My question was closed. Robert Harvey says:

Questions asking for links to external resources are more or less off-topic throughout the Stack Exchange network. That's what search engines are for.

  • "Reference request" is the third most popular tag on cstheory.SE
  • It's is one of the 20 most popular tags on math.SE with over 2,000 questions
  • 114 questions for "reference" on stats.SE
  • Indeed, with the exception of SO itself all of the *.SE I regularly visit have a popular "reference request" tag
  • Even on this site itself, 23 questions have been asked with the "reference" tag, only two of which were closed.

There are questions for which answers not backed by empirical data are fine. But if I want to see a peer-reviewed study instead of someone's personal experience, why is that off topic for this site?

Note there are related questions:

It strikes me that those deal with broad things like "give me a book about C" whereas I'm asking for support of a very specific hypothesis (the relationship between testability and stability).

  • "for... conceptual questions about software development" (faq). How are reference requests "conceptual"?
    – gnat
    Commented Nov 30, 2012 at 16:35
  • @gnat: If the reference was conceptual, I suppose? My point is: I don't want someone's personal experience, I want peer-reviewed data. Is there a better way to phrase this?
    – Xodarap
    Commented Nov 30, 2012 at 18:33
  • what is conceptual in the references?
    – gnat
    Commented Nov 30, 2012 at 18:34
  • 1
    @gnat: we have a question that you agree is on topic. We now append the phrase "please cite evidence." The question has now become non-conceptual?
    – Xodarap
    Commented Nov 30, 2012 at 19:36
  • my comments, as well as your question were about original version in rev 1. Now that things changed, this should be reconsidered
    – gnat
    Commented Nov 30, 2012 at 19:43

1 Answer 1


The first thing to consider is that within the (usually) broad guidelines, the communities of Stack Exchange sites have some ability to define what is acceptable for their community. On a couple of occasions, the community of Programmers has decided that simply asking for resources is not a suitable question for the site. Just because other sites allow for reference requests or resource requests does that mean every site has to.

There are two driving factors against allowing resource requests:

The FAQ states that:

You should only ask practical, answerable questions based on actual problems that you face. Chatty, open-ended questions diminish the usefulness of our site and push other questions off the front page.

The close reason for Not Constructive states:

As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or specific expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, see the FAQ for guidance.

Questions that simply ask for resources are not about solving a problem nor do they allow for answers to be supported by facts or expertise that requires human thought to produce. Searching through libraries or databases for documents is something best done by a search engine, not for a question on a Stack Exchange site. Questions should require some kind of human thought, experiences, or expertise in addition to being able to be backed up by a specific situation or reputable resource.

For your specific question, it seems fairly self-evident that it's true. Especially considering that you say that it's an assertion made in several papers. If no citation is provided, that generally means it's considered to be common knowledge. Of course testable code is going to be more stable - testability is often related to the ability of developers to understand the code as well as be able to write tests. Even if no tests are written, code that is easy to understand can be read and problems seen and fixed quickly, making the system more stable. I highly doubt you'll find this studied because it is evident to anyone with any experience in software development.

  • "nor do they allow for answers to be supported by facts" - That's all I'm asking for! My phrasing was poor, but I don't want an answer like "in my experience..." as I will get answers like you gave ("It's obvious."). I want something like "component coupling is correlated with defect rate at rho = .8". Is there a better way to phrase this?
    – Xodarap
    Commented Nov 30, 2012 at 18:31
  • @Xodarap The key is "supported by". However, there are unlikely to be any of the type of facts that you want for something that is so clearly obvious.
    – Thomas Owens Mod
    Commented Nov 30, 2012 at 18:34
  • Component coupling is measurable, defect rate is measurable, and correlation is measurable. In fact, these are measured extremely frequently. Why do you think it's unlikely that data exists?
    – Xodarap
    Commented Nov 30, 2012 at 18:42
  • PS: I have five hours of dev time. Should I spend it refactoring the code to make it more testable, or fixing an existing bug? The answer may be "obvious" to you, but it's not to me, nor most developers.
    – Xodarap
    Commented Nov 30, 2012 at 18:44
  • @Xodarap I know those are all measurable - my background is in software measurements and empirical software engineering. Your root question is still "Is testable code is actually more stable?" The answer is common knowledge - I did a quick search of the IEEE and ACM databases, and it doesn't seem like there's anything that addresses this because it's common knowledge. If you're looking for a relationship between a given metric or set of metrics and defects or defect density, ask about those specifically. Also, you can't restrict people to only providing references.
    – Thomas Owens Mod
    Commented Nov 30, 2012 at 18:47
  • 2
    @Xodarap Also, that question is a problem that can be solved. Another example of a good question. "Is testable code more stable?" is a bad question for a Stack Exchange site - it's obvious to any professional. "Are there metrics that can be used to predict defects?" is a good question. "Given my time constraints, should I increase testability or fix a defect?" is also a good question since it's about a problem that lots of people may have.
    – Thomas Owens Mod
    Commented Nov 30, 2012 at 18:48
  • I'm certainly open to the criticism that my question wasn't clear, and should've been closed as too vague. I just don't understand why it would be off topic.
    – Xodarap
    Commented Nov 30, 2012 at 19:35
  • 2
    @Xodarap It was flagged for off-topic in addition to the vote to close. Probably just a misclick on my part - it's not off-topic, but it still seems to be not constructive (or at least not well defined) at this point. If it gets flagged or close votes, it will likely be closed again (but easily reopened). If you need more help editing it, jump into chat and I'm sure someone will help you make it better.
    – Thomas Owens Mod
    Commented Nov 30, 2012 at 19:37
  • Searching through libraries or databases for documents is something best done by a search engine, not for a question on a Stack Exchange site. Catch 22 when you consider that the search engine most often points us to stack exchanges.
    – Anon
    Commented Oct 1, 2021 at 9:39

You must log in to answer this question.

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