90
$\begingroup$

I have noticed, of late, people providing answers/copious hints to the question in comments instead of actually adding an answer on their own.

Other people might choose not to answer, as the question has already been answered, and it might offend the person who commented with the answer. In time, this will lead to the site having a lot of questions with 0 answers which actually have an answer in the comments.

Some possible consequences of having answers in comments:

1) Potentially waste time of people. Personally speaking, I read the question before reading the comments, and so I have just wasted a few minutes when I possibly didn't have to. Some people might choose to answer it straight away and find that they have just repeated someone's comment in an answer.

2) The question goes to the unanswered list and they are bumped up periodically by the Community user.

3) Comments are not indexed by the full text search. So we have valuable content just lying around in dead questions.

Here are a couple of examples:

If $f$ continuous and $f(x^2) = f(x)$, then $f$ is a const. This one has an incorrect answer, while the correct answer is in a comment.. The commenter was kind enough to add a detailed answer, upon request.

Connectedness of sets in the plane with rational coordinates and at least one irrational

This problem seems to be more severe with homework like questions. If one is worried about earning trivial rep, one call always add an answer and mark it CW.

So, in case someone adds an answer in comments, what should one do (while not hurting any feelings perhaps)? (Ignore is not an option :-))

I wanted to open this, so we could direct any future comment-answerers to what the community opinion is regarding this (if any).

Note: I am not suggesting we have a policy/faq entry. This question would serve that purpose, if possible.

$\endgroup$
8
  • 3
    $\begingroup$ Here's a related question: meta.math.stackexchange.com/questions/1148/…. In that case it was the asker whose question was answered in the comments not wanting the question to remain "Unanswered", and the accepted advice was for the asker to post a CW answer. $\endgroup$
    – Jonas Meyer
    Jan 27, 2011 at 22:43
  • 3
    $\begingroup$ In this case, a simple request to Sivaram produced an answer post. $\endgroup$
    – Arturo Magidin
    Jan 28, 2011 at 17:05
  • 3
    $\begingroup$ Actually, I think this is somewhat of a balancing act. For instance, sometimes a learning student might ask a question, where it would in fact be more appropriate to give a hint instead. At other times, the OP may find a comment useful and will ask the commenter to make the comment into an answer which the OP is willing to accept (for example: math.stackexchange.com/questions/107731/limit-points-of-sinn/…). In general, where a hint is more appropriate, I think one should comment rather than answer. $\endgroup$
    – user2093
    Mar 13, 2012 at 4:46
  • 6
    $\begingroup$ @WNY: You can always edit your answers (which can start out as hints). They are not set in stone. By adding an answer you let other folks know that you are willing to provide guidance to OP and closure to the question. If you just comment, no one has any clue whether you are going to follow up etc and might lead to the question having no answers. $\endgroup$
    – Aryabhata
    Mar 13, 2012 at 5:28
  • 5
    $\begingroup$ This annoys me to no end, people just posting their answer as a comment instead of an answer. A problem that I'd add is that it comes across as an attempt to have their answer be the first thing people see after reading the question, without having to go through the voting process. I guess people do it because they think their answer is too short or something, but it's still annoying. $\endgroup$
    – Kyle Delaney
    Nov 23, 2015 at 3:36
  • 2
    $\begingroup$ I'm a prime offender for this. I'm aware that other SE sites strongly discourage answers in comments, but I see one difference on MSE; often a hint can lead the OP to find the answer themselves (and they will get more out of it). Is it acceptable to write a hint as an answer? I couldn't find any clarification in the help center. $\endgroup$
    – stewbasic
    Apr 20, 2017 at 3:08
  • 3
    $\begingroup$ @stewbasic: If you are willing to followup and edit to make it a complete answer later, posting a hint-answer should not be a problem... $\endgroup$
    – Aryabhata
    Apr 20, 2017 at 22:20
  • $\begingroup$ @stewbasic See also this math-meta issue, which questions whether hint-comments that are not answers should be deleted. $\endgroup$
    – user2661923
    Dec 4 at 19:39

4 Answers 4

Reset to default
82
$\begingroup$

I think one should first leave a comment asking the user to post their comment as an answer, explaining that this will allow the question to be marked as answered. This often works, and it may encourage the user to post full answers in the future.

If the commenter won't do so, I agree that one should go ahead and post the comment as a community wiki answer. (As of writing, 10 other people agree that this is a good idea, going by the votes on this previous question).

On the other hand, if one wants to add further information and not merely duplicate the content of someone else's comment, there is nothing wrong with posting that as one's own (i.e. not CW) answer. If one is concerned about hurting feelings, one can acknowledge in the beginning that this answer is an elaboration of the answer given by the other user in the comments.

