6

Suppose I asked a question in an area where I am no expert. Then the question is getting answers that point out that the question is ambiguous or has not enough information for a meaningful answer. Comments are added that discuss what the question might actually be, if it makes sense, how ambiguities could be addressed etc.

In such a case, would I start a completely new question "take #2", perhaps with a pointer to the old, confusing question?

Or adding a comment and clarifying. One problem with comments is that they have a size limitation and only limited formatting (no math formulae etc.) and that it's hard to find in the cloud of other comments. Users that see the question for the first time are likely to miss the one comment that adds additional information.

  • 2
    As Script47 says, you should clarify your question by editing it, not in comments. However, your remark that you can't put formulas in comments is incorrect: if a site supports MathJax, so do its comments, eg physics.stackexchange.com/questions/526982/… – PM 2Ring Jan 27 at 9:57
  • I have voted to close as needing more focus. As written, without a link to the problematic question, it's too abstract to be able to give you concrete guidance. – aparente001 Jan 28 at 1:18
  • 1
    Does this answer your question? How can I get answers fast? – Rob Jan 28 at 3:06
6

In such a case, would I start a completely new question "take #2", perhaps with a pointer to the old, confusing question?

No, the edit feature exists for a reason.

Or adding a comment and clarifying. One problem with comments is that they have a size limitation and only limited formatting (no math formulae etc.) and that it's hard to find in the cloud of other comments. Users that see the question for the first time are likely to miss the one comment that adds additional information.

Follow the pattern:

Refactor early, refactor often

You should be directly editing your question to make it better as feedback pours in, not using the comments to explain yourself and leaving the question ambiguous. Yes, you can use the comments to explain but once you've done that, you should go back to the question and edit it to remove the confusion.

Note: Don't edit your question so drastically that it invalidates answers or completely changes the requirements. Folks don't like having their time wasted.

| improve this answer | |
  • 1
    Ok, but when there is an answer that points out that the question is ambiguous, then editing the question will render at least that answer no more matching the question... Thus using "edit" is not only to fix typos then. – emacs drives me nuts Jan 27 at 10:01
  • 3
    @emacsdrivesmenuts An answer that says your question is ambiguous, is not an answer, it never matched the question, it is a comment requesting clarification at best. Such an "answer" should be flagged as not an answer. – Mark Kirby Jan 27 at 10:17
  • Even a minor edit can change the sense drastically. – Mari-Lou A Jan 27 at 10:22
6

If there are existing answers, and not just comments, then whatever edits you make must not invalidate any of them that have a net upvote (i.e. answers deemed useful by the site’s community). If they do, then it would be fair to see them rolled back because those answerers have attempted to understand your question and provided an answer in good faith.

If you cannot make such an edit, then I think it is fine to start a new question that starts by linking to the original one, and explaining that the question you are now posing would have been an edit to the original, except that you wanted to respect the effort that the original answerers volunteered.

| improve this answer | |
  • What about adding an edit tag at the and of the edited question like [EDIT]: Added information XYZ. – emacs drives me nuts Jan 27 at 10:23
  • 1
    I don't think that "Edit:" should ever be used in in a question title or in a question body. In the latter it interrupts the readability that might otherwise entice a potential answerer to read and understand your question. – PolyGeo Jan 27 at 10:27
  • I mostly agree, although an upvote isn't always evidence that an answer was provided in good faith. It's not uncommon for FGITW answerers to take a wild guess and provide an answer to an unclear question, and sometimes those answers get upvotes. – PM 2Ring Jan 27 at 10:37
  • @PM2Ring true but I think it is as good a guide as any and encourages users to downvote answers that are not useful. – PolyGeo Jan 27 at 10:39
  • Fair enough, it is ok, as a guide. The point is to not make edits that would invalidate answers where the answerer has made a reasonable interpretation of the question, and made some effort to provide an answer matching that interpretation. – PM 2Ring Jan 27 at 10:48
3

In addition to the other answers, that all make fair points: ideally, ask for advice right there, with your specific question.

You see, the theoretical rules and practices are all good to know (such as that neat refactor early, refactor often), but in the end, it is always a case by case thing.

For example, sometimes a more experienced user might edit your question for you, so that it is less ambiguous, while also preserving already existing content.

And beyond that: definitely avoid to get into "fights". The other day I came across How can I ask a question when I have a poor grasp on English?. The title of that question was really clear, so was the one-sentence body. I wrote down an answer ... and later found that the question saw an "edit war" between the author and community that lasted over several years.

In other words: identify "the most reasonable" way forward, then stick with that. And when the community disagrees, for example by rolling back your edits, then accept that.

| improve this answer | |

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