Best way to handle suggested edits to accepted answers

Question

I recently was presented with this edit:

https://stackoverflow.com/review/suggested-edits/4003993

In general, what's the feeling on allowing an edit of an accepted answer, and one that has so many upvotes? Should it have been a comment to the answer rather than an edit to the answer? With that many upvotes, and with the code unproven, I don't know if I feel confident in accepting the edit.

The community's guidance is appreciated.

Answer 1

In general

Screw "in general". Evaluate the edit on its merits, in the context of the post being edited. If you don't know enough about the subject to be able to tell whether it's good or bad, skip it. If it made sense to reject edits based on simple rules like this, we could just do that automatically and skip the whole "reviewer" bit. In fact, we already do that, so if you're even seeing an edit there's at least a slim chance there's some worth to it.

The only edits you should evaluate consistently regardless of content or context are the fake ones generated for audits - they're crap 100% of the time or your money back.

P.S. Comments are often hidden, deleted or unread. Posting any non-trivial amount of code in comments is just cruelty to the reader. Don't put anything of value in comments if you can avoid it, and don't encourage others to do so either.

Answer 2

The fact that the answer is accepted and highly upvoted is irrelevant.

In the words of the official guidance:

When should I edit posts?

(…) Common reasons for edits include: (…)

• To correct minor mistakes or add updates as the post ages

This edit is a minor change to the post; it respects the core logic and leaves most of the details intact, and purportedly correctly handles a case that the original case did not handle correctly. It is intended to correct a minor mistake in the original code.

So it boils down to: does this edit indeed correct a minor mistake?

• If it is correct, accept the edit.
• If it is incorrect, reject the edit.
• If it is not quite right but on the right track, improve the edit.
• If you aren't sure, for example because you lack the requisite subject knowledge, skip the edit.

A comment would not be appropriate to correct a mistake in an answer. Comments are only for temporary notes. For example, “X is wrong but I don't know how to fix it” should be posted as a comment, and stay until someone comes along and makes the required fix. On the other hand, “this is wrong and here's the precise way to fix it” should not be posted as a comment, it should be an edit that implements the fix.

Answer 3

Edits are there to improve the presentation of the author's content, not to change it. Their goal is to make the reader better comprehend whatever it was that the author was trying to say.

If you would like to add entirely new content then the appropriate medium to do so is either through a comment to the answer, or if you are providing enough content that you feel it is worthwhile, an entirely new answer to the question. (Just be sure that if posting a new answer, it fully answers the question, and that it's not just a reply to another answer.)

The rejection of the edit was correct, and for the correct reason.

That the answer is accepted has no bearing on this decision.

If a post is marked as Community Wiki, though, that changes matters entirely. The whole point of marking a post as CW is to say, "This post is no longer just mine. Anyone and everyone is free to change the content of the post, not just its presentation." If the post is marked as CW and there is a suggested edit that changes its content, then approve/reject the suggestion based on whether it is correct, and whether it improves the quality of the post. It is under no obligation to maintain the authors intent.