A use case scenario: I am reading a paper and I get stuck on some point. It is obvious to me that everyone else who tried to understand this point would get stuck. I ask around, people either haven't seen the paper at all or didn't care about this particular point. I waste a lot of time, find that maybe the statement is wrong as stated, or it can be fixed in some way, or it is correct but I made a mistake.

Anyway, just having a chance to see if anyone else encountered the same problem with the same paper would be very useful to me. So I would like to have a platform to discuss papers.

When I discuss this idea with colleagues, I find that many oppose it, mostly because of fear that it would be abused, i.e. people would push their political agenda, those who don't understand the content, but have enough influence, would be able to discredit the paper, etc. I would like to stress that I want to discuss papers from a purely technical point of view, in exactly the same way as questions on MO are asked and answered. It is not about discussing if a paper is important, or if the proofs are complete. I would like to be able to exchange ideas with other people who tried to read the same paper, in order to answer the questions I have in mind, and answer other's questions. I think MO's philosophy suits this purpose perfectly, why not adapt it better for discussing papers?

  • 13
    $\begingroup$ Because MathOverflow is for questions and answers, not discussion. However, a reasonable twist on this is the following. Ask question Q about specific point P in paper X on the MathOverflow forum. Along with this, mention chatroom C where you are willing to hold a public discussion about X, emphasizing issues S. Then those who care to discuss can join you in the chatroom. Gerhard "Slot A Into Tab B" Paseman, 2017.08.11. $\endgroup$ Commented Aug 11, 2017 at 23:42
  • $\begingroup$ Perhaps "to discuss" was not a right word. What I really mean is questions and answers. I don't like the chat room idea because it means every participant should be there at the same time. What I like about MO is that a question once asked and answered stays there forever, maybe to be read again after 5 years or so. In fact, what I am suggesting is not so different from what MO already provides. Maybe each paper would simply have several of "Ask question Q about specific point P" grouped together. $\endgroup$ Commented Aug 12, 2017 at 13:08
  • 2
    $\begingroup$ If you suspect a gap in a proof, it's probably a good idea to e-mail the author as well. If it's a point left to an exercise, there should be no problem asking it here. $\endgroup$ Commented Aug 12, 2017 at 14:22
  • 2
    $\begingroup$ While many chats are quick, chatrooms on Stackexchange have some longevity, and have archived transcripts, so in principle a delayed discussion can occur. Simultaneous presence is not required. The single thread format seems more appropriate than QA format for your suggestion. Gerhard "Why Not Try It Out?" Paseman, 2017.08.12. $\endgroup$ Commented Aug 12, 2017 at 16:52
  • $\begingroup$ If I understand correctly - MO is supposed to be a virtual "seminar room" - and the kind of questions asked are questions fit to seminar setting. But the question itself is very interesting - maybe allocating some "discussion room" for specific questions \ discussions of this sort ("virtual group reading"?). Some papers really require a lot of processing - I as a reader would have loved to have such a resource. $\endgroup$ Commented Aug 13, 2017 at 6:02
  • $\begingroup$ why not add a label "discussion" to questions that aim at clarifying hard to interpret passages of articles; or maybe even better add a label "clarification" in order not to arouse nervous MO users $\endgroup$ Commented Aug 13, 2017 at 9:44
  • 2
    $\begingroup$ @GerhardPaseman Perhaps in connection with the claim about longevity of chat rooms, we should mention that a chat room is frozen after 7 or 14 days of inactivity. (In fact, even the main MO chat room has been frozen in the past.) It is still true that transcript is archived even for a frozen room. But it is no longer possible to post new messages until a moderator unfreezes the chat room again. $\endgroup$ Commented Aug 17, 2017 at 5:59
  • $\begingroup$ I think the problem is to find number of people who can support such kind of discussion... We all are busy, have different level of knowledge in narrow subject like a a particular paper, and so on an so forth $\endgroup$ Commented Aug 19, 2017 at 12:43

1 Answer 1


If there is a specific mathematical question about a specific point in the paper, then there is no problem with discussing it on MO. This has been established by consensus on numerous occasions. In particular, there does not seem to be a problem with the use case you suggest. We mostly recommend against broader discussion about the paper as a whole or the author's opus (although there is a bit less consensus here), but that doesn't seem to be what you want anyway.

You also seemed to suggest the possibility of assembling all of the mathematical questions about a single paper into some kind of indexed collection. That certainly sounds like an interesting idea, much like an earlier idea of centralizing textbook errata. However, I do not know of a way we can make that convenient with our available infrastructure right now.

  • 6
    $\begingroup$ As you say, certainly asking occasional specific mathematical questions about particular papers on MO is no problem. -- But much more useful would be a searchable and editable database of comments and errata on papers, which would be something entirely different from MO. This could be a separate site or an extension of zbMATH / MathSciNet, or individual publishers could host the comments. $\endgroup$
    – Stefan Kohl Mod
    Commented Aug 12, 2017 at 20:28
  • 3
    $\begingroup$ Maybe adding a tag with the arXiv reference code for all the questions related to errata or filling in proofs in a particular book/paper? $\endgroup$ Commented Aug 13, 2017 at 1:46
  • 1
    $\begingroup$ Kinda seems like a stretch though. $\endgroup$ Commented Aug 13, 2017 at 2:17
  • $\begingroup$ @HarryGindi Tags immediately leaped to my mind, too. It'd be greatly more useful (for search) if the description contained the paper's title and the arXiv abstract: there could be some kind of auto-importer that would fetch and update the description when the tag is created, but that is a bit of a stretch, as you say. $\endgroup$ Commented Aug 13, 2017 at 8:06
  • 2
    $\begingroup$ MO posts that link to arxiv pages leave trackbacks, so if one clicks the "__ blog links" link on an arxiv page, e.g. any of the examples linked here arxiv.org/tb/recent , there is a complete (?) list of MO posts / comments that linked to it. $\endgroup$
    – j.c.
    Commented Aug 14, 2017 at 21:23
  • 3
    $\begingroup$ @StefanKohl I'll add links to some older posts related to the issue raised in your comment: Are there any good websites for hosting discussions of mathematical papers?, Community Peer-Review in Mathoverflow and Is there a good site for holding online discussions of scientific papers?. $\endgroup$ Commented Aug 16, 2017 at 0:04

You must log in to answer this question.

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