Is it fine to re-ask open questions from ones paper on mathoverflow?

On the one hand, I am interested in finding an answer to the question, and it might help if it is more publicly available. But if I ask such a question here, I would for completeness sake also mention the paper in which it was asked. But then this gets really close to self-promotion, which is rightly frowned upon.

  • 9
    $\begingroup$ I guess it depends on whether the answer is likely to be an MO answer, or another full paper, in your best guess. If an open question is reasonably estimated to requires months of research and 20 pages of typed maths to answer, it's not really an MO question. It might be that the question seems reasonable, and should have a fairly quick answer, but then turns out to be harder than first thought; in that case no one can fault the asker. $\endgroup$
    – David Roberts Mod
    Commented Mar 13 at 12:45
  • 3
    $\begingroup$ Well counterexamples or references are somewhat easier to write down then proofs. But this is true for any question asked. My point was more that this touches the self-promotion part. $\endgroup$ Commented Mar 13 at 13:01
  • 21
    $\begingroup$ I think it's 100% fine to post a question from your paper (at least, if the question is otherwise appropriate for MO). $\endgroup$ Commented Mar 13 at 13:13
  • $\begingroup$ @DavidRoberts "If an open question is reasonably estimated to require months of research and 20 pages of typed maths to answer, it's not really an MO question" How can you possibly determine that until you have solved the problem? $\endgroup$
    – fedja
    Commented Mar 27 at 2:17
  • $\begingroup$ @fedja I mean a good-faith estimation. If I asked something that I knew would amount to a full research paper, then I'm really not asking a good question. I can't ask a colleague in the tea room to supply a quick answer to a question when I think it's going to really need a six-month collaboration. But if I was after some small (as I thought) step that would go into the longer project, but which turned out to be a hard question, then that was at least reasonable to ask. $\endgroup$
    – David Roberts Mod
    Commented Mar 27 at 2:50
  • $\begingroup$ @DavidRoberts I'm not convinced. I hate self-promotion, but isn't mathoverflow.net/questions/461631 a counterexample to your first claim? It generated quite a lively discussion with a few good people participating and sort of happy ending in less than a month, and I see no reason why that experience cannot be repeated. :-) But I sort of see what you mean. I just claim that "good-faith estimation" is utterly unreliable in both directions... $\endgroup$
    – fedja
    Commented Mar 27 at 2:59
  • $\begingroup$ @fedja it's subjective of course. That question is not what I meant. I mean more like the type of question that will generate a 50 page paper in arithmetic geometry, not a fun question about elementary objects that is nontrivial, but not to the point of months of deep collaboration. I totally think the question you linked is a great example of the use of MO. I'm just trying to put my finger on what is clearly not good for MO, and happy to err on the side of generosity. $\endgroup$
    – David Roberts Mod
    Commented Mar 27 at 3:04

1 Answer 1


When I answered Acyclic matching property in $\mathbb{Z}/p\mathbb{Z}$ I didn't realise that it was an open question from OP's doctoral thesis, although I did figure that out afterwards when I paid a bit more attention to the names in the references. We're now hoping to get the result published.

If it's an interesting question and not just a personal hobby-horse, I don't see any issue. Add a disclaimer if you want, but there's a big difference between "Look at my paper: isn't it great?" and "I've managed to get this far but I think it's possible to go further: can anyone see how?"

  • 3
    $\begingroup$ This could also go on the "success stories" thread $\endgroup$ Commented Mar 25 at 21:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .