12
$\begingroup$

Background

We get a lot of posts asking several related but independent questions [1]. Often the posts are even formatted as enumerated lists, making it clear that the author knows that he/she is asking more than one question. I believe that the quality of the question/answer system increases if each post contains a single focused question. It's my understanding that other users generally agree with this.

Question

The site tour does not particularly emphasize the importance of focusing on one question per post. Is there a way for us to make it more obvious to new users that posts should focus on a single question?

Notes

  1. Regardless of the specific rules of the site, I would point out that asking more than one question at a time frequently indicates that further effort by the asker to identify what in his/her own mind is actually unclear would greatly improve the question/answer process.

  2. The site tour is somewhat saturated with fancy animations and has a low amount of information per screen space. This may be related to the main focus of this question.

[1]: Here's a recent example. Note that in the original version the list contained six items.

$\endgroup$
  • 9
    $\begingroup$ I asked the question you've cited. I'm relatively new to SE so I apologize for any lapses on my part. However, why do you (or the community as a whole) think questions of the kind I asked should be avoided? Where I'm coming from: I don't want to post them as separate questions because that would involve me restating the setup 6 times for several closely linked questions, and an answerer who might be able to deal with them comprehensively would be inconvenienced. Alternatively, I could have left it at "Explain the BCS Hamiltonian", but then the answerer wouldn't know what I didn't understand. $\endgroup$ – ApproximatelyTrue Jun 1 '15 at 12:33
  • $\begingroup$ The help center page on questions to avoid emphasizes avoiding subjective questions, and says that I asked the question you've cited. The page in the help center about questions to avoid says, "Your questions should be reasonably scoped. If you can imagine an entire book that answers your question, you’re asking too much." Do you think my questions scope was that broad? A previous question on the meta deals with the same issue but provides no explanation for the policy. $\endgroup$ – ApproximatelyTrue Jun 1 '15 at 13:05
  • 2
    $\begingroup$ If the questions are linked in the same area and if the answer to the first one leads directly to the next, I think it's natural, and pretty much unavoidable, to do a list, but a limit of three would seem a fair balance, imo. I would defer, though, to those users with more experience (and patience) than myself, as a relative newbie. $\endgroup$ – user81619 Jun 1 '15 at 14:50
  • 1
    $\begingroup$ @ApproximatelyTrue I commented on your question directly. The meta is for more general discussion so let's keep comments on your specific question on the main site. $\endgroup$ – DanielSank Jun 1 '15 at 14:56
  • $\begingroup$ The point I was trying to make was that this metaquestion makes certain assumptions about whether questions with subparts should be avoided, and where the line between focussed and unfocussed is drawn. For newer users such as myself, it would be useful to understand what the policy is and where it comes from in order to engage with a discussion about how/whether such a policy should be publicized/implemented. $\endgroup$ – ApproximatelyTrue Jun 2 '15 at 19:03
  • $\begingroup$ That being said, it has become pretty clear to me that my question is indeed far too broad. But, more generally, as you noted in your question, there is something to be said for value added to the question/answer process (particularly for the original asker) if direct answerers towards (possibly multiple) issues that seem particularly confusing to them. $\endgroup$ – ApproximatelyTrue Jun 2 '15 at 19:13
  • 6
    $\begingroup$ @ApproximatelyTrue Perhaps we should start a new meta thread about the reasons for our ways, but one thing to note is that we strive to be a repository of answers to questions, not a forum. Posting all your questions together makes sense in the "these are my questions" mindset of a forum, but here we strive for the condition "these are inextricably related questions such that anyone who seeks the answers to one will simultaneously want to know about the others." $\endgroup$ – user10851 Jun 3 '15 at 23:18
  • $\begingroup$ @ChrisWhite very well said. $\endgroup$ – DanielSank Jun 4 '15 at 0:38
  • $\begingroup$ Suggestion to the title (v2): Replace 'one' with 'only one'. $\endgroup$ – Qmechanic Jun 5 '15 at 13:56
  • $\begingroup$ So whats your question? :) $\endgroup$ – Tdonut Jun 10 '15 at 3:05
  • $\begingroup$ @Tdonut The thing with the question mark at the end in my post. $\endgroup$ – DanielSank Jun 14 '15 at 0:47
  • 1
    $\begingroup$ Related: meta.physics.stackexchange.com/q/13/2451 and meta.physics.stackexchange.com/q/4599/2451 $\endgroup$ – Qmechanic Oct 14 '15 at 22:26
7
$\begingroup$

My experience on other Stack Exchange sites is that the problem of getting a lot of questions that break guidelines is essentially insoluble.

