Not signed in (Sign In)

Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

Etiquette: Should we do anything if a question is answered well in the comments?

Bottom of Page

1 to 11 of 11

  1.  
  1.  
    I've been browsing older questions, and some of them appear at first glance to be unanswered, but in fact have good answers given in the comments. I imagine that once in a while, such a question will be sent to the top as part of User -1's duties. While this isn't by itself particularly annoying to me, I feel that the lack of answers in the answer spots is a bit untidy. I see three options for dealing with such questions:

    1) Vote to close as "no longer relevant". (Have we actually closed any questions for this reason?)

    2) Copy/paste the good answer as an actual answer, with proper attribution.

    3) Don't do anything.

    I was wondering what the meta community thought would be a preferred course of action. There is also the question of whether a copy/pasted answer should be made community wiki or not. I imagine the copy/paste person will get 15 points for free, even choosing CW, if the answer is accepted. Do people have strong opinions about this?
  2.  

    I don't know about others, but when I answer a question in the comments it's usually because either 1) I secretly feel that the question is not appropriate for MO, but it's borderline enough that I want the OP to have an answer anyway, or 2) I'm not confident in the answer and I want someone to tell me I'm wrong. If answers are being given for the second reason then they probably shouldn't be copied as actual answers. Maybe this is a case-by-case thing (or maybe other people don't do 2)).

    • CommentAuthorHarry Gindi
    • CommentTimeApr 3rd 2010 edited
     

    In the case that the user is still active, we should copy the comments over to an answer, and make the answer community wiki if you feel bad about it.

    In the case that the user has not been back in the past n months, add the answer as above and vote to close, for n to be decided by the community.

  4.  
    Good question.

    In the past, I have voted to close questions as "no longer relevant" in the following situation: in my expert -- :) -- opinion, the question is straightforward and what I feel is the ideal answer has already been given. (Fabricated example: Q: "What is a monoid, and how is it different from a group?" A: "A monoid is a set endowed with a binary operation which is associative and has a two-sided identity element. Thus a group is a monoid in which each element has an inverse.") But this was not very popular: I'm not sure if I succeeded in closing any questions in this way.

    I also had a recent experience where someone asked what I thought was a very straightforward algebraic topology question, and I answered it in a comment. Anton G. parlayed my comment into a community wiki answer. I've been mulling that one over since. It wasn't clear to me what made that question different from questions that we label as probable homework and dismiss. (And, to further muddy the waters, it's not even clear to me that we should dismiss questions because we think they are homework or because they are answered on wikipedia.)

    In this regard I agree with Qiaochu -- after having observed several other veteran MOers engage in the practice of answering a "too easy" question with a comment, I have started to do this myself. [It is also true that less thought goes into what I leave as a comment rather than as an answer, but if I am less than usually confident that what I'm saying is true, I generally say so explicitly.]

    Option 2) seems worth trying: if you think a comment is the answer, then you should feel free to post a CW answer duplicating the material in the comment, with proper attribution. It would be great if this answer got accepted -- that would be very tidy -- but in my experience a lot of the borderline questions are asked by people who are not sufficiently familiar with /committed to MO to take the time to come back and accept an answer.
    • CommentAuthorrwbarton
    • CommentTimeApr 3rd 2010
     
    continuing Qiaochu's list: 3) another possibility is that there is a very simple answer/counterexample and it's unclear whether the author meant to ask the question as stated or left out a hypothesis. In such cases it's clearly better to address this in the comments to give the author a chance to fix the question.

    I have been thinking about this a bit too. I noticed someone recently adopted solution 2; I think this is my preference, but 1 is fine too. It would be nice if the asker could unilaterally close a question as "no longer relevant", I guess.

    Another issue is questions that have not actually been answered mathematically yet have responses with upvotes. This is somewhat related in that both issues make "Unanswered Questions" less useful. I don't have any good ideas about this, though.
    • CommentAuthorrwbarton
    • CommentTimeApr 3rd 2010
     
    BTW I've seen some answers posted in comments which were very nontrivial or at least seemed that way to an outsider in the subject like myself. Not sure what to do about this if anything, but I feel it ought to be dissuaded somehow.
  7.  

