13
$\begingroup$

When a user posts a (solution-verification) question on a problem (often a standard problem), I tend to not vote to close as a duplicate, since the user is asking for feedback on their particular solutions.

(Related: this old meta post )

In the recent weeks, there were two exception on my standard behavior: one was a question in which, after feedback was given, another user posted a de nuovo answer. After that, I closed the question as a duplicate using the golden-hammer.

The second was a now-deleted question in which the poster asked for hints for a standard problem (that a semigroup with a right identity and right inverses is a group). I closed as a duplicate, and the poster added an argument for half the problem, tagged as (solution-verification). Another golden-badge user re-opened. But the original poster was still asking for answers to the other half. When I pointed out it was still a duplicate, the user claimed that it was not identical to the previous question, and so could not possibly be a duplicate (as they were only asking about half the problem, not the whole thing). I said that was not the standard, but did not vote to close. The user deleted the post shortly thereafter.

I would like to make the following general policy proposal for (solution-verification) posts:

  1. There should be no immediate votes to close a (solution-verification) post as a duplicate, to allow for suitable feedback.

  2. After a suitable period (two days? three days? some other period), and once feedback has been given, it would be appropriate to vote to close these posts as duplicates.

(In fact, perhaps not only "appropriate", but "desirable"?)

  1. If the solution provided is sufficiently different/of interest, encourage the poster to submit it as an answer to the duplicate question. Otherwise, just closing is appropriate.

Reasoning: I understand not voting to close (solution-verification) posts as an initial matter, so that users will not be turned away before sufficient feedback is given.

But we don't want multiple posts open on the exact same subject, and also when a lot of them will contain only small variations of standard arguments. Hence, the proposal to wait, but eventually close these posts once they have served their purpose.


To highlight/clarify:

What I'm hoping for is consensus that these questions, when they contain correct or nigh-correct answers, be "eventually" closed as duplicates. The precise meaning of "eventually" is perhaps where discussion might be warranted. I'm envisioning it as "not immediately", to (i) encourage/reward the effort; (ii) increase likelihood of engagement; (iii) give the poster the direct feedback they seek. Even a "closed as a duplicate" done pointing to an identical or near-identical argument is likely to be perceived as criticism rather than endorsement/encouragement.

