There's a discussion in the comments on this question about empirical support for commenting code to the effect that "resource request" questions are off-topic. It refers to Why did a question about programming books get closed on Programmers.SE? here on meta.

I've asked on programmers before for empirical evidence in favour of paradigm selection, a question which didn't attract the same response about requesting resources.

Is there a difference between asking for a book/books and for research literature? Should some of these questions be considered on-topic?

| |

My answer, to be treated as a straw man if required:

There is a difference between "is there a canonical book on [x]?", the subject of the question closed in the linked meta discussion, and "are there empirical studies on [x]?".

Answers to questions of the first form are likely to be laundry lists, or single-entry "I like [y]" responses. These are unhelpful; "I like [y]" is not a valuable answer, and "I like [y] becauseā€¦" solicits polling and extended debate which is not valued here. Furthermore, these questions imply the existence of a real question ("how do I do this using [x]" or similar) that is not being asked and answered here, reducing the value of this site.

Answers to questions of the second form give readers information from which to form their own conclusions. If an answer discusses a study, for example "there is [z] published in 1995, which I consider irrelevant becauseā€¦", readers can choose to read [z] and decide for themselves; particularly, they can read [z] in context with the sources presented by other answers. The first-form request for "a canonical book" suggests that the intention is not to read every book, but to be told the one that needs reading.

There are still problems with questions of the second form; the risk of extended debate is not mitigated. See the comments on this question as a demonstration that even whether empirical results have value at all is a contentious issue. There's also the difficulty of choosing an accepted answer; if multiple answers provide multiple, distinct sources then none of them is "the" answer.

| |
  • 1
    I think you've answered you own question quite well. I would temper the second form that you provide by noting the SE isn't an exception path to performing research prior to asking the question. There are some correctly closed questions that fit the second form but are ridiculously simple to answer with a quick google search. Asking for said research && indicating what you've already researched becomes a much more interesting question – user53019 Feb 21 '13 at 18:39
  • 1
    That's a good point, @GlenH7. I think there's a bit of an art to doing literature searches (and Google Scholar isn't always the best place to go), but there is a difference between getting stumped on a search and coming here before trying. Thanks :) – user4051 Feb 21 '13 at 19:05