How to cite comment by unknown user disproving Erdős conjecture?

I believe that fedja has disproved a conjecture of Erdős et al. when commenting on my question Almost monochromatic point sets, but I'm unable to contact fedja regarding this. What shall I do? Shall I just casually drop a footnote in my upcoming paper that an unknown user has solved an Erdős problem?

• Some related posts - although they seem to be more about citations for posts rather than comments: How does one cite a MO post with an anonymous author? Also this post on Academia: How to acknowledge a MathOverflow user? – Martin Sleziak Sep 25 at 20:22
• My impression is that while fedja is in no hurry to link his MO account to his other public details, he is not making any attempt at anonymity. I can't remember if he's ever confirmed his identity but I (and I think many others) have a fairly safe guess as to his identity – Yemon Choi Sep 25 at 21:53
• @GerhardPaseman --- it might be better to remove the comment that discloses the hidden identity... – Carlo Beenakker Sep 26 at 9:34
• Poor Fedja, his/her comment closed a conjecture by Erdös and has only one upvote. It's hard to get recognition... – Federico Poloni Sep 26 at 12:51
• @Federico It was 0 before I've asked this question, but by now it's already 7! – domotorp Sep 26 at 18:59
• I think fedja is probably fine being cited anonymously but if not he will presumably be able to see your response and divulge his identity. However you should certainly be careful to preserve the comment before citing it, as discussed in the comments to Carlo's answer. – Will Sawin Sep 26 at 23:14
• @WillSawin Domotorp's comment dates from August and Fedja has been on the site since then, so he might not have seen the comment; he isn't usually the sort of person to duck a straightforward question ("how should I cite your answer") by ignoring it – Yemon Choi Sep 27 at 12:46
• @YemonChoi In that case "I'm unable to contact fedja" still seems a little extreme - if one were to make another comment there's a good chance he would see it and a small chance he'd be annoyed very much. – Will Sawin Sep 27 at 13:03
• So I have deleted my earlier comment, however there is other information suggesting the name of the entity operating then account fedja. Gerhard "He Is Not Very Anonymous" Paseman, 2019.09.27. – Gerhard Paseman Sep 27 at 16:56
• If you google "fedja math", there is one very clear candidate who appears at the top of the results. Everything on his website seems to match fedja's profile. You might try emailing him and checking if he is the one. – Kim Sep 29 at 10:46

This would not be a first, see arXiv:1909.00177

The citation (which I think follows the best practice) is

[6] Fedja, Communication on the MathOverflow website (2018), available online at https://mathoverflow.net/questions/125861

• Since it is available, citing the edit history of a post is a little better. That way scholars have access to the version that inspired the citation. Gerhard "Making Digital Archaeology Somewhat Easier" Paseman, 2019.09.25. – Gerhard Paseman Sep 25 at 20:31
• And putting that version of the post into the Internet Archive would be a good idea, so it has a permanent-ish url. – David Roberts Sep 26 at 9:09
• I will just add to David Robert's comment that sometimes it might be important that all comments and all answers are shown (if there are many answers or comments). Such version of a question can be obtained using StackPrinter, see this answer. In this case, you can compare the snapshots of the usual version and the StackPrinter version. – Martin Sleziak Sep 26 at 15:28
• Definitely doing one of those two things in this case is a good idea because comment history, unlike edit history, is not automatically preserved. – Will Sawin Sep 26 at 22:56
• @WillSawin ...and, just to remind this fact to those reading here, this is one of the reasons why it is a bad idea to put answers in comments. – Federico Poloni Sep 28 at 9:45
• why not copy the comments in an answer, community wiki, clearly indicating who is the author? – Carlo Beenakker Sep 28 at 10:56
• @WillSawin but many of the answers are in fact, wholly, or partly, inspired by comments on MO, I suppose – vidyarthi Oct 1 at 8:43
• User Fedja is not "anonymous". He is a well known MO user, and he has a name, namely "User Fedja". – Alexandre Eremenko Oct 8 at 15:27
• It wouldn't be a second either: (0,1)-matrix congruence: is it known?. – Rebecca J. Stones yesterday