24

The suspected duplicate question is closely related, but not an exact duplicate: it relates to ever-expanding questions, with follow-up after follow-up, which is not per se the focus of this question, not least because such follow-ups may come in the form of comments rather than modifications to the question.
Because there is substantial overlap, however, I've added the gist of the linked question to this one.

I think the problem is well known:

  • A question is asked.

  • Answers are received.

  • The OP modifies the question substantially (possibly repeatedly), invalidating existing answers.

A less harmful, but still problematic variation:

  • The question itself isn't changed, but the OP realizes their real problem was something else, and they self-answer with whatever they ended up doing or simple ignore answers - either way, answers that addressed the original question don't get accepted.

Per https://meta.stackoverflow.com/a/290704/45375, the first case can be addressed with rolling back to the original form of the question, but that doesn't apply to the second one.

In the interest of guiding posters, however, in both cases I wish there were an official help topic to point them to, which:

  • explains why questions shouldn't be modified in substance after having received answers (due to invalidating existing answers) and that a new question should be asked instead.

  • that it is important to pick an accepted answer to the question as asked (even though they may still post a - properly framed - "afterthought" answer themselves).

  • additionally, to give renewed life to this closely related request:
    explain why users may - for good reasons - stop answering ever-expanding questions, where the OP keeps asking follow-up questions, and that an answer to the original question should be accepted, with follow-up questions raised in a new question.

  • as for changes that do make sense: perhaps say that it's OK to correct spelling errors, syntax errors that are incidental to the problem, fluff, needlessly verbose parts, incidental parts of the code (unless they're being quoted among the answers and removing them would cause confusion).

Update: Having such a help topic is not mutually exclusive with using rollbacks to revert a question to its original form:

You could simply reference the help topic from the rollback edit's comment, which not only saves you from having to explain the same thing over and over, in the limited space of a comment, but carries more weight due to pointing to an official resource.

Similar to how shortcut [mcve] in comments points to http://stackoverflow.com/help/mcve, something like [dontchange] could point to the new topic.

16
  • 2
    It is not always cut and dry, and may be difficult to make clear in a brief entry in the help center. If the OP is clarifying the question, and in the process of doing so invalidates answers, then the edit tends to be more acceptable than someone who is changing the question because they are a help vampire. You would need a way to explain what kind of edit is acceptable and what isn't Commented Jul 19, 2018 at 20:03
  • 2
    @psubsee2003: Fair points - let's add them to the new topic. Obviously, a help topic can never cover all angles and all subtleties of individual cases, but having official guidelines would help, and would be an easy reference point, given how frequently the problem happens. Instead of having to re-explain the issue every time it happens, you can just point to the help topic, perhaps even in the comment accompanying a rollback.
    – mklement0
    Commented Jul 19, 2018 at 20:13
  • 3
    that it is important to pick an accepted answer to the question as asked But the OP isn't obligated to ever accept an answer. And how can the OP state which answer best solved the problem in the question when that wasn't the problem they actually had?
    – BSMP
    Commented Jul 19, 2018 at 20:17
  • 2
    @BSMP: Not obligated to accept an answer, of course, but if they're given a rationale that makes sense to them, they may choose to. Why shouldn't they be capable of knowing which answer addresses the question as asked, even though they realized they needed a different problem solved? These two aspects are unrelated.
    – mklement0
    Commented Jul 19, 2018 at 20:22
  • 2
    Lets clarify something, the purpose of stackoverflow is?_An online community for developers to learn, and share their knowledge_ In many cases the OP may not know the right question and asks the wrong question. The resulting discussion and answers help he OP resolve the real issue. The best answer therefore should be based on the correct way to complete the task upon the question was asked, regardless of the question being correct or not. We have comment boxes for _clarification or to point out problems in the post. _The solution to a problem can always be for a question not asked
    – Gary Allen
    Commented Jul 19, 2018 at 20:24
  • 1
    @mklement0, would you have admin and the community now policing "The correct answer", which is a completely subjective term?
    – Gary Allen
    Commented Jul 19, 2018 at 20:26
  • 4
    @GaryAllen: Can you explain how my request for a new help topic providing guidance to people who ask questions is related to having admin and the community now policing "The correct answer"?
    – mklement0
    Commented Jul 19, 2018 at 20:28
  • 2
    @GaryAllen There's a different term for that. It's normally called upvotes and downvotes.
    – Servy
    Commented Jul 19, 2018 at 20:29
  • 3
    @GaryAllen: What matters to future readers is to have narrowly focused questions of general interest with narrowly focused answers. The narrow focus may come from an XY problem on the part of the OP, but the question they ask a result may still be of general interest - and answers to it may be the right solution for OTHER scenarios. It's perhaps unfortunate that the real issue wasn't properly framed in the original question, but the community is better served by highlighting the answer that answers the question as asked.
    – mklement0
    Commented Jul 19, 2018 at 20:49
  • 2
    @GaryAllen: As stated, accepting the answer that solves the question as asked does not preclude posting an "afterthought" answer that explains the XY problem and a better solution for the question that should have been asked, but there's perhaps a better alternative: create a new question framed the way the original one should have been, and self-answer that.
    – mklement0
    Commented Jul 19, 2018 at 20:57
  • 1
    I never did it, but I think in this case it is not a problem if you rollback, even against the will of the OP. Note, rollback-wars are considered here more seriously as on the wikipedia! If there is a rollback war, don't get into it, rather flag the post and let the case for the mod. And explain nicely to the OP, why is it bad what he is doing. Your votes to the question shouldn't be affected, you should vote the content, not the person!
    – peterh
    Commented Jul 19, 2018 at 21:26
  • 1
    @peterh: Yes, rollback is an option, but even there the help topic could be helpful: simply reference it in the rollback edit's comment. Explaining nicely is great, but I don't want to have to do it over and over, and there's only so much space in a comment; having an official, community-consensus resource always carries more weight.
    – mklement0
    Commented Jul 19, 2018 at 21:30
  • 3
    This might make more sense as a question asking "How much am I allowed to change my question" with an answer that explains it in the answers here. For better or worse, a lot of information about how things work appears exclusively on Meta and Stack Exchange don't seem intent on changing that. Commented Jul 20, 2018 at 8:57
  • @Dukeling: Good point and good idea. Let's see if this goes anywhere first (it may not, as you say), and if it doesn't, we can try to create a canonical answer here the way you suggest.
    – mklement0
    Commented Jul 20, 2018 at 9:07
  • What guidance do you suggest for this question changes?
    – Cœur
    Commented Jul 22, 2018 at 9:49

