Yesterday I answered a question that had a given scope. Today I saw that the question had edits that changed its scope. I saw it because of a downvote to my answer - I believe it was due to my answer no longer being valid. It could even be flagged as NAA and I would agree.

This is not the first time it happens. I usually either edit my answers or let it be. I feel like applying a plain rollback would be authoritary and not nice to new users. I strongly believe that newcomers don't do such edits with bad intentions - they just haven't learned how things work here.

However, this situation keeps happening. It's only a matter of time until another edit invalidates someone's answer.

I am aware if these previous questions in meta:

Is it considered good practice to edit your question after answers have been given?

On editing questions and invalidating answers

But they mention nothing about rollbacks.

Should we have a procedure then? Something along the lines of rolling back edits that would invalidate answers, but also leaving comments about why the rollback was done, asking them to open up a new, different question with whatever new parameters, and point them to the sandbox?

I don't intend to do anything about the question edit I've mentioned - it has already gathered an answer within its new premises, so a rollback now would be more harm than good. I am asking for future cases.

  • Could you still link the question? Examples always help. – dot_Sp0T Oct 16 at 18:44
  • 1
    @dot_Sp0T worldbuilding.stackexchange.com/q/127585/21222 - initially it asked whether and how much vampire bites hurt. I answered that in literature it is udual for it to be not only painless but pleasant. The OP then changed the wording to state that vampire bites hurt, and is now asking how to turn them into a medical procedure. – Renan Oct 16 at 20:29
  • 1
    Same thing happened on this question, because of conflicting tags. – John Locke Oct 16 at 20:42
  • Are you referring to edits made by the OP, by other users based on what the OP has said in the comments, or by other users that make edits the OP has not clarified? – John Locke Oct 16 at 21:44
  • @JohnLocke edits made by the OP. – Renan Oct 16 at 23:03
  • 1
    @Renan Aww, I was hoping that your clarification would invalidate the answer :) – John Locke Oct 17 at 1:09
  • 1
    Two points: 1. those of us who edit questions should do so very carefully so as not to upset the integrity of the query itself or of any answers given. Edits for grammar & spelling & sense are generally quite safe; & 2. those of us who discover a query that has been damaged in some way, or asnwers that have been compromised have to be very careful when rolling back, too! Please make sure that you don't just revert an edit without checking it thoroughly. It could be that useful spelling & grammar edits will get thrown out along with the excessive substance editing! – elemtilas Oct 18 at 7:15
up vote 8 down vote accepted

There is a network-wide consensus: Exit strategies for “chameleon questions”

And I personally like it. In short, it is like that:

  1. Edits are not allowed to invalidate answers
  2. If scope of question changes slightly, and it is still an original problem (aka answers are not wrong, just slightly off) notify OP and answerers and let them figure it out
  3. Use chat
  4. Yes, roll back

Simply rolling back without saying anything does not solve the underlying problem, but sometimes rollback is a part of the solution.

  • 2
    Right. When I make edits to queries, I try to stick to spelling & grammar. Changing the question itself, well, rather defeats the purpose. Unless the OP has a laundry list of questions or sub-questions. I might try to pick the "best of the bunch", but will only do that based on any answers given. I of course do not want to invalidate such answers. The laundry list of questions I would try to work into the main question (a stealth list) or perhaps into "considerations" or "concerns" (also a stealth list!). – elemtilas Oct 18 at 7:10
  • Thans for finding that question :) – Renan Oct 18 at 10:49
  • I believe in this, but it's frustrating that people answer so quickly that the OP doesn't have the chance to improve the question before those improvements invalidate answers. Just as there shouldn't be "chameleon questions," there shouldn't be "anchoring answers." Frankly, answers to underdeveloped questions should be downvoted. – JBH Oct 22 at 17:27
  • 2
    @JBH I believe we simply should close faster, and diamonds shouldn't be afraid to put question on hold and request clarification. – Mołot Oct 22 at 20:15
  • I agree with that completely. – JBH Oct 22 at 21:11

I usually leave the question alone, although I do note to the OP that edits that invalidate previous work are frowned upon. Then I note that my answer is pre-edit and no longer applies at the top of my existing post. Then I leave it completely alone for at least an hour or two so I don't say something I'll regret later. Sometimes I feel that a new answer to the new question is warranted, but usually I'm not feeling like being that helpful when this kind of thing happens. I would personally lean away from using the rollback button myself but I would definitely suggest, assuming there are no answers to the new version, to the OP that it may be both more appropriate and more productive for them to rollback the existing question and ask a new question with the new content.

You must log in to answer this question.

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