview says thanks, and this is also nice (I try not to forget doing this when I am in a PC, but I have already forgotten a few times). In some cases there are rewards for good reviews (best reviewer awards etc.) but it means competition on yet another aspect of research.

During my few years in research I’ve received many high quality reviews. And I have forgotten most of them. But I’d like to take this occasion to thank the reviewers of two papers:

• Graph classes and forbidden patterns on three vertices with Michel Habib, that has a very long (and quite boring but essential to the paper) case analysis, that a reviewer has checked line by line, and even found a few mistakes and ways to patch them.

• Local Certification of Graphs with Bounded Genus with Pierre Fraigniaud, Ivan Rapaport, Éric Rémila, Pedro Montealegre, Ioan Todinca, that needed a lot of notations, technical lemmas etc. and was generalized and simplified before we could publish it. A reviewer pushed us to try to simplify the approach, or at least explain why simpler approaches could not work. This was an insightful discussion, that really improved the paper.