Not signed in (Sign In)

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

    • CommentAuthorMartin B.
    • CommentTimeJun 21st 2012
     
    Why do so many people on mathoverflow use the "comment box" for actual answers?!

    It seems to me that this happens more often than in the early days of mathoverflow. One reason might be that one is in a hurry and doesn't want to spend time writing a detailed answer. But: Often comments actually contain whole answers. And it's no harm to write a short and concise answer!

    What do you think? And above all, how can we change this (if you are in favour of a change)?
    • CommentAuthorplhersh
    • CommentTimeJun 21st 2012 edited
     
    I'm happy to see your question, because I was wondering what to do if one knows how to solve a question but doesn't have time to post the full solution right then -- my inclination was to leave a comment telling the OP how to get started, and then fill in the full solution later.
    • CommentAuthorAngelo
    • CommentTimeJun 21st 2012
     
    I sometimes write a comment for something that is easy or sort of standard; I guess that the (irrational) motive is a feeling of embarrassment at the idea of getting credit, and reputation points. Of course, both "easy" and "standard" are highly subjective terms.
    • CommentAuthora-fortiori
    • CommentTimeJun 21st 2012
     
    I usually post a comment if I feel that the question has not reached its final form yet, like http://mathoverflow.net/questions/99563/existence-of-regular-sequences This is sometimes combined with "this is too easy/standard, probably I misunderstood something, or the OP intended to ask something else".
    • CommentAuthorArtie
    • CommentTimeJun 21st 2012
     
    Like Angelo, I sometimes answer in comments for things that shouldn't really be MO questions. But I think people _only_ using comments for answers is pretty rare: Jason Starr has lately filled the vacancy left by BCnrd, but I can't think of another user who does it.
    • CommentAuthorMartin B.
    • CommentTimeJun 21st 2012
     
    @Angelo: This was also my impression. But this may offend the questioner and suggest that his question is so boring/irrelevant/etc. that the experts don't bother to write an answer, even if they actually know an answer and put it into the comment box ....... yes, this is irrational.

    I've observed this for months, maybe I will collect example MO threads here.
    • CommentAuthorbsteinberg
    • CommentTimeJun 21st 2012
     
    This practice has the disadvantage that MO will bump the question back because the software thInks it is unanswered. In theory one can make a CW answer if you don't want to gain rep.
  1.  

    As mentioned by Artie, the case of BCnrd is relevant to this. This has been discussed before http://tea.mathoverflow.net/discussion/950/questions-spoiled-by-comments/. Briefly, the consensus reached at that time was that this behavior is discouraged but not unacceptable. It is also perfectly appropriate for another user to expand the comments into a (CW) answer, as Harry Gindi did for several BCnrd nswrs.

    • CommentAuthorArtie
    • CommentTimeJun 21st 2012
     
    By the way, in case my earlier post gave a different impression, let me make clear that I don't have any objection to people only writing answers in comments. If too many people started to do this, it would decrease the usefulness of MO, but at the moment it doesn't seem to be a widespread practice. Moreover, beyond the basic rules of conduct, it doesn't seem wise to regulate too much the way people use the site --- I would much rather have people like Jason Starr contributing their expertise in the comment box than not at all.
  2.  
    Since I'm occasionally guilty of this, I guess I don't have a problem with the practice. My reasons are similar to those given by Angelo. I feel obligated to make a much greater effort to write an answer, whereas with a comment I feel free to merely give a hint or suggestion which may nevertheless be helpful.
    Finally, and this may sound petty, if the person asking is anonymous, I'm a lot less likely to make the effort to write a complete answer.
    • CommentAuthorHMŠ
    • CommentTimeJun 24th 2012
     
    Just my 2 cents: The question http://mathoverflow.net/questions/100430/how-to-interpret-this-class-of-numbers and my comment which seems to satisfy the OP just as much as an answer would (judging on the fact that he was online since then and gave no feedback). I just couldn't have put a wikipedia link in an answer and consider it a legit answer.
  3.  
    Actually I think that would have been a perfectly legitimate answer, although I understand where you're coming from.
    • CommentAuthorHMŠ
    • CommentTimeJun 24th 2012 edited
     
    Actually, I wanted to ask for future reference - how to formulate such an answer without making it look like RTFM (since the wikipedia article on tetration is something the OP would have found after 5 minutes of googling even if he never heard of it before)?

