Proper way to deal with a question that contains an arithmetic error?

In this question Given $d \equiv 1 \pmod 4$, can $\mathcal{O}_{\mathbb{Q}(\sqrt{d})}$ have a fundamental unit that is not a "half-integer"? the OP made some inexplicable arithmetic error that led him to believe that the fundamental unit of $\mathcal{O}_{\textbf{Q}(\sqrt{17})}$ is $\frac{5 + 3 \sqrt{17}}{2}$, which is clearly wrong when you compute the norm. The only answer so far does point out this error, but I feel like it's not enough, as the mistake derailed one of the commenters. But to simply correct the error seems like quite a bit of overstepping. What about appending an editorial note to the end of the question?

• At least for this specific question the asker has realized his mistake and added a note about it yesterday. Jul 25 '15 at 1:58

In the current situation I would not correct the error in the question; the answer corrects it any way. The question (and the answer!) will make no sense anymore if it is corrected. We usually asks questioners not to edit their questions in a way that would invalidate answers; this should go for all edits.

If anything, one can add a note pointing out the error also in the question-post (I do not find this really necessary though; the answer seems clear and visible enough and that comment was written before the answer).

If somebody would have a true aversion against having that error, correcting it would still not be good. Then, one might take out that example, and adapt the answer accordingly. But this seems like some work for not that much in return.

• Agreed. Plus in this particular case, the asker even wrote "If I've done it correctly" in his question, which points up the possibility he might have made a mistake in his calculations.
– user153918
Jul 16 '15 at 15:47

I'm inclined to agree with quid, but neither he nor Alonzo posted answers to that questions (though it could be the case that they, like me, started on an answer but have not posted it). Those who do post answers then feel obligated to point out the mistake (unless they fail to notice it, in which case they go down the wrong path with their answer).

For this reason, I think an unobtrusive note at the end of the question might not be quite such a bad thing. Something like:

Editor's note: The OP made a mistake which prevented him from finding an example of such a domain himself.