A posted question says:

Find the following integral: $\displaystyle \int\cdots\cdots\, dx.$

A big problem with this way of phrasing posted questions is that someone may have copied a homework question without any understanding at all, so that no question about mathematics is present in the user's mind, and the user may just want to copy and paste one of the posted answers to turn it in as homework without understanding any of it.

Someone recently asserted here that the user's understanding of the question must be exhibited in the actual posted question and not in comments under it. But comments under the question may be just as cogent in showing that the poster has a brilliant grasp of the question as is material in the question itself.

It is of course better for such information to be in the question itself, in part because it's less work to understand that.

Is there some reason why the poster's comments under the question should not count as making it clear that the user understands the question and wants to understand the answer? Is such understanding demonstrated in the initial posting supposed to serve some purpose that cannot be served by comments posted under the question?

  • 13
    $\begingroup$ I'd say that adding context directly in the question rather than in comments is better mainly because it is much better visible. (So that people reading that question do not overlook it. IIRC if there are too many comments, they can be automatically moved to chat - in such situation they are even less likely to be seen. Of course, that does not mean that some clarifications or some minor details cannot be explained in the comments. $\endgroup$ – Martin Sleziak Sep 5 '17 at 6:35
  • 6
    $\begingroup$ And I'll add that when I see that a user posted in comments something which can contribute to better question, I suggest quite often that they should include the information provided in the comments also into the question - precisely for the reasons I've mentioned above. $\endgroup$ – Martin Sleziak Sep 5 '17 at 6:36
  • 12
    $\begingroup$ +1 from me - although I disagree with the POV that comments are good place to include information which is very relevant to the question. I upvoted mainly because I think that this is a reasonable question which deserves to be discussed and it is good that somebody brought it up. $\endgroup$ – Martin Sleziak Sep 5 '17 at 6:41
  • $\begingroup$ Usually, if a user makes a comment that is very relevant to the question : for example, altering the nature of solutions he wants to an equation he has provided, then I myself ask him to move it into the question space, so that at least users in the future will not need to look at the comments to understand all the essentials of the question. But there is no reason, as you say, why comments (or a lack of it) should reflect the user's understanding of the question. In fact, I have seen users who do wait for many comments before choosily replying to one. This never implied a lack of interest. $\endgroup$ – астон вілла олоф мэллбэрг Sep 6 '17 at 3:29
  • 1
    $\begingroup$ In continuation, such users were interested in their question, but comments usually flowed once I posted an answer, in response to it. However, posting a complete question (with context/appropriate to stackexchange norms) in the starting indicates that "preparation" has been done before asking the question. The comments are more of an afterthought : a-la "this came to mind after you posted the question", which means you did not think about this question "enough", hence you are perceived to have a lack of interest in your question and it's answer, or seek a hasty no-arguments solution. $\endgroup$ – астон вілла олоф мэллбэрг Sep 6 '17 at 3:36
  • 2
    $\begingroup$ In my opinion, yes, the comments should count towards showing that the OP knows what he's talking about, but there should also be something in the question itself that gives a hint of this. If it's just "solve this problem for me," then it's understandable that people might assume he hasn't even bothered to read the class handouts. $\endgroup$ – Mr. Brooks Sep 9 '17 at 21:38
  • $\begingroup$ Find the integral... There it is! This reminds me of i.pinimg.com/736x/6f/a2/b3/… $\endgroup$ – user349106 Sep 10 '17 at 18:35

If there is a piece of information that readers need to know before answering the question, or before moving on to see the answers, then that information needs to be where the readers will find it.

Stackexchange isn't a message board where threads represent an active conversation between users, and thus newcomers are expected to catch up on the conversation before participating.

Instead, stackexchange is a question and answer site. A posting is expected to have questions and possibly answers. Additional information and clarification may appear in the comments, but that is not where essential information is expected to be.

(and anything really useful in the comments really should be migrated to the question/answer it is attached to anyways!)

If "the user's understanding" is is essential information that readers need to know, then it needs to be where they'll read it: the body of the question.

