I know about similar requests, but after stumbling over this question (just an example; question poster answered himself using edits instead of an answer) I thought you should at least add an option in the flagging menu like "problem is solved in comments/edits". Then, moderators (or high-rep users) should be able to check this and maybe accept the question itself or convert the edit/comment into an answer that could at least get upvoted to show it is solved. What does the community/staff think about this?
Like you mention, the question has been asked before, and there are good reasons not to do that:
- If you only look at the huge stock of questions, waiting to review, you can imagine what happens if you add another category of jobs to be done. Judging if comments are good "candidates" to become an answer, editing them into an answer, those are mind- intensive tasks and decisions. Even more than reviewing, where you only need to read properly.
- It places moderators in a vulnerable position; even if they are of the highest integrity (which I am sure they are), they could be accused of being subjective, since deciding if an answer or comment should be an accepted answer to someone else's question, is subjective by definition.
- It would create two "classes" of acceptance, questions which are "really" accepted, and cases in which an answer or comment is promoted to an accepted answer. Not really making things clearer.
- It pollutes the voting and reputation system.
We all know good answers are neglected frequently by the original poster. To decide whether I upvote a question or not, I take into account how a poster looks after his question after he or she posted it. I'd rather see an (initially) mediocre question, edited during the process into a good question, and a poster communicating in a clear and alert way, then a seemingly good question, but no one can tell if the answer(s) were useful to OP or not.
I am afraid taking over the job from OP is not improving the process, and does not contribute to clarity at all.