2
$\begingroup$

Sometimes it happens that the answer to a question is contained in the answer to a different question.

Should the former question be closed as a duplicate?

Here's an example, (read the comments). I've encountered more, but I don't think more examples are needed to make my point.

$\endgroup$
9
$\begingroup$

I think that we should be careful when closing as duplicates of more general questions; and that it is quite often not easy to decide.

Even if it is true that the more general question answers the less general one, there might be easier answer to less general question. By closing question we might prevent MSE users from posting an easier answer (which does not apply to the more general question).

Since you posted example with integrals: I would have no problem with closing $\int 2x+1 \,\mathrm{d}x$ as a duplicate of $\int ax+b \,\mathrm{d}x$ or even $\int P(x)\,\mathrm{d}x$, $P$ being a polynomial. But if we a questions asking about $\int P(\sin x)\,\mathrm{d}x$, with answers explaining Weierstrass substitution, we should not close $\int \cos^2x+\sin^2x \,\mathrm{d}x$ as a duplicate. Many integrals including $\cos x$ and $\sin x$ might have much simpler solution than using Weierstrass substitution; if we close all of them as a duplicate of the more general questions, it is true that the more general questions answers them, but we have provided OP with a much more difficult answer than he could have obtained if the questions would not have been closed.

$\endgroup$
  • $\begingroup$ Good point.${{}}$ $\endgroup$ – Git Gud Jun 27 '13 at 5:54
4
$\begingroup$

In principle, I'd say not. It prohibits giving answers to the newer question.

If the other question is linked in the comments, that puts it in the "Linked" section. If more than this is deemed appropriate, perhaps a CW answer pointing to (the relevant part of) the answer to the other question can be posted.

In general, my condition for marking duplicate is:

Any answer to the old question will (mutatis mutandis) answer the new question.

This includes cases where the old question is more general; if OP doesn't understand the answers, they can edit to make this clear, and I'll support reopening (for the nature of the question will have changed to a follow-up question).

Of course, in specific cases I may deviate from the above as I deem fit.

$\endgroup$
  • 1
    $\begingroup$ I think you're contradicting yourself, depending on what exactly you mean with This includes cases where the old question is more general. If question 1 is strictly more general than question 2 and answer 1 answers question 1 perfectly, then answer 1 will not mutatis mutadins answer question 2, as it will exceed it. $\endgroup$ – Git Gud Jun 26 '13 at 20:56
  • $\begingroup$ I meant "more general" as in "What is $\int(a+bx)^m \,\mathrm dx$?" encompassing "What is $\int(1+2x)^4\,\mathrm dx$?". It is precisely because one can argue this to fall outside of mutatis mutandis that I included a remark about it. $\endgroup$ – Lord_Farin Jun 26 '13 at 21:00
  • $\begingroup$ I stand by my statement, I don't think those cases are included in the mutatis mutandis case. $\endgroup$ – Git Gud Jun 26 '13 at 21:02

You must log in to answer this question.

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