  • $\begingroup$ You misunderstand. I think what this question is asking about is not the issue of lack of information. I think they were referring to the rules regarding cheating whereby people cannot use SE to cheat on homework. I believe their point is that comments should serve as evidence for moderators to not ban people for cheating. This isn't about whether or not the post is missing information but rather whether or not the asker has proved that they are not a cheater taking advantage of SE. $\endgroup$ – user64742 Sep 10 '17 at 20:53
  • 1
    $\begingroup$ This answer misses the point. There was never any suggestion that essential information should not be in the question. It is better for such things to be in the question than in comments. But when comments do clearly indicate an understanding of the question, then it should not be treated as if it's one of those cases where the reader copied the question and will copy the answer without any interest in understanding. $\endgroup$ – Michael Hardy Sep 10 '17 at 21:29

My .02 small monetary units.

I agree with the sentiment that all the necessary information should be in the question body. Askers leave those bits in the comments only at their own peril. For example the users reviewing the question in a review queue usually don't see the comments. And there is no need for them to check a comment section, should one exist. This is the reason why I, when trying to help a newbie asker, insist that they edit the question. If I like the question, I may (grudgingly) do that myself.

(tangential!!) Having said that I also strongly feel that, for their part, wanna-be-answerers also ignore the comment section at their own peril (as well as other extras that can be inferred from the choice of tags). This goes double when the question is new, and other users are busily pointing out the omissions and/or asking for clarifications in the comments.

To give an obvious example:

Assume that we have a question asking to prove that $xyz=0$ for all the solutions of $x^3+y^3=z^3$. Assume further that the question has been tagged . A bad question all right. But, if some busy beaver with over 10k rep posts an answer: $x=y=1, z=\root3\of2$ is a counterexample, IMHO they fully deserve the downvote (from me). Either somebody has already asked, in a comment, the OP to clarify whether the intention is that all the variables should be integers. Or, they should do that themselves.