2 Answers 2

-1

I think most reasons invalidating this suggestion are invalid. Reactions to this post along the lines of

Maybe OP used wrong terminology

or

Maybe OP misunderstood what their actual problem was

are irrelevant in my opinion. The questions was asked as is, and people put their effort into it - I think making clear you should not invalidate their effort is a no brainer here. There is no "wrong terminology" from the Q&A perspective unless the question becomes nonsense - if a question was asked using some wording, and a good answer for that appears, than I think SO has got another good question for the database. The fact that OP considers the wording bad in retrospect means they want to ask a new, different question (often).

I also do not agree any consideration should be taken to other possible bad behavior - if the OP has a substantially different question they should ask it. If people mis-identify the new question as a duplicate or spam, I think the community should fix that separately (a single meta post usually gets this done).

There is also the matter of an evolving question, which is another, subtle yet annoying variant of this. As such I think the suggestion is great, and I really do not think many words are needed to explain this, nor do they need to be harsh:

Substantial editing of your question Please refrain from substantially editing your question in ways that may change the essence and meaning of the original question, especially after receiving an answer. People have already put effort into the original version, and it would be impolite to invalidate their work. If your realize you meant a different question, by all means, post another one within our guidelines.

Note the community may rollback any such edits in case of already submitted answers, in order to keep the questions within our guidelines.

But what if I realize there is another problem after receiving the answer? You may comment on the answerer's post to request expansion, but be advised they are not obliged to answer something different than the original post. Again, no harm in posting a second, good question.

I think this would work across the SE network, since the problem is the same on all sites.

Suggestions from comments

mklement0

Suggests a more formal wording - we do not want to address "impolite" behavior, but confusion from future readers.

7
  • In other parts of the network, they aren't afraid of reformulating and even turning around questions if it helps to bring the quality up.
    – Braiam
    Commented Dec 10, 2018 at 14:14
  • @Braiam are you sure? I have seen conflicting posts, with people debating this (on IPS for example). Do you have specific meta posts from other sites with such a decision? In that case your last point is strong - the help files would not be able to be cross-network.
    – kabanus
    Commented Dec 10, 2018 at 14:15
  • @Braiam good chance this question needs to be migrated to meta SE now that you raise the point. I think this is a debate cross-network.
    – kabanus
    Commented Dec 10, 2018 at 14:29
  • I wouldn't hold my breath waiting for migration.
    – Braiam
    Commented Dec 10, 2018 at 14:30
  • @Braiam :D I was getting blue.
    – kabanus
    Commented Dec 10, 2018 at 14:31
  • 1
    @kabanus: The part of your answer with the suggested wording is promising, except I wouldn't frame refraining from substantial edits as a matter of politeness; while that is also true, the more important point is that future readers should be spared the confusion of seeing answers that do not match the question. I don't fully understand the points made at the beginning of your post. Yes, why they ended up with the question that should no longer be edited is irrelevant with respect to that question, but the suggested help topic would address possible motivations for wanting to edit.
    – mklement0
    Commented Dec 10, 2018 at 15:10
  • 2
    @mklement0 I was trying to get a "soft" version out. I agree with you, but I am not sure a new user will understand all the "big picture" issues, and being polite is usually enough incentive for many people. Regarding the beginning of the post, I was addressing some of the issues raised in the comments that suggest we allow users extensive edits. I disagree they are points "pro" editing in this case - If someone used wrong terminology (from their point of view) and got an answer - than that answer is good, and so was the original question (hopefully).
    – kabanus
    Commented Dec 10, 2018 at 15:16
-7

explains why questions shouldn't be modified in substance after having received answers