$\endgroup$
44
  • 3
    $\begingroup$ To me, it's at least partially a function of the quality of the proposed answer. In some (many?) cases, the proposed solution is so vague that little is served by trying to patch it up. In those cases, closing the thing as a duplicate seems (to me) to make sense. $\endgroup$
    – lulu
    Oct 22 at 14:42
  • 3
    $\begingroup$ @lulu: I'm actually advocating for (eventual) closure. "Suitable period" may be an elastic standard depending on the quality of the proposal, of course. I guess I'm trying to see if there is consensus about a general site policy for them. $\endgroup$ Oct 22 at 14:44
  • 5
    $\begingroup$ @lulu Mis-tags should be re-tagged, of course. And incoherence can be voted to close. What about reasonable efforts? That's the ones that I'm mainly concern about. I don't want to golden-hammer them too soon, but I don't think it makes sense to just let them be indefinitely. $\endgroup$ Oct 22 at 14:54
  • 7
    $\begingroup$ I will put forward---again---that I think that the best way to deal with unclear questions (including solution-verification questions) is to close them, then give the asker time to correct the problems before folk can provide a large number of irrelevant answers. Such questions should be closed as duplicates until and unless the asker can clearly articulate why their question is really novel. Closure is not punitive---it is meant to either direct an asker to an answer, or to give them time to make improvements. $\endgroup$
    – Xander Henderson Mod
    Oct 22 at 19:15
  • 5
    $\begingroup$ @XanderHenderson: And that I agree with; but how do we deal with clear, and let's say correct solution verification questions? This is what I am talking about, though the comment thread seems to have gone into what to do with those that are mislabeled or unclear while claiming solution verification. I know I brought up one such case, and I may have muddied the waters; it was more along the lines of why I decided to bring this up. $\endgroup$ Oct 22 at 19:20
  • 3
    $\begingroup$ @XanderHenderson E.g., here. Perfectly fine and clear question, received adequate feedback, and does things that are already done elsewhere; certainly lots of posts on the irrationality of $\sqrt{2}$, and even one on the fact that an $n$th root is either an integer or irrational using the same argument. Shouldn't the first one be closed as a duplicate now? $\endgroup$ Oct 22 at 19:22
  • 9
    $\begingroup$ @ArturoMagidin Honestly, I am of the opinion that solution-verification questions are a very poor fit for this site. They tend to be very narrow, and of little long-term use to others, while cluttering the database. I am of the opinion that the tag should be burninated. However, my impression is that there is zero consensus about this in the community here, so I try to ignore the tag as much as possible. $\endgroup$
    – Xander Henderson Mod
    Oct 22 at 19:23
  • 4
    $\begingroup$ @ArturoMagidin There are probably thousands of questions on this site about the irrationality of $\sqrt{2}$, and $\sqrt{3}$ and $\sqrt{2}-\sqrt{7}$ and whatever else, all of which come down to three or four standard arguments. The should probably all be closed as duplicates of one another. $\endgroup$
    – Xander Henderson Mod
    Oct 22 at 19:25
  • 8
    $\begingroup$ At the very least I move that answers to a solution-verification /proof-verification question MUST first and foremost discuss the attempt. Similar errors do recur, and finding duplicates is problematic. In a sense, a duplicate to a failed attempt would be another question where a similar logical error is explained. Possibly in a vastly different context. $\endgroup$ Oct 23 at 8:38
  • 3
    $\begingroup$ @TeresaLisbon: For answers that have some fatal error that can re-occur, an actual answer addressing this would be suitable and closure may be discussed (but I would strongly discourage what happens sometimes, where you post explaining the error, and then you get into a back-and-forth where the poster tries to correct their argument by further editing the post). $\endgroup$ Oct 23 at 19:00
  • 7
    $\begingroup$ @TeresaLisbon I'm not okay with posted answers that consist only of "yes, this works." In fact, I will flag such answers as being "not an answer", or comment saying so. I'm okay with the post receiving only comments saying that (and then would like the community to agree to close them afterwards). $\endgroup$ Oct 23 at 19:22
  • 3
    $\begingroup$ It feels like most of the time, if a user posts something of this form where the question has a solution and they're looking for verification, it's often too open ended for it to be on topic and really it ends up being little more than flame bait that wastes time, frays nerves and detracts from the purpose and reputation of the site at large, because it causes a duplicate/non-duplicate conflict. We're on here to help each other solve problems and not to insult each other's intelligence. $\endgroup$
    – JustKevin
    Oct 26 at 4:43
  • 3
    $\begingroup$ @XanderHenderson I respect your position, but at the same time I think you underestimate the potential for question closures and deletions to come across as punitive, even if the curation process isn't intended to be punitive in itself, and that's why I argue that the tags are perhaps, too blurry in scope and give users the wrong impression about what is allowed and what is not allowed on the site, particularly given the way the knowledge base has grown since the site was founded. $\endgroup$
    – JustKevin
    Oct 26 at 4:51
  • 6
    $\begingroup$ My view is similar to this comment of @Xander. Most SV questions lack sufficent context, i.e. they don't say precisely where the doubt lies (and why). So usually the best choice is to quickly (gold) close as a dupe (lacking context closures are too slow to prevent floods of dupe answers). If OP clarifies then either we can refine dupe links (or answers) if need be, or else reopen the novel good (clarified) question. $\endgroup$ Oct 28 at 16:56
  • 4
    $\begingroup$ @Divide1918 Mathematics is not about performing rote computations to solve specific problems. It is about creating reasoning which allows one to solve many problems at once. There is no reason to have questions about the irrationality of $\sqrt{2}$ and $\sqrt{3}$ and $\sqrt{5}$ and $\sqrt{6}$ and $\sqrt{7}$ and so on when there are one or two arguments which will answers all of those questions at the same time. $\endgroup$
    – Xander Henderson Mod
    Nov 4 at 14:41

You must log in to answer this question.

Browse other questions tagged .