$\endgroup$
7
  • 2
    $\begingroup$ What if one decides to expand a bit on the answer? Especially if it is a new user, they might need the rep to gain certain privileges. Marking it as CW probably won't help in that regard. $\endgroup$
    – Aryabhata
    Jan 27, 2011 at 23:09
  • 4
    $\begingroup$ @Moron: I edited my answer to the effect that if you are adding your own contribution, it should be OK to make that a non-CW answer. (I thought I should put that in the answer so people can vote it up or down appropriately, while one can't vote down a comment.) $\endgroup$
    – user856
    Jan 27, 2011 at 23:40
  • 27
    $\begingroup$ answers posted as as comments are fair game to be reposted as true answers. If users continue to post answers as comments over time, I would consider that a somewhat abusive use of the system. $\endgroup$
    – Jeff Atwood
    Jan 28, 2011 at 1:28
  • 1
    $\begingroup$ I am posting community wikis for questions resolved in the comments as recommended in this article, but I am receiving reputation points when people upvote them, which seems unfair, since I really didn't do anything (often it is my question being answered)! $\endgroup$
    – Eric Auld
    Jul 28, 2013 at 15:20
  • 1
    $\begingroup$ @EricAuld That can't happen, if the post is CW. Possible explanation: after you posted CW answer to your question, more users saw it and upvoted the question along with the answer. In the reputation report, it's not easy to tell what brought you points: question or its answer. $\endgroup$
    – Post No Bulls
    Dec 16, 2013 at 2:51
  • 1
    $\begingroup$ Would it be a faux pas to ask people to post their comments as answers if it's someone else's question and not your own? $\endgroup$
    – Kyle Delaney
    Nov 23, 2015 at 3:38
  • 1
    $\begingroup$ @Kyle: Not at all, people do that all the time. Heck, people often just repost other people's comments as answers too, so you could just do that and cut out the middleman. (As you can tell, in the ~5 years since I wrote that answer my opinion has become much less conservative.) $\endgroup$
    – user856
    Nov 23, 2015 at 4:00
4
$\begingroup$

I actually wanted to post the same question since this issue has been bothering me for a long time, but I found this post (and lots of duplicates), so let me just add my 2 cents here.

I think we have a huge problem with people writing answers in comments. My assumption is that sometimes they think their answer is too short to be an answer, even though it is an answer. Even worse, they might feel that they are not supposed to write answers for "trivial questions" - but why do they comment then?

I always ask the person to post it as an answer then. It does not always happen, though. Sometimes I then write a CW answer myself, copying the comment into it (even on questions that I asked myself).

Here are some recent examples where this happened, showing that it just happens too often.

It also happens on mathoverflow on a regular basis.

Personally, I think that we need to have a better solution here. For the start, it would be good to have a reference (EDIT: the Help Center already provides such a reference) where it is explained clearly that and why answers should not be written inside of comments. Maybe accompanied with a teaser text that links to this reference. Then we might be able to copy paste this text.

But in the long run, this will not change much, since there are a lot of poeple who are new to the site and just do not know this recommendation. To new users, there should be a popup before they are submitting a comment. Alternatively, there should be a flag option for comments, saying that it is actually an answer.

Let me start something that may be copied (edits are welcome!):

This is an answer to the question, not just a comment. Please post your answer as an answer. This brings extra visibility to the answer und puts the question off the unanswered list (otherwise it will be bumped up periodically). Also notice that comments are not indexed by the full text search, don't have any revision history, can only be edited for 5 minutes and allow only limited markup. Comments should only be used to clarify, not answer the problem. See What are comments and How do comments work for more information.

New contributor
Martin Brandenburg is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
$\endgroup$
6
  • 7
    $\begingroup$ Making a CW answer certainly is a step up from a comment, and a gesture of personal humility maybe, but honestly it doesn't seem like an intended use for CW. Really in that case what is called for is just a toggle for an answerer to hit when they write a solution saying "I don't want rep for this." It would get the effect one wants without the attachments of making the solution easier to edit by newcomers. Anyhow I have no objection to this use of CW, and I think there is an argument it is not necessary. $\endgroup$
    – rschwieb
    Dec 1 at 15:06
  • $\begingroup$ I agree, but it's maybe the best we can do for now, right? Qiaochu also proposed this (long time ago) here. $\endgroup$
    – Martin Brandenburg
    Dec 1 at 20:46
  • $\begingroup$ Request clarification: what is your take on the issues presented in this math-meta article? $\endgroup$
    – user2661923
    Dec 4 at 19:37
  • $\begingroup$ I don't have any take on this issue. But thanks for pointing me to the excellent answer by Xander Henderson. $\endgroup$
    – Martin Brandenburg
    Dec 4 at 20:28
  • $\begingroup$ In my opinion, the linked comment from mathoverflow is a prime example of a reasonable answer in ths comments. It does not meet the criteria for an actual answer (link only) and the commenter may not have the time and/or motivation to fledge out the details, but the link may help the OP or others to give a proper answer. $\endgroup$
    – MaoWao
    2 days ago
  • 1
    $\begingroup$ I think the course of action proposed here, nudging the commenter to give a full answer and otherwise writing an answer yourself, is a good one. But I think enforcing a ban on such comments would result in a net negative for the site. $\endgroup$
    – MaoWao
    2 days ago
2
$\begingroup$

If you answer a few questions that get closed, you get a note from the Moderation Team, telling you to stop doing that. If you answer a few more questions that get closed, you get suspended. If, instead, you write answers in comments, nothing happens to you.

