23
$\begingroup$

I noticed that many users of MSE (especially newcomers) have the tendency to propose deeply bugged question under the semantical point of view, like:

  • How to compute the numerical value of this (clearly divergent) integral?
  • What is a universal technique for computing the primitive of (take a random function that depends on at least three different parameters, one of them being completely useless, removable through a change of variable)?
  • Given two sides of a triangle, how to compute the length of the third one?

and so on. Usually, such questions are soon commented and the OP is made aware that his question has no sense at all, in a polite way. Nothing wrong about that. However, if not closed, such questions continue to populate the MSE database being unanswered, and being essentially useless to other users. Moreover, among the reasons for closing a question, there is no polite equivalent of this question makes no sense at all.

So my question is: should we introduce it?

And, in any case, how to deal with/get rid of such issue?

$\endgroup$
  • 23
    $\begingroup$ Let's call the new close option Zen questions: What is the sound of one hand clapping?, and suggest that they be moved to Buddhism.SE :-$)$ $\endgroup$ – Lucian Sep 13 '15 at 15:52
  • 3
    $\begingroup$ @Lucian: I love and I have always loved your irony, but I think this issue is quite substantial, I do not like MSE to be "flooded by garbage" (sorry in advance if I sound rude). $\endgroup$ – Jack D'Aurizio Sep 13 '15 at 15:59
  • 5
    $\begingroup$ @Lucian it sounds like this $\endgroup$ – achille hui Sep 13 '15 at 16:17
  • 11
    $\begingroup$ I sometimes use «unclear what you're asking» [for such questions] $\endgroup$ – Grigory M Sep 13 '15 at 16:27
  • 2
    $\begingroup$ @GrigoryM: understandable, but assume that a question was: what is the numerical value of the improper Riemann integral $\int_{0}^{+\infty}e^{x}\,dx$? The question is perfectly clear, but still makes no sense, so unclear what you're asking is not a good match. $\endgroup$ – Jack D'Aurizio Sep 13 '15 at 17:18
  • 2
    $\begingroup$ I think the volume of this stream of questions is sufficiently small that a custom close reason should be adequate. The unofficial SE AutoReviewComments extension supports snippets for that box as well, so you could use it to prevent having to write the same thing over and over, without adding yet another close option to the dialogue. $\endgroup$ – Lord_Farin Sep 13 '15 at 18:12
  • 7
    $\begingroup$ I wouldn't close any of the example questions you gave. They make perfect sense. The answer to the first one is "it diverges", and the answer to the third is "it depends on the angles", along with some examples to illustrate. $\endgroup$ – Jack M Sep 17 '15 at 8:33
  • $\begingroup$ Your examples cannot even begin to compare with math.stackexchange.com/q/734583 and math.stackexchange.com/q/1433736 and math.stackexchange.com/q/790578. @JackM: These are the kind of questions that make perfectly no sense because the asker has a long history of attempting to pretend to know a lot but yet not putting any effort in understanding mathematics. Honestly, Lucian's analogy works well; these are like asking "How can we prove that one hand can clap without moving?" $\endgroup$ – user21820 Sep 27 '15 at 9:03
  • $\begingroup$ @user21820 If the question was "Can one hand clap without moving?", would it be a better question? If the asker asks to prove something false, then the correct answer is to point out that it's false and explain why. $\endgroup$ – Jack M Sep 27 '15 at 9:37
  • $\begingroup$ @JackM: Yes indeed exactly why I point out the existence of really unambiguously nonsensical questions. "Can one hand clap without moving" is either an honest question that deserves and honest answer, or what some people call trolling, but still arguably answerable. $\endgroup$ – user21820 Sep 27 '15 at 13:12
29
$\begingroup$

This is similar to the case of questions asking to prove false claims, which has been previously discussed on meta. In line with that, I would suggest, if such a question is otherwise well written, one should post an answer explaining why the question cannot be mathematically solved. For instance, one could show that the integral diverges or that distinct triangles can share two side lengths. This still serves the purpose of our site, since it provides a mathematical answer to a natural question. That is, answering is preferable to commenting & closing.

$\endgroup$
  • 2
    $\begingroup$ "natural" is kind of an abused word nowadays; the naturality of something is a highly subjective matter. Do we really want MSE being a repository for question&answers explaing why the sum of two even numbers cannot be odd? I think the overall quality of the site should increase by closing such questions, but I am open to discuss my position. $\endgroup$ – Jack D'Aurizio Sep 13 '15 at 17:16
  • 6
    $\begingroup$ @JackD'Aurizio I think the subjectivity is largely contained the question of: "Does this question provide adequate context?" On one extreme, I can see questions like "How can I prove all Jordan curves have measure $0$?" being genuine and coming after lots of thought & experimentation. On the other extreme, "What are two even numbers who sum is odd?" strikes me as something unlikely to be asked by someone who's tried solving it. In my opinion, the difference would largely be that, in one case, someone has made attempts in the wrong direction and, in the other, someone hasn't tried anything. $\endgroup$ – Milo Brandt Sep 13 '15 at 17:28
  • 1
    $\begingroup$ There is another option, namely there is a typo or other simple oversight. For instance in the integral a minus could have been omitted. Answering in such cases has serious drawbacks; sometimes it could even be seen as somewhat rude as it implies OP could have asked such a question. $\endgroup$ – quid Sep 13 '15 at 20:03
  • $\begingroup$ @quid: it is also true that sometimes to fix the question is almost impossible. For instance, here. If we try to make such integrand function a $L^1(0,1)$ function by changing the numerator into $e^{\arctan(x)}-(x+1)$, we lose the integrability over $(1,+\infty)$. I believe nonsensical questions really exist, and their volume is not so small (I am partially replying to Lord_Farin, here). $\endgroup$ – Jack D'Aurizio Sep 14 '15 at 22:53
  • 2
    $\begingroup$ @JackD'Aurizio my point is that I believe that most cases are covered by: a) the question can be fixed, b) the question is of little value. For example, a "Prove that [some false claim]" can simply be turned into a "I believe one can prove [some false claim]. Is this the case?" $\endgroup$ – quid Sep 17 '15 at 18:30
  • 4
    $\begingroup$ @JackD'Aurizio If a question is on-topic but you dumb, downvote it. Indeed, that's what downvotes are for. $\endgroup$ – PyRulez Sep 18 '15 at 16:37

You must log in to answer this question.

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