Consider a question of this type: “My task is to prove […]. Here's my attempt: […] Is my approach correct?” I've seen several questions like this being closed as duplicates. To be more precise, being closed as a duplicate of a question of the type “How to prove […]?” In my opinion, that doesn't make sense. It would only be a duplicate if the other question consisted basically of the same attempt of proving the statement.

An example can be seen here. I've re-opened it because, by the reason explained above, it doesn't make sense to me that this is classified as a duplicate. And much less an exact duplicate.

I would like to know what other users think about this issue.

Please note that this question is not a duplicate of this one since that other question is, in fact, not a question. It is a proposal in the sense that questions which consist of verifying a proof should be closed, after the verification of the proof has been dealt with. I made no proposal. Instead, I have expressed the wish of knowing how other users deal with this situation.

  • 2
    $\begingroup$ Previous discussion about this: Proof-verification and duplicates. and Closing as a duplicate if the post contains OP's own proof/solution. $\endgroup$ May 13 '18 at 15:03
  • 1
    $\begingroup$ My two cents. The OP's proof attempt may often be correct (or close enough for Math.SE). Or with apologies to Tolstoy, correct proofs are all correct in the same way; incorrect proofs are potentially more distinctive. $\endgroup$
    – hardmath
    May 15 '18 at 17:21
  • $\begingroup$ Two cents: if a person has a question and proof that they have no reason to think something is wrong, then they ought to post the question as a question and the solution as a solution. Flaws will be worked out in the review of their solution. If the question already exists, it should be a solution on that question. If they have a proof and they're pretty sure there is a mistake, then they don't have a question and a solution, they have a question about their flawed solution. At that point we have a basis for a post and answers that aren't "everything's correct." TLDR they can be duped. $\endgroup$
    – rschwieb
    May 16 '18 at 13:23
  • $\begingroup$ @rschwieb What you wrote is suggestion meant for those who wish to have their proofs examined. My question is basically what to do with those users who don't follow your suggestion. And I am not sure about your idea. Suppose that someone posts this question: “How to prove that $n\times n$ complex matrices with $n$ distinct eigenvalues are dense in the space of all $n\times n$ complex matrices?” followed (as an answer) by a proof. Than it is quite likely that the question will be closed as a duplicate. $\endgroup$ May 16 '18 at 13:31
  • 2
    $\begingroup$ @JoséCarlosSantos The implicit suggestion is that things not following the advice can be duped. As I said, if the solution which needs scrutiny is to a question that already exists, it belongs on that question (all other things equal.) . There is no sense in duplicating the question post and putting a correct solution in the body or in the solution space. It does not help coalesce helpful answers to the question. $\endgroup$
    – rschwieb
    May 16 '18 at 14:34
  • $\begingroup$ @rschwieb Right! I hadn't thought of that. $\endgroup$ May 16 '18 at 15:05

It would only be a duplicate if the other question consisted basically of the same attempt of proving the statement.

This is arguably a valid point of view. However, in my opinion the answer you gave is inconsistent with that view. You do not really engage with the attempt itself, pointing out what exactly was wrong about it and so on.

My view is that you need to decide:

  • either the question is mostly about the attempt, in which case it is not a dupe and your answer is incomplete,
  • or it is rather about getting an actual answer and the attempt is provided as context in which case it is a dupe and your answer is fine.

Put differently, in a way the form of your answer made it a dupe.

I focused on the specific question, but the reasoning typically runs along these lines. Is it about the attempt itself then different attempts are different questions and should receive different answers engaging with the attempt. Is it about the underlying question then different attempts can still give the same question (except if they are so fundamentally different that the change of context gives a different question). In case of doubt one should ask OP what they intended, both are admissible but different.

  • 1
    $\begingroup$ I agree with your comment about my answer: I did not analyse the specific attempt by the OP to prove what was supposed to be proved. $\endgroup$ May 13 '18 at 14:52

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