(0) How to find out for an MO question at hand whether it already has anything like its canonical meta-thread?

(1) Should, in your opinion, every MO question automatically generate a meta-thread of its own, reachable by a "meta"-button on its non-meta parent (and vice-versa, the parent reachable via a "parent" button on the meta-thread)? Has this meta-meta question been discussed on meta already?

EDIT: no, there should not be any automatism.


  • This is very likely to have been asked on MO.meta before, but I did not find it, despite a (cursory) search.

  • One of my motivations for asking this is that on several occasion I have encoutered comments on MO about which I think something along the lines of " one should get someone to correct this comment."

Then, one finds oneself at a loss what to routinely do: comment oneself (hoping for the pinging to work), flag (which most of the time seems excessive-one should flag sparingly), or turn to meta. But where on MO.meta to turn to for a particular question at hand?

  • Of course, my question (0) rests on an implicit premise, which is questionable, namely that there is any such thing as the canonical meta-thread to a question if it has one.
  • 6
    $\begingroup$ Every question automatically generating a meta question sounds like a terrible idea. It sounds like a surefire recipe for clogging the meta from being actually useful. I'm not even sure why this would be even useful enough to merit being sure such a proposal was discussed before. $\endgroup$ – Asaf Karagila Jul 17 '17 at 9:45
  • $\begingroup$ @AsafKaragila: thanks for the response. Now, especially since you summarize it so nicely, it clearly seems a thing to avoid. Re question (0): one apect of question (0) is: there obvioiusly is no problem if the MO thread somewhere says something clickable like "this thread has grown its own meta.descendent [here]()"; but, in situations where (0.0) there is no such comment, (0.1) one would like to discuss a thread on meta, what should one routinely do? In a sense, this question amounts to nothing more than "Is absence of a meta-thread-here-comment sufficient for absence of a meta-thread?" $\endgroup$ – Peter Heinig Jul 17 '17 at 11:02
  • 3
    $\begingroup$ There is not that much activity on meta, and meta posts on specific question are usually posted within a few days of the question (afterwards the issues are already either resolved, or the question is closed). You are overthinking this. $\endgroup$ – Emil Jeřábek Jul 17 '17 at 13:02
  • $\begingroup$ Brief summary of hints kindly given by the commenters and responders: (0) try the serial number of a thread, (1) try searching for the title of the thread, together with "site:meta.mathoverflow.net" in a search engine, (2) when bringing up a thread in a meta.discussion, do not even name the thread, let alone a particular user, rather start by a precise yet general discussion of the matter at hand. $\endgroup$ – Peter Heinig Jul 18 '17 at 4:29
  • 1
    $\begingroup$ If a specific question on main has lead to starting to a discussion on meta, quite often somebody posts link to the met thread in the comments. (I mean, the comments on the main.) $\endgroup$ – Martin Sleziak Jul 18 '17 at 7:11
  • 1
    $\begingroup$ This seems tangential to the question in the title, but could you clarify what you mean by: "one should get someone to correct this comment"? If a comment is rude or should be removed for some other reason, you should flag it. If you mean editing a comment, this can only be done by moderators (if the comment is more than 5 minutes old). $\endgroup$ – Martin Sleziak Jul 18 '17 at 7:14
  • $\begingroup$ @MartinSleziak: "someone" was used in reaction to thinking of the two alternatives (0) author of comment deletes and possibly corrects-and-reposts comment,which is possible even after the time-interval, (1) moderator deletes comment. $\endgroup$ – Peter Heinig Jul 18 '17 at 9:58

If an MO question has spawned a meta thread, chances are the title of the MO question will appear somewhere in the meta thread, so typing that title into your favorite search engine along with site:meta.mathoverflow.net should find it

  • 5
    $\begingroup$ The serial number of the question might be even better to identify the question; there is no way to avoid its occurrence in any link to the question, and, presumably, a meta post about a specific question should at least link to the question. $\endgroup$ – Emil Jeřábek Jul 17 '17 at 15:28

I will add to the suggestions already mentioned in answers/comments posts so far that you might try to look at recent posts on meta tagged . (This tag is used for discussion related to a particular question. If a question is about a general issue, the tag is not used - although even in such cases some questions from the main site might be linked as examples.)

And I definitely agree with Emil Jeřábek's comment that searching for question id seems like a reasonable thing to do - if some post here on meta links to a question of the main, it will contain the question id (whichever format of the url the poster used). So you could search for some thing like url:"*118202*". (Of course, replace 118202 with the id of the question you're interested in.)


MathOverflow (to speak in a fuzzy sense) is for issues in mathematics, while meta.mathoverflow is for issues about the practice of using MathOverflow. As a matter of courtesy, in talking about the latter issues, it is best to pick one issue and describe it in general terms without naming a particular question or user or sets of questions or users. This is especially true when the issue is considered problematic; if the meta post is well written, future occurrences of said issue can include a comment directing one to the meta post. (Example MathOverflow questions where good behaviour is considered can and sometimes should be highlighted, on the other hand.)

So the correspondence, should there be any, should be between issues of practice on MathOverflow and the associated meta question. Things are more fuzzy than that though, as some things have many aspects which (in the current format) take several meta questions to consider.

If you need to search meta therefore, search by issue. It may be good to have more than one name for the issue and insert a relevant comment in the MathOverflow question for the convenience of future readers which takes them to the meta post.

Gerhard "This Has Been An Issue" Paseman, 2017.07.17.

  • $\begingroup$ Thanks for the constructive comments. Taking, for the sake of argument, the downvote seriously (which does not seem fair to me by the way), I feel inclined to delete the question, but there is a double bind: two answers have been given, with hints (e.g. " title into your favorite search engine along with site:meta.mathoverflow.net" and the etiquette of not naming users in a meta discussion, at least not usually and initially, hints which may be of use to others. So, it seems to me, deleting the question would destroy these hints. What would you do? Just leave it at that? $\endgroup$ – Peter Heinig Jul 17 '17 at 18:20
  • $\begingroup$ I would ignore the downvotes (none from me) and place faith in the intent of the question (On meta, downvotes have no effect on reputation and have in my view more correlation with popularity than with intrinsic value.) Take note of the commentary and see about how to modify behaviour for the good of the forum. A nice philosophy is to act as a policeman/mediator to encourage good behaviour rather than a vigilante to prosecute bad behaviour. I think keeping your question is fine. Gerhard "Forum Is Like Raising Children" Paseman, 2017.07.17. $\endgroup$ – Gerhard Paseman Jul 17 '17 at 18:49
  • 1
    $\begingroup$ According to my understanding, downvotes on meta are supposed to signal "I don't like the proposed change" rather than "this was a bad question". So don't worry about them in this case. $\endgroup$ – Hugh Thomas Jul 18 '17 at 21:08

You must log in to answer this question.

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