Recently I saw that there is a new Computational Science SE site and it seems to overlap considerably with scientific and statistical computing. Those questions seem to be welcomed on both sites. How do I distinguish where each question goes?

For a motivating example, I recently came up with a question where I couldn't figure out which site was appropriate. I wanted to compare scipy's implementation of PCA with my own. (The question can be found here.) How would I know if this question was better for Cross Validated or computational science?

(Note that I ended up submitting it to Stack Overflow even before these two sites occurred to me. That may have been a mistake on my part. Feel free to contribute more or better examples to distinguish where questions go where.)


1 Answer 1


Note that in some cases a post can be on topic at more than one site. Generally you should just choose the best one (which may relate to the kind of answers you seek).

In other cases a post may be excluded from the scope of one or even both sites, in spite of the fact that you think it should belong there.

Our help/on-topic page discusses the scope of our site in some detail. See the opening section and the later part that mentions programming.

In particular, to be on topic here a question should be statistical in nature, rather than be about implementation details in some language. If you want to compare two algorithms, it might be on topic. If you want to compare two code-blocks, it's likely off topic here.

If your question relates to an error message it's almost certainly off-topic here (with some exceptions where the underlying issue is a clearly statistical one).

  • $\begingroup$ now on reflection, my question is in the comparison of the algorithms. Wether if the difference is in statistical nature, its still unknown to me. Maybe it would have been fine here in stats overflow? $\endgroup$ Commented Aug 6, 2016 at 2:01
  • $\begingroup$ @Charlie with some edits to focus on the algorithms themselves rather than implementation in code, and without discussion of error messages, it's possible that it would be seen on topic (depending on how it was presented). $\endgroup$
    – Glen_b
    Commented Aug 7, 2016 at 2:00
  • 1
    $\begingroup$ @Charlie, if you are talking about this question stackoverflow.com/questions/38799205 then it would have been completely fine here on CV. $\endgroup$
    – amoeba
    Commented Aug 7, 2016 at 22:31
  • $\begingroup$ thanks! @amoeba would it have been better in CV? just curious (trying to learn). $\endgroup$ Commented Aug 7, 2016 at 22:42
  • $\begingroup$ @Charlie What is "better" in this context? Having higher probability of a good answer? I don't know. It would probably get answered here, but I am not following SO and can't say if questions like that get good answers there. I suggest that if you don't get a satisfactorily answer there after some time (a week?), you can re-post it here. [I have even less of an idea about scicomp.SE.] $\endgroup$
    – amoeba
    Commented Aug 7, 2016 at 22:46
  • $\begingroup$ @amoeba I've seen more nearly on-topic/ and less "why do I get this error" type questions close here for being of a similar form. I don't think it's survival here as is would be at all clear cut without some edits to alter the emphasis. $\endgroup$
    – Glen_b
    Commented Aug 8, 2016 at 1:43
  • $\begingroup$ Perhaps. On the other hand, I've seen far more off-topic questions staying open. $\endgroup$
    – amoeba
    Commented Aug 8, 2016 at 9:44
  • $\begingroup$ However an off topic staying open is easy (people fail to notice or fail to act); one closing requires one or more people to conclude it's actually off topic and act on it, so the closed ones give relatively more information about what's considered off topic. $\endgroup$
    – Glen_b
    Commented Aug 8, 2016 at 10:05
  • $\begingroup$ @Glen_b Agreed. But in this specific case there might actually be some misunderstanding: the linked question is not at all of the "why do I get this error" type; OP is not mentioning any software errors at all. There is an "error" mentioned in the question text, but it refers to the "reconstruction error". $\endgroup$
    – amoeba
    Commented Aug 8, 2016 at 11:54
  • $\begingroup$ Thanks; I think you're right. $\endgroup$
    – Glen_b
    Commented Aug 10, 2016 at 11:50

You must log in to answer this question.

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