Not signed in (Sign In)

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

Feature requests: minimising/restricting edits in some way

Bottom of Page

1 to 14 of 14

  1.  
  1.  
    I have now seen several examples on MO of the following thing: questioner posts question, someone posts a correct solution (usually of the form "this is trivially false because of the following counterexample"), the questioner then substantially edits the question (something which the solver may not even notice) e.g. by adding in extra assumptions ruling out the counterexample, and then someone adds a comment to the solution saying "this solution is nonsense because the counterexample doesn't fulfil the conditions of the question". This leads to confusion. One simple way of stopping this from happening would be to not allow people to remove anything from their original question when editing it, and only to let them add stuff at the end of the form "oh yeah, furthermore assume k has characteristic zero" or whatever, which is then automatically flagged with "EDIT" and possibly also dated. This would lead to less confusion. If a poster really wants to substantially edit their question then they simply ask another one; that way the comments made about the original question (which would no longer apply after the edit) could not be misinterpreted and confusion would not reign.

    Of course there are people who don't "abuse the system" and carefully flag their edits. But I've seen sufficiently many super-confusing edits now that I think it might be time this possibility is considered.
    • CommentAuthorcritch
    • CommentTimeDec 6th 2009 edited
     

    In my opinion, I think it's unfair to edit a question if a correct answer has already been given to it, even if the question was asked incorrectly by accident. So when I edit a question, I always make sure I'm not invalidating any correct answers to it as stated. Perhaps this or something similar should be adopted as general ettiquette.

    I do think editing a discussion to clean up unenlightening errors is important to making MathOverflow readable "for posterity". However, ever since Dave Brown suggested it, I discourage completely erasing "enlightening errors", i.e. mistakes I or others have made that other would probably learn something from.

    As well, if people make comments as to how a question/answer could be improved, if it's valid, I think the best thing to do for future readers is to just implement the suggestion and reverse-engineer the content, and <i>include accreditation</i> to the people whose comments helps with the improvement. This has the effect of making some comments obsolete, which I think is not unfortunate, as long credit goes to those who deserve it.

  3.  
    If there's a trivial counterexample, perhaps it should be first posted as a comment, not an answer.
  4.  
    I agree that this is a problem, but I'm not convinced that restricting the technology is the right solution. I know this has happened at least twice for questions I've asked -- both times I've had a certain "model" of the question in my head, where there was an implicit assumption which didn't even cross my mind could fail and so I didn't state.

    I feel like it should perhaps be "unofficial policy" to somehow acknowledge in these kinds of situations that your question wasn't correctly posed. In fact there are at least three ways to be polite and interoperable here that would render the need for some kind of restriction on edits moot.

    1. If you ask a question and someone points out a trivial counterexample that shows that your assumptions weren't what you wanted them to be, acknowledge that you screwed up at first, either in a comment to their post (or a comment to the question if their counterexample was in a comment), or in the body of the edited question.

    2. If you answer a question with a trivial counterexample, suggest what they might be implicitly assuming but not writing down! If you're right, then someone who reads the answer will (hopefully) go "oh, this answer only applied to an earlier, less precise version," and even if you're wrong you'll hopefully get the questioner to consider his/her unspoken assumptions about the problem.

    3. If you come across a post that isn't explicitly stated as being this kind of trivial, question-sharpening example, but looks like it might be, treat it as if it were. In particular, it'll probably do more harm than good to make some comment along the lines of "this is wrong, it goes against the assumptions of the problem!!" If you absolutely must, check the edit history and leave a comment saying something like "This was a counterexample offered to an earlier, incorrect version of the question, which no longer works because the question has been refined

    As Python hackers say, "We're all consenting adults here." At least for now, it's probably easier and better to assume good faith and try to make flagging edits standard MO etiquette than to restrict edit capability on the chance that people might get confused.
  5.  
    But we're not all consenting adults. Just because you and your friends follow sensible rules doesn't mean everyone else does. Somehow that's what I am suggesting is the problem: because many people don't do the sensible thing, that's why one might consider coercing them in the right direction. Maybe people with less than x points can't edit their own posts much or something?
  6.  
    Hm, for some reason I was under the impression that people with not many points couldn't edit their own posts at ALL, but checking the FAQ that doesn't seem to be the case.

    Of course just because I do something doesn't mean that everyone else will, but if we actively tried to make "flag substantive edits to your posts" a part of the general rules of etiquette, then hopefully new members of the community would pick up on it after asking a couple of questions. So the worst-case scenario would be that someone asks such a question, gets a counterexample, edits to exclude the counterexample, and leaves forever; and then there are enough people (about 20 now, I think, and growing) with editing privileges to clear things up if it's really confusing.
  7.  

    Here is an idea. Right now, when I go to edit a post, I see a message in the side bar reading

    Good edits:

    Fix grammatical or spelling errors.

    Clarify meaning without changing it.

    Correct minor mistakes.

    Add related resources or links.

    Always respect the original author.

    Could we add a line to address this issue? Maybe something like "Clearly indicate any edits that change the meaning of the question."?

  8.  

    That list already holds the solution. Number two on the list is:

    Clarify meaning without changing it.

    Adding new conditions which excludes a counter-example does change the meaning.

    Perhaps vigilant people with edit privileges should step in to re-edit if it is thought that this situation has occurred.

  9.  

    Perhaps vigilant people with edit privileges should step in to re-edit if it is thought that this situation has occurred.

    Absolutely agree. If somebody breaks a sensible convention, the best way is to just show the person what the convention is.

    Also, sometimes people essentially start a new question/answer thread in the comments to a correct answer. In that case, I tell the asker that he/she's likely to get more by asking a new question.

  10.  

    I think we should try to encourage people asking new questions, if they realise that the original question was broken for some reason. I'm not sure how to do this except through comments and community editing practice.

  11.  
    What if the trivial answer is "the zero ring" or "the empty set"? I mean as adults, surely we can tell when this is just being obnoxious, no?
  12.  

    @Andrew- my interpretation of that line was that it was mainly aimed at when you're editing other people's questions.

  13.  

    @Ben: no reason why it shouldn't apply to your own questions as well. Can we change the wording to make this clear (assuming others agree with me)? It may be "hard coded" into the site.

    To the main point: I'm against technological solutions for issues of this type. Perhaps the FAQ on voting down should say something about this possibility as I guess it just might not have occurred to people that it can happen - it's not easy to see how major an edit was and most seem to be simply retagging so it's a fair bet that an answer is an answer to the current form of the question. In general, the community system will sort this out: if you notice an answer you gave has garnered a vote down then go and check it and see if someone has said something like "Doesn't answer the question.". Then edit your answer to make it clear that this was for an earlier form of the question.

    Perhaps also we should be better at voting for questions because of how they are asked, not just their content. So if a question is edited in a good way, vote for it and say "+1 for gracious edit" (or whatever).

    Always better to praise good behaviour than stamp down on dubious behaviour.

  14.  
    If one of my posts gets to -1, I just delete it.

1 to 14 of 14

  1.  
Back to Discussions Top of Page