Related, but not exactly the same, to this question: "Should I rewrite my question, or create a new one?"

Recently I gave a "quick and dirty" answer (in fact I just googled for getting it) to this interesting question. The asker reviewed it and commented very politely on why it misses to answer the question, without downvoting (or insulting me, which perhaps it would have been the case...).

After that, I remembered some interesting facts on simple iterations and fixed point of holomorphic functions in a single variable, and worked out an entirely different answer. Meanwhile another user gave the question a very interesting answer, and they cite the paper I cited: so, quoting exactly the title of this question,

Should I give a new answer or change the old one?

Note that my answer was not downvoted nor upvoted, so there would be no harm in deleting it and, for example, adding its reference to the other answer.

  • 6
    $\begingroup$ If you change the old answer, try to preserve the previous version so the comment makes sense. At the very least, acknowledge that your new version addresses the concerns of the questioners comment as well as the question. Gerhard "We Can Learn From History" Paseman, 2020.02.17. $\endgroup$ Feb 17, 2020 at 16:55
  • 5
    $\begingroup$ "Best practice" according to SE philosophy is probably to edit the old answer; you could do this by having the old answer above a fold (type three dashes on a line after the old) and then start the new response on a line below the three dashes. Or, you could buck the philosophy and make a second post if you'd like; usually no one here at MO minds. In that case, if you don't want to receive "credit" for the quick and dirty answer, you could make it community wiki. $\endgroup$
    – Todd Trimble Mod
    Feb 17, 2020 at 17:17
  • 3
    $\begingroup$ I think replying to the comment that you completely rewrote the answer is enough without trying to preserve the previous answer if you think it's not worthwhile (one can always go into history of answers). $\endgroup$
    – YCor
    Feb 17, 2020 at 17:51
  • $\begingroup$ I thought @ToddTrimble's advice was actually contra the general SE philosophy of having answers in their final form, but can't find documentation of that. I did find this old MSE question, which seems to be essentially a duplicate: meta.stackexchange.com/questions/40318/… . $\endgroup$
    – LSpice
    Feb 24, 2020 at 0:28
  • $\begingroup$ @LSpice Whenever one attempts to add a second answer on top of an answer one has already given, a pop-up box asks whether you're sure you want to do this instead of editing the old one. I'm not sure where you're getting that general SE philosophy from -- it sounds like a discouragement from editing. $\endgroup$
    – Todd Trimble Mod
    Feb 24, 2020 at 0:54
  • 1
    $\begingroup$ @ToddTrimble, I don't mean no editing, I mean that (I think) the guideline is to have the answer after editing look like "Statement X is true.", rather than "Statement X is false. EDIT: Actually, it is true." I can't find documentation for this guideline, so I might be wrong, but I thought it was the SE best practice (even though I don't like it, because it has the effect of invalidating comments). $\endgroup$
    – LSpice
    Feb 24, 2020 at 1:24
  • $\begingroup$ @LSpice In that case, let me clarify by suggesting you should read my first comment in the light of my second comment. I don't think I was really suggesting anything pro or con this alleged SE guideline (and actually, I'm not seeing how my suggestion for this particular case would result in the rather-than scenario you described, but I think I've said enough on this for now). $\endgroup$
    – Todd Trimble Mod
    Feb 24, 2020 at 1:41


You must log in to answer this question.

Browse other questions tagged .