There have been a few shopping/recommendation-like question on CogSci.SE lately:

Actually, nearly the whole Software tag falls under the category of these recommendation questions:

Despite the above questions being well accepted, we seem to reject most book recommendations:

Can we make a policy on what questions that take the form of a recommendation we should accept and develop some heuristics? How can we avoid questions that are too localized (they'll become obsolete within a year or are only useful to one person's set of requirements/experience)?

2 Answers 2


I think we should consider the goal of the questions, when determining whether it is off- or on-topic here. I would like to suggest that we make a distinction between goal oriented and tool oriented questions.

Goal oriented questions are typically posed as "how can I measure/analyze phenomenon X?" or "What is a good book to learn about topic Y?". Questions that are posed in this way are on-topic because more than one person can benefit from it. Secondly, more then one answer can be given this way.

Tool oriented questions, on the other hand, are more focused on one book, software or hardware package. These questions are often too specific and will not benefit more than the OP itself. Therefore, tool oriented questions should be considered off-topic.

Note: The goal oriented questions still need to follow the other guidelines of course. A question that is too broad, unclear or self-help still needs to be improved (or otherwise closed).


I propose the following heurisitcs below.

For books:

  • Answers asking how to understand a book are too broad. There are too many books in the world and their notoriety shifts at a rate that makes it unsuitable to answer these questions.
  • Recommendations for books on a certain topic should be changed to reference requests and treated as such.

For software/hardware:

  • Software/hardware for accomplishing a single step of analysis or part of methodology are acceptable, but should still strive for the guidelines indicated in this blog post
  • Tutorial recommendations are too broad, as they fall under the same curse as asking how to understand a book.
  • Questions about how to accomplish a certain goal with software are acceptable, but are unlikely to be answered at this point.
  • Questions asking how to combine hardware inputs for analysis are too broad.
  • Questions asking to compare two softwares/hardwares are too broad.

This should be reviewed after a year of implementation, if adopted.

  • Not fully sure what you mean by: "Answers asking how to understand a book are too broad." In relation to what are you posting this?
    – Steven Jeuris Mod
    Oct 11, 2016 at 23:53
  • This question which asks about "The Mind that Changes Itself".
    – Seanny123
    Oct 12, 2016 at 0:17
  • I think the heuristic can be much easier: is it goal oriented (wanting to learn about topic X, wanting to measure/analyze data type Y) or tool oriented (focused on book, software, or hardware)? In the latter case, it would be too specific and useful for one. In case of the former, there is potential use for others. Oct 12, 2016 at 22:06
  • @RobinKramer that makes more sense! I love the idea of validity being considered on a spectrum between those two poles. If you write that up as an answer, I will accept it.
    – Seanny123
    Oct 12, 2016 at 23:20

You must log in to answer this question.

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