  • $\begingroup$ Don't think the question is really about answerers. While I would agree with the example you posed, I do not think it is related to the issue at hand. I believe that if someone posts a serious answer that well answers the question as posed, but a comment from the OP below their own post invalidates this answer, then it is not the answerers fault, but the OP's. I think that most of these content-comments are not posted without prompt, and that if there is a serious answer to be posted, you should not have to ask for clarity. $\endgroup$ – Simply Beautiful Art Sep 5 '17 at 16:38
  • $\begingroup$ @SimplyBeautifulArt I agree that parts of my post are tangential to Michael Hardy's question. But, I believe that answerers should also do their part in quality/clarity control. I am willing to cut newbies a bit of slack, but the more experienced users should look around. I'm afraid I may be soapboxing a bit here, but it is IMHO wrong to concentrate on the role of the askers alone. $\endgroup$ – Jyrki Lahtonen Sep 5 '17 at 16:53
  • $\begingroup$ As a counter example, consider someone asking about how to solve $xe^x=5$ and someone provides a numerical solution involving fixed point iteration as well as a lot of analysis of the problem, but by the time they've posted their answer, the OP has mentioned in the comments that they are searching for algebraic solutions. Do you really believe that this answer should be down voted? As I see it, this answer was in good spirit, and much more practical than the other answer that uses the Lambert W function. $\endgroup$ – Simply Beautiful Art Sep 5 '17 at 16:54
  • $\begingroup$ @SimplyBeautifulArt Yes, this is a grey area, and not all users agree. I may be an extremist here rather than mainstream. IMO the answerers should not rush to post in such a case. Or, they can, but don't necessarily have the right to complain, if the question was edited while they were typing. This applies to all the answers posted withing the grace period = a five-minute-window, during which the OP can fix all the OOPS. $\endgroup$ – Jyrki Lahtonen Sep 5 '17 at 16:59
  • $\begingroup$ I can definitely agree with changes via edits, but to do such only through comments? Likely, the one posting an answer involving numerical approaches, giving an analysis of the situation as well, did not do so within the grace period, but not so late either. And I do not think that, in such a scenario, that one should ask for more clarification as to whether or not such an answer would be accepted, being perfectly valid in its own right. $\endgroup$ – Simply Beautiful Art Sep 5 '17 at 17:10
  • 4
    $\begingroup$ That's exactly why I also disprove of the practice of posting essential details only in comments. It's just that that doesn't IMHO relieve the answerers from their duty of checking also. Ideally we should IMHO all also spend 5-10 minutes dupe hunting (in the cases where there is likely to be a dupe) before typing an answer, but I have given up on that ghost. You see, there's a long list of things that annoy me :-) $\endgroup$ – Jyrki Lahtonen Sep 5 '17 at 17:20
  • $\begingroup$ In your crossed out example, why would I ever think that x y or z must be an integer? I fail to see where the logic of that lies. Is that a well known question or something because I would assume it could at least be any algebraic number and to go further I'm sure many transcendental solutions also exist. Why would one think that needs clarification? It's not self-apparent. Questions should be written exactly how the asker wants them. If they get answers that "don't answer the intended question" then it is their fault for omitting information. $\endgroup$ – user64742 Sep 10 '17 at 20:45
  • 1
    $\begingroup$ @Typhon: Because the question was tagged number-theory. Number theory largely deals with integers (and generalizations) so that would be the default to many users. Similar to multivariable calculus where a reader has the right to assume that $x,y,z$ refer to the Cartesian coordinates in 3-space (unless otherwise specified). And, yes, it is also the simplest $n=3$ case of Fermat's Last Theorem (or rather Wiles' theorem). Calling that a well known question is an understatement :-) $\endgroup$ – Jyrki Lahtonen Sep 10 '17 at 20:48
  • 1
    $\begingroup$ @JyrkiLahtonen fair enough. In my experience number theory primarily deals with the algebraic integers and various algebraic extensions to the integers. Hence, it would be natural to assume in my eyes that algebraic (likely real) solutions are being examined. I would not personally assume integers as in my eyes that's what the elementary number theory tag is for. $\endgroup$ – user64742 Sep 10 '17 at 20:54
  • 1
    $\begingroup$ A fair point @Typhon. The asker should still specify, sure, just in case. But that tag tips the expectations of a casual reader the other way, and they should check before typing an answer. You are right in that algebraic numbers still might be of interest. On the other hand, if the asker used the tag diophantine-equations, then my case would be even stronger :-) $\endgroup$ – Jyrki Lahtonen Sep 10 '17 at 20:58
  • $\begingroup$ @JyrkiLahtonen indeed. That should bring up a red-flag. I just wanted to point out that not all readers will be aware so I wouldn't be so bold to say they deserve a downvote as if they broke a site rule or custom. I for one, will generally assume questions ask in the broadest sense imaginable and maybe comment on what circumstances a conjecture is and is not true (so as to do a complete analysis of the situation). I tend to take answers that limit to a narrower scope as incomplete as they assume a weaker situation then what I asked (which would be for all numbers/functions). $\endgroup$ – user64742 Sep 10 '17 at 21:24
  • $\begingroup$ Shouldn't that be "My $2$ small monetary units" or "My $.02$ large monetary units"? $\endgroup$ – msh210 Sep 18 '17 at 9:13

Based on my initial impressions and some comments, I think you are specifically interested in the comments being "fact of" information that a user comprehends the question.

I expect you are missing the point of "someone"'s assertion. For example, we recently had a question — maybe even the one you have in mind — where the posted question was "Do this exercise", and buried in the comment thread of the second answer was information "I've already solved this exercise. I didn't think the answer was right."

This is an example of how the additional context an completely change the character of the question — the question had two answers, one indicating how to solve the exercise and the other pointing out a loss of precision due to rounding, neither of which did anything to address the actual issue the original poster was having.

(the latter answer was later edited to provide a complete answer that also completely fails to address the user's actual issue, but still likely satisfied the asker's need by accident, through seeing someone else getting basically the same answer he did)

In this example, the important part of the comments expressing the "user's understanding" of the problem is the details of that understanding, not the fact of understanding.

  • 6
    $\begingroup$ Incidentally, these sorts of problems are endemic of problem statement questions, which is a large part of why they are so severely frowned upon. $\endgroup$ – user14972 Sep 5 '17 at 19:28

My take on the issue is fairly short:

Why should I have to read the comments to understand anything?

