Answering your own question, but don't accept answer

Question

Is it bad etiquette to answer your own question, but then not accept that answer? I realize there have been numerous questions about answering one's own question, but I'm specifically wondering about not accept one's own answer for the sake of attracting better answers.

Is there any consensus on whether this good etiquette?

---

Example

I asked a question over on the emacs SE, and later on found a workaround for the problem based on another 4-year-old SE answer. I then updated the question with this new information. Although this included information solves the original problem I had, it doesn't actually answer the question I had.

A well-respected member of the community later asked that I move the addendum to an answer and accept it.

This is a variant of case 2 in etiquette for answering your own question:

Provide the answer in the question and ask the question phrased as "is there a better answer". This allows others to gain reputation, and doesn't look so greedy but makes your answer harder to vote on.

Except in this case, the answer I'm providing is something I'm not really happy with; I don't accept my own answer.

This answer is in favor of this approach:

I think #2 is a strong answer, respectful of the SO community.

Asking a question you know the answer to (especially if you answer int he same breath) is of dubious value and intent, but if you approach this with the mindset that what you think you "know" might not really be the best answer, I think you'll be illuminated most of the time.

But then a comment to this answer has an equal number of votes says:

#2 is a bad idea. What if the asker's answer is a good answer? Then we end up with an answer in the question, and no answer as an answer. It completely abuses the format of Stack Exchange.

Answer 1

I've had the same angst as you about whether people would look at my posts as some sort of attempt to generate somehow-illegitimate rep. There will always be people who are suspicious of the motives of others, and SO users looking at possible rep gain are no exception. I wouldn't try to tailor or optimize my behavior to appease those people.

To address your broader question, I think it comes down to a judgment call. There is no bright line separating "I have a workaround but it's not really a good solution" and "this technically solves the problem but I want a better answer"; it's a very gray area. One person's "this should be in the question as a 'what I've already tried'" is another person's "this should be a self-answer".

And finally, to answer the question you initially asked, I don't consider it bad etiquette to self-answer and not accept the self-answer quickly (or even ever; it's unusual but entirely possible for answer to be helpful enough to post yet still not quite complete enough to be accepted). I actually endorse this behavior a good signal that you're looking for additional answers. If you want to be extra clear, you might post a comment to that effect when you answer that way, explaining what specific additional info you're looking for in an additional answer, or what about your self-answer you think is imperfect.

Answer 2

I've answered several of my own questions (e.g. here, here, here, and here). One time I noted in the question that I had a potential answer that wasn't very satisfying and I would post it as an answer. I posted it as an answer but I was hoping someone else would come up with a better answer. The rest of the times I only came up with the answer after asking the question. One time I accepted my own answer; the other times I did not. The reason why I accepted one and not the others is the same reason why I accept some answers from other people and don't accept other answers. If an answer is demonstrably correct, or completely addresses my question in a satisfactory way, I accept it. If an answer is useful but does not entirely answer the question, I upvote it and hope for a better answer to come along. If an answer is not useful or is actually incorrect then I downvote it and hope for a better answer to come along.

In short, I think the "accepted answer" should always be based on the content of answers. It makes no difference who wrote the answer.

One exception, perhaps, is where a different answer has so many more upvotes than my answer. Even if I think my answer is better, I would be hesitant to accept it in such a case because it just looks like I am acting biasedly (e.g. here.)

Answer 3

Is it bad etiquette to answer your own question, but then not accept that answer?

I'd say it's bad etiquette to answer your own question. I've seen this happen time and time again, and I don't like it. Nor do other people I've spoken to. I'd also say it's even worse etiquette to then accept your own answer.

I realize there have been numerous questions about answering one's own question, but I'm specifically wondering about not accept one's own answer for the sake of attracting better answers.

Sounds good to me. Accepting an answer discourages any new answers.

Is there any consensus on whether this good etiquette?

This makes me think about morality. I know it's wrong to beat up some little kid, and I'm not listening to any "consensus" on that. I know it's wrong, because I wouldn't like it to happen to me. In similar vein I'm not listening to any consensus on your question. If I answer a question and then the guy who asked it answers his own question, I feel cheated. Especially if he/she then accepts it as the best answer.

I asked a question over on the Emacs SE, and later on found a workaround for the problem based on another 4-year-old SE answer. I then updated the question with this new information. Although this included information solves the original problem I had, it doesn't actually answer the question I had.

Sounds good to me.

A well-respected member of the community later asked that I move the addendum to an answer and accept it.

Who?

This is a variant of case 2 in etiquette for answering your own question... Provide the answer in the question and ask the question phrased as "is there a better answer". This allows others to gain reputation and doesn't look so greedy but makes your answer harder to vote on.

It says provide the answer in the question.

Except in this case, the answer I'm providing is something I'm not really happy with; I don't accept my own answer.

Good. Even if you are happy with the answer you provided, I don't think you should accept it.

This answer is in favor of this approach:

I don't care. I think Stack Exchange is too prey to "consensus" anyway. Science is not a democracy. All the upvotes in the world won't make a wrong answer right. Upvotes are used to indicate that one answer or option is better than another, but the truth of it might be just that one guy has got a lot of friends. For all you know, they might even be friends like these:

LOL! In similar vein all the downvotes in the world won't make a right answer wrong.

But then a comment to this answer has an equal number of votes says "#2 is a bad idea. What if the asker's answer is a good answer? Then we end up with an answer in the question, and no answer as an answer. It completely abuses the format of Stack Exchange".

Phooey. Answering your own question and then accepting your own answer completely abuses the format of Stack Exchange. Because then other people won't want to answer your future questions. If everybody did that the website would die a death.