"Well," you say, "you shouldn't be answering questions that are going to get closed." OK, but some questions are borderline, it's hard to tell whether they'll get closed or whether they'll survive. Some questions, after they get closed, the community reconsiders, and decides they shouldn't have been closed, and they get re-opened. It's not always easy to tell whether or not a question meets community standards.

And some questions get closed as duplicates. With $4,000,000$ questions on site, there are a lot of places for a duplicate to hide, and even those of us who know some of the tricks for finding them are going to miss them sometimes. And, again, sometimes questions get closed as duplicates by mistake, and the community votes to re-open, so even duplication isn't always easy to decide.

All in all, if you have any doubt, it's safer to answer the question in a comment than in an answer.

Well, somebody had to say it.

$\endgroup$
8
  • 2
    $\begingroup$ Yup. Glad you said it. $\endgroup$
    – JonathanZ
    2 days ago
  • 3
    $\begingroup$ This is against the site policy and doesn't make any sense. When you answer a question in a comment because it might be a duplicate, you essentially bypass the rule that duplicates should not receive answers just by sending your answer in a different format. If you think that the question is a duplicate, don't answer it, and if you have time: mark it as a duplicate, done. $\endgroup$
    – Martin Brandenburg
    2 days ago
  • 4
    $\begingroup$ @Martin, if I know the question is a duplicate, of course I vote to close it as a duplicate. If I think maybe it's a duplicate, but I don't become aware of an earlier target, I might answer it in a comment. This makes perfect sense, and site policy is occasionally pretty stupid. $\endgroup$
    – Gerry Myerson
    2 days ago
  • 2
    $\begingroup$ This sounds like an effect of the Enforcement oQS. Borderline questions receive borderline comments that are not answers but clarification requests or guides to improve the post. Because the blame of low-quality questions is shifted or extended to authors of carefully-written answers. $\endgroup$
    – peterwhy
    yesterday
  • 1
    $\begingroup$ I think EoQS is badly conceived indeed that it puts the blame of low-quality questions on those who give [usually well-written] answers in the answers section, and encourages "clarification requests" which are really online tutoring sessions that end up answering the question anyway and also end up rewarding the bad questions. I also think that at this point, about everything this side of a successful proof of the Riemann Hypothesis is really a "duplicate" on some level or another, and that marking a question as a "duplicate" is really just a remedy for the low-quality questions on this site. $\endgroup$
    – Mike
    10 hours ago
  • $\begingroup$ I myself will often either (A) answer the question in the answer section, and then delete my answer if the question turns out to be a duplicate, or (B) [much less often] say in the comments 'I know this is a duplicate try this approach'. $\endgroup$
    – Mike
    9 hours ago
  • $\begingroup$ * marking a question as duplicate is often just a remedy for the low-quality questions...some otherwise good questions do get marked as duplicates $\endgroup$
    – Mike
    8 hours ago
  • 1
    $\begingroup$ @Mike I read that deleted answers can negatively affected the author, though I am not sure what the effect is for self-deleted answers. So another reason to not put transient information as answer? $\endgroup$
    – peterwhy
    8 hours ago
-1
$\begingroup$

I find answering in the comments to be annoying, to the point where I have reported the answer-comments to moderation, and gotten many of these answer-comments deleted. The thing is, answering in the comments completely changes the way MSE works.

To the best of my knowledge going by the Help Section, this is how MSE is supposed to work. Someone asks a high-quality question--one that is thought-provoking and interesting to a wider community, not just to the person asking. Then someone else thinks about it and comes up with a clearly expressed answer to that question. The person who answers the question is rewarded not only with a sense of accomplishment--and an uprated skillset for figuring out a solution to the question, but also upvotes which translates to an increasing set of voting privileges for that user on this site. The site meanwhile benefits from this in that it becomes a repository of interesting questions with well-written answers. Sounds great to me. I mean, it isn't perfect--as pointed out by Gerry Myerson above, it isn't perfectly clear what counts as a duplicate or not--but overall it is a sound model I think.

However, answering in the comments, throws a wrench into all this. There is no more rep or satisfaction to be gained from answering a question and writing it up nicely for posterity, as someone likely already presented the crux of the answer in the comments. I say it kind of takes the fun away from participating, and that can't be good for the site. There is a reason why "rep" was put in for answering the questions when the Stackexchange sites were built--as silly as it sounds on some level. Also, low-quality questions are getting rewarded with an online tutoring session on here, which means....more low-quality questions. That can't be good for the site either.

I already can see the backlash from those who say that my views are elitist, that MSE should really be an online office-hours where the goal should be the student answering his own question. I actually don't agree with this at all. For one thing there is no doubt a lot of students will use this site in a way their professors did not intend to, to try to get an unfair advantage on their fellow students. The joy of helping someone, for me, is also tinged with a 'wait did I just help someone on their take-home exam'. For another, I believe one should get rep by answering questions that are of interest to the wider community e.g., answering questions asked by someone else for one thing, instead of by themselves. Why should you get points on here for doing your own homework?

My 2 cents.

$\endgroup$

You must log in to answer this question.

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