On this site, we have a well-known policy on editing code in questions: Don't. After all, that code is up for review, in all its glorious ugliness and buginess. But clearly, such reasoning cannot apply equally to code in answers.
I was therefore surprised to find that a nice edit on one of my answers had been rejected for reasons that would only make sense if this were code in a question. The edit was attempted twice unsuccessfully (review 1, review 2), and rejected for the following reasons:
This edit deviates from the original intent of the post. Even edits that must make drastic changes should strive to preserve the goals of the post's owner.
No, it did not change any intent, but merely fixed the faulty expression of my (the answerer's) intent.
Fixing the author's code for them removes a potential point for reviewers.
No, this is an answer, not a question. Fixing simple mistakes should be all right.
Please don't edit OP's code. Comment on the post instead. Code blocks are for OP to fix on this site.
Citation needed! Why wouldn't it be OK to make an edit to the code in an answer?
(In the meanwhile, I invited the editor to re-submit the edit, and have already approved it. Ergo, this is a discussion on general site policy using that edit only as a post-mortem example.)
Since the last reason was given by a mod, I trawled through the help center and meta in search of policy guidance relating to answer edits, but there is none. All I could find were discussions on question edits.
What is the site policy on answer edits? Are they held to the same strict standards as question edits, or do we allow a more collaborative style of answering?
Regarding this specific edit, is there a community consensus on whether the edit was reviewed correctly? If so, could you please link to existing codifications of answer editing guidelines? As it stands, I assume all those edit rejection reasons to be unfounded.
In either case, I agree with RubberDuck that “we really should write down our editing guidelines anyway. Otherwise, we'll someday end up with 2k users who don't know any better.”