This following question was motivated by this question, though many of similar nature can be found throughout MO.

The Problem: The question asks whether any bound exist on an expression, and conjectures what it might be. An answer partially disproves the conjecture, i.e, only part of it. Now, and for all future viewers, the question still contains the now known-to-be-wrong conjecture.


I don't think that the question or the answer are bad, or counterproductive. I only think that it is not clear that although the answer is helpful for the OP and is interesting for anyone else, there's still half of the OP question unanswered. My Question, therefore, is how to make it more visible, so that other people will try to answer it?

  • $\begingroup$ Here is a similar situation. A asks a question. B gives a trivial counterexample, "constant functions satisfy that" or something. A modifies the question to eliminate the counterexample, such as adding "nonconstant" to the text. Surely that is to be permitted? $\endgroup$ – Gerald Edgar Apr 13 '16 at 14:23
  • $\begingroup$ I actually think this should be the best way to go. This is the kind of edit I wanted to do, but I wanted to consult first as it is a bit intrusive. $\endgroup$ – Amir Sagiv Apr 13 '16 at 14:26
  • $\begingroup$ But the question you linked doesn't seem to fall into this category. The OP thanks the writer of the answer as having helped them understand something. Editing such a question doesn't seem warranted. If you think of a refined better question, why not just ask that separately, adding a link to the first one. $\endgroup$ – Lucia Apr 13 '16 at 15:58
  • $\begingroup$ @Lucia see the editing. Is it clearer now? $\endgroup$ – Amir Sagiv Apr 13 '16 at 18:22
  • 1
    $\begingroup$ That's easy enough: if you have a more interesting question that needs answering, then you ask that question! $\endgroup$ – Lucia Apr 13 '16 at 18:29
  • $\begingroup$ @Lucia but the question is already there - it's just not clear enough that it is unanswered, because another (very related) question from the original OP was answered. $\endgroup$ – Amir Sagiv Apr 13 '16 at 18:31

A question was asked. An answer was given. For the benefit of future readers, edits should only contain relevant information and corrections. Changing the question enough so that the answer becomes less relevant is discouraged. (Deletions may be appropriate, but will not be discussed in this post.)

As a further service, one can add to the question, or even better link to a new question, which asks the modified conjecture. Now there is a path which gives the readers the option to continue on the path of discovery or not. To muddle the signposts by changing the conjecture after it has been answered properly is not doing a service to the readers of this forum.

Edit 2016.04.13: If you have a contribution to make on part of the original question, even if it is only a literature reference where the (at present unanswered) part is discussed, that will fit nicely as a brief answer. If you have some additional material that might flesh out the part while not answering it, making a new question and editing the old question (and putting your initials near the link to show your emphasis) to link to the new question seems appropriate. Clarifying the path of discovery may be important to you and to some others, but it should not be presumed or made to appear that it is a motivation or path of the original poster. End Edit 2016.04.13.

Gerhard "It's Not All About You" Paseman, 2016.04.12.

  • $\begingroup$ I somehow disagree with the first statement - A question was asked and and an answer was given. There was a question, and a possible answer - which was disproved. To be strict- The first answer is not really an answer to the question. $\endgroup$ – Amir Sagiv Apr 12 '16 at 19:01
  • 2
    $\begingroup$ If the answer or the question are bad to begin with, then editing/deleting should occur. The culture here is to encourage people to contribute well; if the answer made a contribution and then was made invalid because of a change in the question, that discourages that and other contributors. I made the assumption that the answer was a contribution. If it isn't, then what is the issue? Gerhard "There Is Also The Commentary" Paseman, 2016.04.12. $\endgroup$ – Gerhard Paseman Apr 12 '16 at 19:06
  • $\begingroup$ @AmirSagiv I don't follow you. In the linked question, it seems clear that a question was asked and a satisfying answer was given (both question and answer were received very well also). $\endgroup$ – Tobias Kildetoft Apr 13 '16 at 5:03
  • $\begingroup$ You have a double negative in the first paragraph that was probably not intended ("less irrelevant"). $\endgroup$ – Tobias Kildetoft Apr 13 '16 at 5:03
  • $\begingroup$ @GerhardPaseman, the "less irrelevant" comment is for your answer. $\endgroup$ – Amir Sagiv Apr 13 '16 at 6:42
  • $\begingroup$ @TobiasKildetoft , I don't agree. The question is (1) what can be said about X and (2) I conjecture that X' is Y. The answer says that the conjecture is wrong, and that not much can be said about X'. But question (1) still remains unanswered. I don't think the question is bad, rather that it consists of two parts, and it will be counter productive to cancel it, or branch it. Maybe the edit should be that part (2) is now disproved, but still show it so people will have more context. $\endgroup$ – Amir Sagiv Apr 13 '16 at 6:46
  • $\begingroup$ @AmirSagiv The OP accepted the answer, so it was clearly helpful to them. $\endgroup$ – Tobias Kildetoft Apr 13 '16 at 6:48
  • $\begingroup$ @TobiasKildetoft, I must have been misunderstood. It is a very helpful answer. The OP had a conjecture and someone helped him understand why it is wrong. My only claim is that, by the sheer presentation of the question, it is unclear that there's still an open question - namely (1). My only wish is to find an editing scheme that will make it more obvious. $\endgroup$ – Amir Sagiv Apr 13 '16 at 6:52
  • 1
    $\begingroup$ @AmirSagiv Why do you want to do this with an edit, when a new question would clearly be a great way to do it? $\endgroup$ – Tobias Kildetoft Apr 13 '16 at 6:53
  • $\begingroup$ @TobiasKildetoft That's a good answer. I think that the "new" question is not new, it is well phrased withing the "old one", and people can benefit from the existing discussion, conjecture and its counter example, in order to answer the question. It only needs some refurnishing to be more visible - something like "Part 1 xxxxxx. Part 2 - Conjecture yyyy. Edit - I now know from answer zzz that conjectures of the form of mine can not be true", $\endgroup$ – Amir Sagiv Apr 13 '16 at 6:56

You must log in to answer this question.

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