How far should edits go before a new question should be asked?

Question

Somewhat following on from this recent meta question, how much of a change should someone make to a question before they should simply ask a new question?

As a case study, this question was asked some months ago, then apparently abandoned by the original poster immediately after asking. A few hours ago (at the time of writing) a second person edited the question in such a way that the meaning changed significantly (from "what is this property" to "how do I compute this property") and invalidated the answer. Some things to note:

• for full disclosure the invalidated answer was mine (this was how it came to my attention)
• the original question was, in my opinion, trivial
• the new question is much more useful (again, in my opinion)

These points aside however, it still seems like a step too far to change a question (with answers, well, one not-that-interesting answer) in such a significant manner (especially after such a long fallow time).

It's obvious in the extreme case of course, where a sensible question is replaced with nonsense, but here it also seems like an evaluation that we should not be making: "this question is too trivial, and I know a better, related question, so I will replace it". Part of my unease may also stem from the fact that a second party is making this judgement, rather than the original author working out the real question the wanted to ask.

So what is everyone's opinion on where we should lean on this kind of editing? Should editing always preserve intent, or more pointedly, should we replace bad questions with better ones?

Answer 1

I agree with you. It's bad practice to make a large change to the question that changes its meaning, particularly when the question has already received an answer and the change invalidates the existing answer. (I consider this to be true regardless of whether it is the original author making the edit or someone else.)

In this case I am sympathetic to the motivation behind editing the question -- it was an attempt to rescue a bad question and turn it into something reasonable for this site -- but ultimately I agree with you that's not good to change the meaning of a question, particularly when that invalidates an existing answer. It's better to just let the bad question die an honorable death, and ask a new question.

In this case, the question is apparently copied directly from a book, and without citation. That's not good, either.

Answer 2

I apologize: I deleted your answer because it didn't answer the question as it then was, but I hadn't realized that the question had been edited beyond recognition and your answer was perfectly legitimate for the question as it was at the time.

It is definitely bad practice to edit a question in such a way that existing answers are invalidated. The only case in which this can be justified is if such an edit was necessary to prevent the question from being closed (for example, when different answers interpreted the question differently, and the choice is to close the question as unclear and eventually delete it, or to clarify the question in one way or another, matching one answer but making another answer irrelevant).

In this case, we had a mediocre question with a boring answer — not a fault of the answer, but a fault of the question. It would have been best to leave the question alone and ask another question.

Given that the edited question is a lot more interesting, and that you've answered it, I'm inclined to now leave this thread alone, and keep the revised question with your new answer.