  • 1
    $\begingroup$ No one suggests that you should have to read comments to understand the question. In the first place, this isn't about understanding the question, but about knowing that the poster understands the question, as opposed to having merely copied a homework question. In the second place, "have to" is not what this is about. I'm saying that in cases where the fact that the poster' understanding of the question and desire to understand the answer is made clear in comments, should the question be treated like one where those are not clear? $\endgroup$ – Michael Hardy Sep 5 '17 at 14:39
  • 1
    $\begingroup$ My bad, but either way, my opinion is that no-one should have to read the comments to gain any information, hence why many questions are closed due to lack of context despite useful information in the comments. $\endgroup$ – Simply Beautiful Art Sep 5 '17 at 16:24
  • $\begingroup$ You can't understand this comment without reading it. $\endgroup$ – Wildcard Sep 7 '17 at 23:08
  • $\begingroup$ @Wildcard But do I want to take the time to read your comment? Especially if I believe the majority of the information of the question is in the question itself, and need no further information to answer? $\endgroup$ – Simply Beautiful Art Sep 7 '17 at 23:11
  • $\begingroup$ Then don't read the comments, just answer the question. Right? But if you understand the question and how to answer it, but want to know additional information about the asker first, then reading the comments is a good place to start. It doesn't seem very complicated to me. (I agree with your take, by the way.) $\endgroup$ – Wildcard Sep 7 '17 at 23:16
  • 1
    $\begingroup$ @Wildcard Ofc, by all means, read the comments for additional information. $\endgroup$ – Simply Beautiful Art Sep 7 '17 at 23:18
  • $\begingroup$ @Wildcard I believe that the point here is that the question shouldn't need any comments to convey all information pertaining to the post. Comments are to point out flaws in posts. Those flaws must be addressed in some fashion at risk of the OP being rude and/or disrespectful in my eyes. $\endgroup$ – user64742 Sep 10 '17 at 20:50

I do consider Comments left by the OP as possible indications of their understanding/wrestling with a Question. It would be better to surface these thoughts in the body of the Question (sometimes the context they provide is essential to shaping responses), but in the worst case (for an interesting problem) I'm willing to do the editing of the Question to include that material.

I even will look (from Review queues) for clues in the OP's interaction with Answer posters for such clarification. This can be somewhat time intensive and occasionally I'll be late to cast a vote from Review (possibly due to leaving a Comment myself).

My view is not to judge the Reviewers and voters who are faster than I am. I don't know that they did or didn't consider exactly the same "facts" in drawing a conclusion. But if you also consider those Comments in your reviewing, more power to you.


How many elements would have to be looked to locate these elements?? is a question I've been helping answer for the OP (admittedly through long winded comments for the most part) but at first the tags for the question were wrong which likely didn't help the audience that came to view it. the problem is about comment sections is they can hide a lot of the details needed to help. in this case, that the person is a visual learner, so most of the comments I made may not help. but we all know math problems can be picky if someone just says how many combinations are there, is that with or without repetitions ? permutations asked about , with or without fixed points ? without these pieces, you can get wildly different answers that only add to the confusion. the comments are a great help, but they don't help people who can't be bothered to look at them. so it's important to not getting multiple answers, that all relevant information be posted in the question itself.

  • $\begingroup$ It's a good example. The OP of that linked Question is clearly engaged but needs help articulating the problem statement. Ideally some agreement reached in the Comments there about a precise problem statement will result in more forceful edits to the current body text. $\endgroup$ – hardmath Sep 10 '17 at 16:47

In my mind this is primarily a stylistic issue, albeit an important one. It is in the questioner's interest to write their question in a way which most quickly produces good responses, and one way to do that is to put all the information into the question itself.

But some questions don't come out that way at first, not until the questioner is prodded by pertinent comments, and then it finally occurs to the questioner to add some further important information in the comments (in which case I might say something like "please, please put that information in the body of the question, please", but it doesn't always work). I doubt that this happens only when a questioner is simply copying a homework problem, or even primarily in that situation.

So it remains useful for use readers/answerers to be flexible on the issue.


You must log in to answer this question.

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