It seems we've started the habit to complement particular meta questions with corresponding "test" questions on the main site. See for examples:

As explained by Dave (who should get full credit for this very question - sorry Dave), this isn't how it should work, and such debates should be fully resolved on meta before questions are asked (if allowed).

Of course it's unavoidable when a question on main is what causes the debate on meta, but this deals with the specific opposite case, which is creating a question on main to illustrate a debate on meta. So this is the place to discuss this practice.

  • 2
    Maybe this question should actually belong to meta meta.
    – Joubarc
    Aug 23, 2011 at 7:24
  • Could you (@Dave) please explain your reasoning? I have no idea why you would think “this isn't how it should work”. Aug 23, 2011 at 17:33
  • As I was saying elsewhere, my understanding is based on previous use of betas where this was an undisputed rule. But rules aside, this seems like rather common sense to me in that it is quite annoying to see questions while browsing, only to realise they are "test" and liable to be closed from the start. A basic courtesy to users of the site would be not to ask a question that one knows may be closed right away (thus wasting the time of anybody who tried answering). And this is what Meta is for.
    – Dave
    Aug 24, 2011 at 0:39

4 Answers 4


There are two common ways to handle scope discussions. You can handle it in the "theoretical space", that is, by noting the kind of question on Meta in advance and having a discussion ahead of time. Or, there is the "practical space", in which you run into the question on the main site and it segues into a Meta discussion due to opposition. Over the course of the site growth, you will generally encounter both rather equally, but usually you only intentionally run into the former.

During the beta phases, it's an acceptable thing, as Benjol notes, to test the waters by posting such a question. However, it's highly recommended that you never ask a question for the sake of asking. That is, if you want to test if a particular question type or scope is appropriate for the site, ask if it's something you're actually interested in knowing the answer for and believe belongs on the site. When making a point, make a point with something you care for.

If you are not vested in the question, it is probably more useful to just discuss a theoretical example question on the Meta (basically post a sample question body along with your summary of what the question type is). Otherwise, even if the resolution of the Meta discussion is in the question's favor, you'll be left with a question that you don't care for, and that's not often a useful expenditure of people's time. On the flipside, if it's something you don't think belongs, or to which the conclusion is that it doesn't belong, having to clean up and delete the "sample question" is just extra work that isn't necessary.

Beta is where it is important to hit the most kinds of question types so you can concretely shape your scope, especially before you launch into public beta. Asking good questions, if being constructive, is very conducive to helping shape scope. Once you exit beta into a graduated site, generally these questions that inspire revisiting your scope will come over the natural course of the site and will not need forceful evocation. And if you do have concerns about a potential problem, at that stage it is usually more productive to discuss it on the Meta site directly.


My own feeling is that it's OK to do so.

The reason I feel that way is precisely because it is bound to happen anyway with questions asked in all honesty which will create debates. Maybe it won't happen much, but it will happen.

And since it will, why make a difference on whether the question or the debate came first?

I understand other SE sites may have chosen another path, but do we need to follow them blindly? (Should we use a picture of a lemming for the site logo? Evpok has a nice one.)

That said, questions asked that way should be genuine questions and have a real content value. So far I feel it has been the case, as the "test" questions were all valid (I'm still wondering about this arabi/mosquito thing, by the way).

  • 2
    A lemming? Moi qui croyais que ç'allait être une grenouille... Aug 23, 2011 at 7:54
  • I am too lazy to go sift through meta.SO, but I am pretty sure the Official Rule™ is to never use Main as a testing ground for topics that should be in Meta (discussions on form, discussions on what's on/off topic, help requests, FAQ topics etc.)
    – Dave
    Aug 23, 2011 at 8:43
  • 1
    @Dave Should we threaten you with a needle to cure that lazyness? ;-)
    – Joubarc
    Aug 23, 2011 at 8:57

You seem to assume that just because someone knows that a question is borderline and proactively starts a meta thread that they are not actually interested in the question. This is a wrong assumption.

I agree that people shouldn't ask questions that they aren't really interested in just to make a point, but I haven't seen such a question here.

The fact that someone knows that other people might want to close their question does not obligate them to not ask their question.

Quite the opposite, the people who want to close questions that obviously are liked by some people should start a meta discussion first before the close the question, especially in private beta. If the question asker proactively does your job for you, this is not a reason to complain that they should not ask the question at all.

  • Sorry if I left that impression, I didn't really assume that, no. On the contrary, I do agree that the "test" questions we've been seing so far are genuine questions and that the people who asked them are interested in them. But you're right, there would be a big difference, and I can only assume I'm ok with these particular questions because of that. I someone were to ask a "test" question he wasn't interested in, I might also want to close it. Note though that the "test" aspect does influence the question; witness the German question which could have been asked in French to start with.
    – Joubarc
    Aug 24, 2011 at 9:47

Sorry, I think I'm with Dave on this one. I don't think it's a good idea to knowingly ask borderline questions if you are fully aware that it is something that needs debating on meta.

With the following potential exceptions:

  • During the beta phase, I think it might be acceptable for 'testing the water'
  • Also during the beta phase, I think it might be a sort-of acceptable way of 'pulling' people into the discussion on meta.

You must log in to answer this question.

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