    Qiaochu's description of why questions get answered in the comments sounds right to me. Basically, I think that if a question has been answered in the comments and is now dormant, you should do whatever you think is best for future generations. Imagine that somebody has come to the question through a google search, decide what would be optimal for them to see, and move the thread in that direction (which may mean doing nothing).

    If it looks like a question was answered in the comments because the question wasn't appropriate for MO, it's still not appropriate for MO, so you should vote to close for whatever reason fits the question. However, I don't think it's right to vote to close a question simply because it has been answered.

    If it looks like the person answering wasn't very confident in her answer (or just wanted to get a terse answer out), but you think she should post it as an answer, I think you should either post it as a CW answer or leave a comment to the effect of "@username: I think you should post that as an answer." (this will be more effective when we finally get replies to comments merged from SO). The advantage of CWing the answer is that people with <2000 rep can then edit it to make it better if it was incomplete in some way.

    I've also been mulling over the question Pete mentioned. I'm not completely sure that posting Pete's comment as an answer was the right thing to do, but I don't regret it. For questions that might be homework, I feel like it's reasonable to post an actual answer which consists of a reference. This gets around the problem of unanswered-but-actually-answered questions. If it wasn't homework, this is exactly what the asker was looking for, and if it was homework, the asker still has to work through the reference. The main reasons for discouraging homework questions are (a) we want to get the right audience (research mathematicians, not calculus students), and (b) it's annoying when somebody asks you to do his work for him without having put any effort in. Somehow the question Pete mentioned didn't feel to me like somebody gaming MO to do their homework, and it was an appropriate subject matter, so it didn't trigger my "inappropriate question detector".

    3) another possibility is that there is a very simple answer/counterexample and it's unclear whether the author meant to ask the question as stated or left out a hypothesis. In such cases it's clearly better to address this in the comments to give the author a chance to fix the question.

    If the question is unclear, I would vote to close and leave a comment asking the OP to clarify (and perhaps link to http://mathoverflow.net/howtoask). Once the question has been sharpened, vote to reopen. It's really frustrating when any answer to the question has to start off with a guess about what the question actually is. Typically, the OP realizes that something is wrong after a couple of answers have been posted, then edits the question so the answers look totally bogus. Closing the question until it's cleaned up prevents this sort of thing from happening.

    • CommentAuthorgrp
    • CommentTimeApr 3rd 2010
     
    If we have reached the point that it is clear what constitutes a good MathOverflow question and a good MathOverflow answer, then good answers which might be overlooked as comments should be made accessible as answers, even as accepted answers so that those both inside and outside of MathOverflow can see them.

    I suspect that we have not reached that point yet, and that the best method is to encourage the original poster of the good answer to post in answer format, or give permission for it to be posted in that format.

    Perhaps MathOverflow in its mature stages can produce (or spawn) an editorial board which will compile the best questions and answers in a nice reference format, with links to the original posts for sake of completeness and correct attribution.

    Gerhard "Ask Me About System Design" Paseman, 2010.04.03
  9.  
    Here's a possible test case:

    http://mathoverflow.net/questions/13919/finding-the-base-of-exponentiation-subject-to-range-constraints

    The user has recently been active, the question was borderline, and a simple answer was given in the comments. Any thoughts on what should be done? (Actually, this is why I don't particularly look forward to having 3000 points, because it would be hard to decide in many cases whether or not to close.)
    • CommentAuthorBen Webster
    • CommentTimeApr 4th 2010 edited
     

    I would say that if you wouldn't vote to close the question, absolutely you should post the answer. part of the point of MO is to have a canonical place on the internet to stick a question and its answer in a way that makes it immediately apparent what the answer is. I think the community norm which makes us feel as though we are "stealing" an answer if we weren't the originator of it is a counterproductive one. If you feel uncomfortable "claiming" the answer, you can make it community wiki, but it is a positively thing for later users of the site to have the answer in the answer box, hopefully voted up a few times.

  11.  
    I think you should post the answer but I think ethically you need to 1) make it community wiki and 2) explicitly mention which user posted the commented.

1 to 11 of 11

  1.  
Back to Discussions Top of Page