21

Say I've asked a question to which someone added an answer they put quite some effort into. At first glance the answer seemed to perfectly answer my question, so I accepted it. Now, after using the approach in my real code, it turned out there's a major flaw in it and the answer, in its current state, actually doesn't answer my question. Moreover, properly implementing a solution seems even harder than expected at the beginning.

Is it fair in this case to "unaccept" the answer as long as the problems in it still persists? The fault of early acception is clearly on my side here, so "stealing" credits back from the answerer could be considered unfair behavior. What would be the best practice here?

  • 5
    Just ask a new question with the previous one linked. – CarLaTeX Jul 12 at 19:56
  • 15
    it depends on your question. if your question was flawed and e.g. didn't specify an important requirement it would be quite unfair to add this requirement in an after-thought and so nullify a valid answer. Then as CarLaTeX write better ask a new question. If the answer is simply faulty you can naturally unaccept it. – Ulrike Fischer Jul 12 at 23:16
  • 6
    I think this depends a lot on whether the unaccept is based on your having to change the question itself. It's really not very good to change the question when answers have already been given, especially (as you describe) your question has changed substantially. But if the question doesn't change substantially, then it may be fine to unaccept with a comment explaining why. – Alan Munn Jul 13 at 2:35
  • 3
    It depends on the case. If the original answer is self-consistent (and may potentially useful for some), do as @CarLaTeX suggested: ask a follow-up question. If, on the other hand the answer had an inherent flaw that you just didn't catch, and may not be useful to anyone, then you unaccept it IMHO. – user121799 Jul 13 at 5:38
  • Btw, the question I'm referring to is this one. In this specific case I accepted the other answer, incidentally by the same author, which uses a slower approach but meats the specifications of my question. The assumption actually is that the specifications of the question didn't change. – siracusa Jul 15 at 16:00
19

Let's assume the real problems are not with your query (say, unclear wording) but solely with the answer you accepted prematurely, before you discovered the flaws.

Before un-accepting that answer, I believe that sheer politeness and good manners require leaving a comment first, detailing the flaws you've discovered and asking the answer-giver if it's possible to revise his/her answer. Be prepared to receive, and reply to, follow-up questions from the answer-giver as he/she grapples with the new information and tries to figure out how to best fix the flaws.

If you don't receive a reply and/or an improved answer within 48 hours or so, I suppose it would be ok to unaccept. Remember that people travel at times and might not have access to their computers at all times. It would be grossly unfair to set a tight deadline that they simply cannot meet.

Do contemplate the possibility that a rapid un-accept, done before the answer-giver has had a reasonable amount of time to think about improving his/her answer, might well alienate and antagonize that answer-giver. The answer-giver's reaction might well be not only to ignore your request for an improved answer, but any and all questions you might pose in the future on TeX.SE.

I've experienced many instances of OPs contacting me after they had already accepted an answer of mine, asking for an extension and, yes, sometimes a correction. I almost invariably address these follow-up requests with a positive attitude, even if the follow-up is mostly caused by unclear wording in the initial query. But here's how I handle what I consider to be impolite, inconsiderate, and sometimes downright belligerent [!] requests for follow-up work: I simply ignore the person's request, except maybe to delete the answer entirely (to make it less straightforward for them to continue their harassing behavior), and I make a note to myself to ignore any and all queries this particular user may post in the future, ever, no matter the merits of the query. Life is too short to waste time on being harassed by bullies.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .