Possible Duplicate:
Does it make sense to answer old questions?

I'm asking in general, but this question is what made me think about it. The accepted answer is perfectly correct, but since Python 3's os.makedirs function has the exists_ok flag that the python 2.x function does not.

This is the basic rationale I thought up for each position:

ANSWER: Since this is the first question that comes up for a google search for "python mkdir if not exist" and similar queries, it should have new information included. Otherwise, when similar questions are posted, users will close them as duplicates referring to this question, and the new information is never provided.

DON'T ANSWER: This question is quite old and thorough, so users are probably safe being directed to it. The relevant information could still be added as a comment.

Thoughts? I saw this question, but since the accepted answer is still correct, although not 100% complete, I still wanted to ask.

  • Important info really shouldn't be left as a comment unless it's just a minor correction to an existing answer
    – Ben Brocka
    Jul 17, 2012 at 15:55
  • I would propose asking a new question stating the version difference and self answering Jul 17, 2012 at 16:12
  • @SerkanArıkuşu I'm skeptical about that option, since I think I would most likely get results telling me to the read the documentation (which is fair, because it's a simple problem to solve by doing so). Jul 17, 2012 at 16:14
  • @RicardoAltamirano If the information is all readily avaliable in documentation is it not showing up on the first page of the relevant query results? If it is, there's no need for a SO Q/A for it. If not, then self-answering a new question would benefit the internets. Same logic applies to answering the old one too.
    – Servy
    Jul 17, 2012 at 16:20
  • @Servy In my search, only the Python 2 documentation comes up in the first page, not the Python 3 documentation. Numerous duplicate SO questions come up, and I'm on the fourth page of search results and the python 3 documentation still has not appeared. Jul 17, 2012 at 16:22
  • Numerous existing questions touch on this topic to varying degrees.
    – Pops
    Jul 17, 2012 at 16:30

1 Answer 1


Behold, the Decision-Making Filter™:

  • If you think that your answer can help others as a solution to the problem, absolutely you should answer. One of the goals of Stack Exchange is to create an encyclopedia-like repository of information, that can be updated as the world around it changes.

  • If you don't really have an answer to the problem, but you think it's worth noting that something has significantly changed, edit the accepted answer.

  • If you think that your information is really just an afterthought, something that people should know exists, but isn't so significant that it's worth editing the answer, comment on the answer.

  • If you've gotten to this point in the Decision-Making Filter™, then your information is probably not worth attaching to the answer at all.

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