What if the answerers misinterpreted the question due, I don't know, well the question not using the correct terminology to explain the problem? What if there was someone that actually saw this issue (the not using the correct terminology), commented asking for clarification and OP clarified by editing its post?

Now, lets say the thing gets implemented. What happens when OP tries to ask another similar but "different" question and it's closed as duplicate? Because I assure you, it already happens.

In summary, this would only cause catch 22 situations, where the wealth of information would be restricted due the adamant reaction of answerers to recognize that their answer wasn't valid to being with.

If instead answerers asked for more clarification and used their close votes before answering vague or ambiguous questions, such situations (where OP is forced to edit their questions after being answered) would not happen.

Oh, btw, remember that the help center is universal for all SE sites. On other sites, they need to be able to change their question due the nature of the questions themselves, like on Interpersonal Relationship.

11
  • 1
    Yes, people answering questions prematurely rather than using comments to guide the asker toward a question that is worth answering is a problem, but it's a separate one that perhaps deserves its own help topic. If an asker asks another question, then whether to close it as a duplicate depends entirely on the content of that question, and if closure happens, why is that a problem? It would imply that they didn't understand the help topic, but the only argument arising out of that is to make the topic as clear as possible.
    – mklement0
    Commented Jul 22, 2018 at 13:19
  • What do you mean by due the adamant reaction of answerers to recognize that their answer wasn't valid to being with?
    – mklement0
    Commented Jul 22, 2018 at 13:19
  • That's an interesting point re shared help content, but I'd say the guidance applies site-independently (why do you think Interpersonal Relationship is different?). That said, I've just added a few points that would be SO-specific (pertaining to code), but I still think they could be part of a shared topic; if not, then it's time to create SO-specific topics. That there are none at the moment doesn't strike me as a good argument not to write this topic at all, if the need for it is recognized.
    – mklement0
    Commented Jul 22, 2018 at 13:20
  • @mklement0 "but it's a separate one" it cannot be separated, sine you need to define how to recognize if the edit is in fact another question and not just your misinterpretation of both text. "why is that a problem?" Why it is not a problem? Lets put an example: say an edit change too much the question, someone comes along an ask OP to ask a new question instead, reverting the edit. OP then asks a new question with the text of the edit, someone also comes along and close it as duplicate of the previous one, because they are too similar. How would you resolve the situation?
    – Braiam
    Commented Jul 22, 2018 at 15:36
  • @mklement0 "What do you mean by" I, for one, am not bothered by OP editing their question in a way that invalidate my answer. If that happens, I just ask a new question that specifically ask for the answer I provided, since I was the one with the misunderstanding. Some people prefer to impose themselves to OP instead.
    – Braiam
    Commented Jul 22, 2018 at 15:37
  • @mklement0 "Interpersonal Relationship is different?" because is a non-technical site. You can't give a technical objective answer to questions about relationships, there are too many variables and bias on all members.
    – Braiam
    Commented Jul 22, 2018 at 15:38
  • Re non-technical site: It's not about technical objectivity, it's about the answers matching the questions - that is subject-matter independent.
    – mklement0
    Commented Jul 22, 2018 at 16:18
  • Re separate problems: They're separate, because one addresses askers (don't answer a question until it's stable and worth answering), and the other answerers (don't change your question substantially/continually after the fact) - though cross-linking them makes sense. If the new question is indeed too similar to the old one, then it should be closed; I see no problem there, given that I'm sure it is happening already. A help topic to point to can only be beneficial in this situation.
    – mklement0
    Commented Jul 22, 2018 at 16:18
  • Re invalidating answers: The premise of this issue and suggested help topic is not a misunderstanding on the part of the answerers - that's yet another matter. The premise is ensuring that the answers match the question, for the benefit of the community at large, whether or not it is what the OP ultimately needed - that can be pursued via a new question.
    – mklement0
    Commented Jul 22, 2018 at 16:24
  • @mklement0 "answers matching the questions" well, what if the question evolve to match answers OP is looking for?
    – Braiam
    Commented Jul 22, 2018 at 17:10
  • 1
    I don't understand your question.
    – mklement0
    Commented Jul 22, 2018 at 17:12

You must log in to answer this question.

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