    P. S. What to do if a question gets an answer through comments - since it makes an impression when searching or browsing MO that the question has 0 answers - but it's in fact resolved? Should the user who answered through comment be encouraged to place an answer containing the comment and therefore offer a chance to the OP to accept it, for the sake of completeness and easier browsing?
  4.  
    If you come across a question that has been answered in the comments and it bothers you that it hasn't been answered in an answer, you can always copy the answer from the comments and post it as an answer yourself, being careful of course to note that all you are doing is copying over comments made by someone else. If you are worried about gaining undeserved points for doing this you can urge people not to vote your answer up, or you can designate it Community Wiki.
    • CommentAuthorHMŠ
    • CommentTimeJun 24th 2012
     
    Sounds fair. Maybe it was a silly question for me to ask though - but it's nice to have an "official" guideline for cases like that, in case one feels the need for it.
  5.  
    To follow with Gerry Myerson's most recent comment, I think it's better to designate it CW than to "urge people not to vote your answer up." My reason is that often those who ask questions which can be answered in comments are not the best about coming back and accepting answers, especially if the answer contains nothing new. But if a question has neither an accepted answer nor any upvoted answer then the software will continue to bump that question to the front-page. So it's better to make a CW answer which people can vote up just to stop the software from bumping the question.
  6.  

    Given this question, one is then led to ask, "why do people who complain about leaving answers as comments then delete such answers once they are given?".

    • CommentAuthorMartin B.
    • CommentTimeJun 30th 2012 edited
     
    @Scott: This is offtopic. I've exchanged mails with Pavel and Damien. And meanwhile I've posted an improved version of this question. The original one didn't make clear enough what my question was (and obviously in the new one the people still cannot read and are happy to give references to other questions they would like to answer more ... same story as usual).
    • CommentAuthorMariano
    • CommentTimeJun 30th 2012
     

    And editing the old one was impossible for some reason :)

    • CommentAuthorDamienC
    • CommentTimeAug 17th 2012
     
    Hi,

    I just saw this discussion here and I must say that, despite our e-mail exchange, I agree with Scott.
    I was happy to spend some time turning (on request) a short comment into an answer, but after your question was deleted and re-posted I prefered not to re-post an answer.
    I guess it would have been more efficient to edit the old question, as suggested by Mariano.
  7.  
    If someone adds something to "comments", it does not seem to be flagged as active. This strikes me as a bad idea, given that there isn't a whole lot of difference between answers and comments. Is it considered acceptable to move a comment to an answer?
  8.  

    If you want to bring a post to the top, Gene, just edit the original question or an answer to it. Be careful not to do it too many times to avoid turning it into a community wiki question.

  9.  

    Gene, what I've sometimes done if I feel that someone's comment deserves to be promoted to answer status is post that comment as an answer, in my name but in CW mode and with credit to the commenter. I feel this is acceptable. I've seen other experienced users do the same.

    • CommentAuthorplhersh
    • CommentTimeOct 6th 2012
     
    I figure if I leave a comment within a day or two of when people are posting answers to the question, it's likely that the people most interested in the question will revisit the question and see the comment. It seems like there's often a discussion that takes place in the comments among the people most interested in a question. I'd rather wait until this comes to a conclusion before updating an answer, if I update the answer at all. There are times I purposefully leave things as comments to avoid rebumping the question -- for instance, when I suggested a possible way of attacking a question to a student, it was clear he was reading the comments, and that was what to me seemed important. Personally, if I leave a comment of substance and someone else wants it to be an answer, I'd rather the person ask me in the comments if I would move it to an answer than have the other person post it as an answer without asking.
  10.  

    I'd rather the person ask me in the comments if I would move it to an answer than have the other person post it as an answer without asking.

    Yes, people do that too, and I agree it's polite to do so. But if a comment just sits there for a long time, then under the present software, a commenter will not be alerted in mail to the fact that someone has made that request of him or her. Some things about MO are really inefficient!

    There are times I purposefully leave things as comments to avoid rebumping the question

    But if the question is never answered except in comments, then the MO bot will eventually come around anyway and bump it to the front page. (Hm, is that how it works? or does this happen only if no answer is ever accepted?) Anyway, I believe it's much more in line with the spirit of MO to give good answers (not just comments) to good questions.

  11.  
    @Todd At least one problem will be solved with the new stacke exchange software, one can adress comments directly to other commenters which then get notified.
  12.  

    So I've heard, Michael. Say, when is this supposed to take place?

    • CommentAuthorplhersh
    • CommentTimeOct 6th 2012
     
    @Todd: I agree with you on a lot of what you said, and I definitely do try to answer questions fairly often. But I don't always have time to do that when I feel like I do have something potentially useful still to say; I'm also not always convinced that it is in the best interest of the person asking the question to get a full answer as opposed to simply a suggestion of a paper to look at or something the person might try (which might or might not work). I suppose sometimes it's probably better if one doesn't have time to give a full answer just to keep quiet and let someone else do that. I'm probably going to keep doing a mix of things, but will think twice about just leaving a comment.
  13.  
    Tood, I guess this depends on the big migration timeline. I don'nt know and would like to know myself when this will be. There were no recent updates I think.
  14.  

    @Patricia: I appreciate your thoughtfulness, both here and at MO. Actually, I think a reference to a paper, or a lightly tendered suggestion, can make for an excellent answer, often better than a highly detailed answer. But I understand these are sometimes tough calls.