Not signed in (Sign In)

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

Feature requests: editing after voting/accepting/answering

Bottom of Page

1 to 4 of 4

  1.  
    • CommentAuthorESQG
    • CommentTimeDec 13th 2009
     
    This is in part the same issue raised on <a href="http://tea.mathoverflow.net/discussion/74/minimisingrestricting-edits-in-some-way/#Item_14">restricting edits</a>, but I think that my focus is different.

    It is currently possible to edit a question after it has been answered, rendering an answer obsolete. It is also possible to change the contents of an answer after it has been voted for or accepted (though not to delete it after acceptance, which causes <a href="http://mathoverflow.net/questions/8707/knots-that-unknot-in-a-manifold">a problem in this question</a>).
    The theoretical problem with this is obvious: when an answer or a question has major edits done to it, not only other answers but also votes and acceptances no longer reflect what they should. The minimum on reputation combined with the seriousness of MO is probably why we haven't had pranks like changing a good answer to something outlandish after it's been accepted, but I think there is still a problem when people misjudge and overedit their own answers.

    I just found that I voted for a concise partial answer I liked, went back, and found the answer edited beyond all recognition and expanded to a full and clunky one. I no longer wanted to upvote it, so I took my upvote away. I also found that I almost suggested that rather than delete the answer to the question linked above (which the poster wanted to do after realizing it was incorrect) the user should merely change the entire contents of the answer. But this is theoretically possible, is it not?

    Here's a solution that's worse than the problem: when a post is "significantly" (how to track this?) edited it could lose all its votes. But is there some modification of this solution that will work? Or perhaps something like the notifications we receive when people comment or reply to our comments, or something so simple as a sign appearing next to an answer or a question to indicate that it has been edited since it first appeared and got most of its attention (tracked by views or votes).
  2.  

    I think people should be encouraged to leave old stuff as it is and just post new questions/answer. Consider an example scenario:

    Alice posts a question she thought about for the last two weeks but omits the crucial requirement. This makes question easy, though not obvious. Bob soon replies with a simple answer and Carol posts a textbook reference. Alice comments that without the assumption question is indeed easy and that she plans to edit the question. Dilbert, a very experienced and wise user, suggests that Alice posts the new question instead. Alice does that, accepting the simple answer and linking the new question from an old one, and the whole company moves on to work on the new post, which quickly gains 20 upvotes and 10 partial solutions. Everyone is happy.

    As for your suggestion, once people agree that the above is the social convention, it can be easily enforced by anyone with 2k+ reputation reverting the post to its pre-big-edit stage, or, in extreme cases, by closing it. The date of most recent edit is clickable and will reveal the edit history which can be examined to see which edits were significant and which not.

    • CommentAuthorBen Webster
    • CommentTimeDec 14th 2009 edited
     

    I'll be honest: I just don't think this is a serious problem. I agree with you that it's annoying, but the ability of people to edit their answers does so much more good than harm that I just don't see it as getting worried about. If nothing else, there are much higher priority issues in terms of adding features.

  4.  

    I just took the dramatic step of rolling back an edit which changed the intent of the question after ithad been answered, and writing to the author asking them to post a new question.

    "Are graph manifold groups residually finite" --> "residually hyperbolic".

1 to 4 of 4

  1.  
Back to Discussions Top of Page