6
$\begingroup$

Thread

Are questions about solving practical data science problems (memory etc.) on topic?

here on Meta is discussing the interesting problem of questions which lie at the boundary between on-topic and off-topic, because they ask for software, but they can be answered in a more general way, which intersperses the discussion of a statistical method, with some code showing a practical application.

This is an instance of a more general problem: if we see a question where it looks like that the OP is principally interested in an answer which would be off-topic (for example, code only, but it could be off-topic for other reasons), but which can be answered in a way that the OP probably wouldn't like, but which can be very useful for generic users, what do we do? For me it's a difficult problem:

  1. giving the on-topic answer seems "the right thing to do". However, the OP will probably complain and/or not accept it. This may lead to 1) an increase of off-topic questions and 2) an increase of unanswered questions, which I think is bad for what it concerns the site statistics.
  2. just closing the question might be a bit of missed opportunity for the whole community, which would have benefited from such an answer.
  3. write another, better question and answer it myself? Possible, but not completely satisfying.
$\endgroup$
  • 4
    $\begingroup$ Have you contemplated writing comments to ask the OP to clarify their intention? That's the main purpose of the commenting mechanism, by the way. Note, too, that closing a question does not delete it or in any way limit the process of improving it: it's a way to suspend posting answers until the OP has had an opportunity to formulate the question appropriately. $\endgroup$ – whuber♦ Feb 23 '18 at 15:53
  • 1
    $\begingroup$ @whuber you're right, that's the first thing to do and I should have said that in my Q. Suppose the OP doesn't reply, or replies in an ambiguous way (not the other thread's case, but it happened to me sometimes). Suppose also most of the edits to make the question more on-topic are done by someone else, not the OP (this is the other thread's case). We cannot be 100% sure that the OP won't like the on-topic answer, but we have a strong suspicion. Your suggestion is then "close and reopen only if the OP modifies the question so that it's clearly on-topic". Right? $\endgroup$ – DeltaIV Feb 23 '18 at 16:09
  • 5
    $\begingroup$ Yes, that's right. My perception of closing has evolved. Initially closing seemed pejorative: "this is a closed question, meaning we think it's bad or you're stupid." Now I see closing as a service and as a kindness: it staves off downvotes; it prevents answers from appearing that may be irrelevant, confusing, or even wrong (not because they are inherently incorrect but because they address some issue that wasn't asked); and it provides the OP some time in which to reflect on, discuss, and refine their question before it's subjected to any more criticism or voting. $\endgroup$ – whuber♦ Feb 23 '18 at 16:15
  • $\begingroup$ @whuber it's an interesting perspective, which I never thought of. I'm not sure I can see it as a kindness, but I do see your point in it being a service, and giving the OP time to think. After all, if s/he doesn't want to invest some time to reflect on, and refine her/his question, it's probably best left closed. $\endgroup$ – DeltaIV Feb 23 '18 at 16:21
  • 5
    $\begingroup$ It's a kindness in that sometimes downvotes accumulate quickly on unclear or ill-posed questions. Those can only be perceived by the OP as negative. Moreover, many downvotes are never reversed even after a post is edited. I would like to prevent such downvotes, if I can, in circumstances where there is a prospect for the question to be turned into a good one. The message I intend is "Welcome to our site; we see you have taken the effort to formulate a question for us, but it currently doesn't fit our format; let us work with you to get it to the point where we can offer useful answers." $\endgroup$ – whuber♦ Feb 23 '18 at 16:26
  • $\begingroup$ @whuber got it, it makes sense. Thanks! $\endgroup$ – DeltaIV Feb 23 '18 at 16:39

You must log in to answer this question.

Browse other questions tagged .