  • Most people don't read the tour or the help centre before posting so editing the text there can only have a minor effect.

  • On Theoretical Computer Science SE, only research-level questions are on topic. To try to reduce the number of undergrads asking for homework help on CS questions that weren't even necessarily theoretical, we changed the title text prompt for new questions to "What's your research-level theoretical computer science question? Be specific", I guess about a year ago. I didn't notice that having any effect.

  • It's possible to add warning messages when specific tags are used. That would probably be useful for specific subjects that attract a lot of bad questions but doesn't help in this case, since the issue is with the form of the question, not its subject.

Unfortunately, prevention isn't feasible. The focus needs to be on improving or closing inappropriate questions. Improvements, ideally, need to be done quickly. On a recent question on Computer Science SE, I edited out a paragraph that added a vaguely related but independent question and left a comment saying it was a great question and encouraging them to ask it separately, explaining how they could retrieve the text from the edit history. Unfortunately, somebody had already provided a sketchy answer to that part of the question on the back of an excellent answer to the main question, so the asker never bothered to repost it.

$\endgroup$
  • $\begingroup$ It is, however, useful to have something to link to in comments to new users when they fail to follow site guidelines. I'd rather paste a canonical link instead of having to type the same basic message over and over and over and... $\endgroup$ – DanielSank Apr 19 '16 at 0:29
  • $\begingroup$ @DanielSank That's definitely useful, yes. On CS, we have a meta thread with canned comments for the various situations that come up often. There's also a userscript that avoids the copy-pasting; see the same page for details. $\endgroup$ – David Richerby Apr 19 '16 at 2:25
  • $\begingroup$ The other thing is I figure that if a new user is directed to the help center for good advice relating to one of their actual posts, they just might be inclined to read more of the help center. I'm an optimist though. $\endgroup$ – DanielSank Apr 19 '16 at 2:33
6
$\begingroup$

I don't mind If a post contains more than one but essentially related questions.

For instance if a questioner asks Is there a phase transition in So and So model of statistical mechanics? What is the Order Parameter that drives the phase transition, if any? Does any thermodynamic quantity diverge at the phase transition?

and so on. I would not like such a question split for sure.

In the question referenced, clearly Question 2 Had to be a separate one because it was completely orthogonal to the discussion.

I could go either way about separating 1 and 3, I would lean towards not breaking it up because it asks why this interaction and not that, which is related to why this coupling constant and not that.

Now regarding stack exchange's model of accepting answers, I would accept the most helpful answer, that's all.

$\endgroup$
  • $\begingroup$ I think the important distinciton here should be are these questions logically parts of one another ore are they veering of into their own topic. This is naturally a grey area problem sometimes but often its quite obvious I think. $\endgroup$ – pindakaas Jun 13 '15 at 8:41
2
$\begingroup$

More than one question per thread is of course allowed, but let me here list some reasons why there should preferably only be one question per thread:

  • The meaning of voting becomes unclear. Should one upvote/downvote a question/answer if the subquestions/subanswers are not uniformly good/bad? How should one compare an answer that only answers subquestion 1 versus an answer that only answers subquestion 2? Etc.

  • The question becomes too broad. Potential answerers may only answer partially, or get discouraged before they get started.

  • Often readers only read the title question, and ignore other subquestions hidden in the main text.

  • If a question is not already in the title, it becomes harder to search for similar questions.

  • The combinatorical number of possible threads on the site becomes unnecessarily large, especially if the subquestions are unrelated.

  • Certain subquestions may already be duplicates of other posts, while others are not.

$\endgroup$

You must log in to answer this question.

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