9
$\begingroup$

A questioner asked me about re-asking/re-trying after an attempted proof.

What is the correct advice here? The questioner's proof was somewhat bungled. Should they ask another question, or edit the existing question? I told them to edit the question but I'm not sure whether that is right or not.

Also what should I do with my answer pointing out the proof errors? Presumably the questioner will resolve those issues, making the answer irrelevant.

$\endgroup$
4
  • $\begingroup$ This is too difficult to answer, but in general I'd break it into three cases (1) When the proof is "almost" or completely correct (2) When the proof is genuinely wrong, but the mistake is somewhat natural to expect, even if it looks sloppy (3) The OP genuinely has no clue about what they're doing. If I assign to each post a score in $[1,3]$ on this basis, then for $[1,1.6]$ and $[2.4,3]$, I'd ask the user to not post another question (in the first case they already have a grasp, and in the second they're reaching too far ahead). This one looks like a $2.2$ to me, honestly. However,... $\endgroup$
    – Sarvesh Ravichandran Iyer
    Commented Sep 27, 2022 at 11:12
  • 1
    $\begingroup$ ... it's a good solution verification question, because at least the author points out a particular issue with their answer which can then be checked. So with posts whose score is in the range $[1.6,2.4]$ and where the asker has posted a "good" SV question, I would encourage a new question with a fresh solution. I'm not sure I would encourage the "I will edit my question, please check the new attempt as well" approach : that could make the question too broad. $\endgroup$
    – Sarvesh Ravichandran Iyer
    Commented Sep 27, 2022 at 11:13
  • 5
    $\begingroup$ Just to clarify as the OP of the post in question, I never overwrite my original posts because I like to keep them as future reference, mistakes and all :P I would just add the new attempt underneath the original one (preferably as a collapsible spoiler, so as to better separate it from the original question). $\endgroup$
    – user51462
    Commented Sep 28, 2022 at 2:25
  • 1
    $\begingroup$ First of all, go through the original attempt step-by-step and try to find where the reasoning fails. That can be instructive. Maybe it was a misunderstanding of some theorem. Maybe a basic arithmetic error or an inequality that got switched. It will really depend on the field of math and the level. Undergraduate real analysis? Graduate abstract algebra? $\endgroup$
    – jdods
    Commented Oct 4, 2022 at 16:15

1 Answer 1

Reset to default
13
$\begingroup$

The question is primarily about a verification of a solution. So you did just fine addressing the shortcomings of the OPs attempt (+1 from me). If the OP chooses to make another attempt, and includes that into the question body, it is, in my opinion, also fine, but it does create a problem to future readers.

If I had control of all of it, as the asker I would add a horizontal line, separating the 1st attempt from the 2nd. You would then have the option to either do the same, and review the second attempt. Or, simply add a clause that you address the first attempt only. Doing that would safeguard your answer from possible future criticism. Ideally the future readers would be able to see that without such a clause, but some readers may make mistakes.

The OP may overwrite their first attempt with another one. In a classroom setting that would be fine, but here we disprove of such antics. Exactly because it would void your first answer. A new user may do something like that simply because they didn't think through. Should that happen I would simply add the first attempt back to the top (and politely point out the faux pas).

So there is potential for something to go wrong in the future of that thread. All the solution/proof -verification queries suffer from this as well as other maladies (such as answerers jumping in providing their version of solution without discussing the OPs attempt). That may warrant another discussion. Asaf started one in 2018. Don't know if that is the best place to discuss general policies?

$\endgroup$
1
  • 5
    $\begingroup$ Initially I wanted to make this a comment only, but it became too long, so an answer it is. I'm aware that I'm not really answering everything Suzu Hirose wanted us to discuss, but others can add their points. $\endgroup$
    – Jyrki Lahtonen
    Commented Sep 27, 2022 at 11:15

You must log in to answer this question.

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