Every time I see a question asking to verify a proof (usually some exercise), the tag is just not there. I often edit the question to add it but... does it really make sense to keep it?

Since every proof is about some topic in maths, this topic would already be covered by another tag. I can't think of a question who could have the unique tag "proof-verification".

Moreover, who is ever going to search questions by the "proof-verification" tag?

I really don't see its usefulness. I understand it could be a way for some users to blacklist those questions, but in my experience almost nobody ever uses that tag.

  • $\begingroup$ I'll add that the possibility of completely abandoning solution verification questions was mentioned here on meta a few times, for example, see the links at the beginning of this answer. If the math.SE indeed reaches consensus to completely abolish this kind of question, then the tag automatically becomes obsolete. $\endgroup$ Nov 24, 2016 at 4:52
  • $\begingroup$ Although it is not explicitly mentioned in this question, (proof-verification) is one of meta-tags - it cannot stand on its own. $\endgroup$ Nov 24, 2016 at 5:35
  • $\begingroup$ Your last paragraph don't quite make sense... there is no way you experience other users blacklisting a tag (And I am one of them). $\endgroup$
    – user99914
    Nov 24, 2016 at 8:58
  • $\begingroup$ @Martin: I'm surprised that you didn't link to this answer. $\endgroup$
    – Asaf Karagila Mod
    Nov 24, 2016 at 9:11
  • $\begingroup$ @Asaf Indirectly, I did. At the beginning of the linked answer, several posts related to this issue are mentioned. One of them is your link. The exact quote is: "...previously discussed here and here. In fact, even a possibility of a separate SE site for this purpose was discussed about two years ago here." $\endgroup$ Nov 24, 2016 at 9:31
  • $\begingroup$ @Martin: Oh, yeah, I meant directly. I forgot about the transitive property! $\endgroup$
    – Asaf Karagila Mod
    Nov 24, 2016 at 9:42
  • $\begingroup$ @John, I meant that most users don't use it anyway, so people end up seeing "proof questions" anyway $\endgroup$ Nov 24, 2016 at 9:44

2 Answers 2


I actually think that the "proof-verification" tag is a virtual necessity, perhaps not for searching purposes, but rather as an alert to users that the asker is not asking for others' proofs, but rather, is asking for users here to assess the quality of the asker's proof, specifically.

In these cases, for example, it is not appropriate to flag as a duplicate of a question asking for a proof, because the asker is making explicit that she/he is interested in feedback on the work they did.

Good answers then, address the asker's proof, point out possible improvements; offer feedback. If after addressing these points, some may go on to offer to fine-tune the asker's proof, or suggest an alternative proof, that's just icing on the cake.

  • $\begingroup$ I also support an "un-synomym-izing" of "solution verification" with "proof verification." But that's for another time. $\endgroup$
    – amWhy
    Nov 23, 2016 at 21:13
  • 1
    $\begingroup$ I agree with this answer, and I think it highlights an important distinction that needs to be made. Many questions posted on MSE have the form 'where did I go wrong' or 'does this make sense', and they then receive answers ignoring that that simply solve the problem the asker was working on in a different way. $\endgroup$ Nov 24, 2016 at 12:13
  • 2
    $\begingroup$ @AntonioVargas But that is not necessarily a bad thing, esp. nowadays that the OP is, alas, forced to try some method even if they are completely lost (to avoid rapid closure). The OP's attempt may be very far from optimal, may have fatal errors, or may lead nowhere. In such cases they OP may learn much from being steered in a more fruitful directions. And doing so makes the answer more useful for other readers (who rarely are stuck at exactly the same spot). $\endgroup$ Nov 24, 2016 at 17:33
  • $\begingroup$ @BillDubuque, sure, I agree. On a case-by-case basis we should indeed try to determine if we should answer the question OP asked or they question they didn't know they wanted to ask. $\endgroup$ Nov 24, 2016 at 18:34
  • $\begingroup$ I agree with you, @Antonio. I've never discouraged any answerer from, after addressing the OP's proof, adds a link to another proof, or provides one theirselvelf. I don't think it's an either-or situation. $\endgroup$
    – amWhy
    Nov 24, 2016 at 18:35
  • $\begingroup$ I do think that alternative proofs, are well-manicured proofs are useful...I simply believe that if a question is looking for "proof-verification" (whether through a tag, or in the actual posted question), we need to first answer that. In addition, it'd be great to add a link to a similar post, and/or to write your own approach. But this is NOT ABOUT trying to show off one's brilliance; first and foremost: it's about helping an OP. $\endgroup$
    – amWhy
    Nov 24, 2016 at 19:12

How about no? Without all the campaigning, I would offer that its mere existence implies that math.SO is a venue for hobbyists to dump long proofs or partial proofs on the community and expect that they be reviewed for correctness or completed by an at large entity, which... really isn't the purpose of SE. Or at least that's what's been communicated to me, either implicitly or explicitly through flags and other moderation.

Previous history aside, math.SE and MO are sites for commentary on existing reviewed literature rather than sites for proposing new theory, based on the experience gradient and the fact that all users may not be "at their peak" when asking questions or producing answers and commentary.

As far as the "asker's proof" part? Perhaps the proof they are suggesting already exists in a more rudimentary form and is thereby obsolete through Occham's razor, unless the proof they offer is in a "shorter" form, in the sense of "code golf".

Further, if a user really needs to ask, then perhaps they need to review both the vocabulary they employ and the equations implied by their theory.


You must log in to answer this question.

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