Subjective questions, community wiki. What are the rules and why?

Question

After this question was closed today, and someone suggested a question of mine will become a community wiki, because it's "a poll" (which I'm not sure it is) I'm really confused.
Are there clear criteria as to what should become a CW and what is too subjective for SO?

Are those criteria being followed (unlike the links to questions I gave in the first link, and really a lot more)?

Added: And specifically - Should my question be a CW? Why?
And also, what's so unacceptable in subjective questions? Obviously if someone would ask "Which programming languages is best?" (or something of that sort) That's a useless question, but if you can learn something from a subjective question, and if it's interesting, what's wrong with it? SO is surely not just about solving technical code issues.

More Added: Oh, now I see clearly, my question was closed but at least I got a good explanation as to why subjective questions are evil:

It's impossible to objectively answer this question; questions of this type are too open ended and usually lead to confrontation and argument.

My question was closed because it may lead to "confrontation and argument"????

Thanks.

Answer 1

Reputation on StackOverflow is supposed to be as strict an approximation as possible of your technical contribution to the site. Anything along the lines of "Most useful" is not, strictly speaking, a technical question (especially if it ends with "rule of thumb"), because there is no right answer to the question; it's a matter of opinion. Other questions may also have more than one right answer, but whether or not those answers work is not a matter of opinion; anyone can test them and see.

Your question definitely sets up a poll, in that people will vote on answers in the same way they'd vote in a political election or respond to a political poll. A side-effect of Community Wiki is that is does not give reputation, and so it can mitigate the issue somewhat. Nevertheless, such questions do not really belong on StackOverflow itself; it's not what Stack Overflow is about.

Occasionally a question such as yours will bring in so many good answers or be so useful that it manages to succeed on Stack Overflow anyway. I have one like this myself, from the early days of Stack Overflow before the rules were as clear. These questions are the exception. They are akin to winning the lottery (and hopefully you're smart enough not to play that, too).

Recently there has been an upswing in subjective questions on Stack Overflow, thanks largely to this post on the Stack Overflow blog. Near the bottom of the post there is a collection of six guidelines for a good subjective question. Yours fails on at least three points (#s 2,4, and 5).

Answer 2

No, there are no clear criteria and there are no rules. All efforts to establish them have failed.

You have two choices:

1. Ask the questions on SO and hope they garner enough popular support to be re-opened. If you try this, marking them CW may help your cause, as CW questions don't collect rep and therefore it may seem less like you're just asking poll questions as a way to increase your on-site reputation score.

2. Ask the question on Programmers.SE (and hope they garner enough popular support to be re-opened). P.SE was originally created for these types of questions, as strictly-speaking they are wholly inappropriate for SO... but their current appropriateness for P.SE is also now in question. Regardless, marking them as CW on P.SE will not help you.

Answer 3

Math Overflow has a rather strict policy on these questions, which I summarised on math.sx as follows:

1. Questions that don't look like they will have a single, correct answer should be tagged either [soft-question] if that's because the question isn't precise enough, or [big-list] if that is because we expect the question to invite many different, incomparable answers.
2. These questions should be made Community Wiki.

This policy is of course one that explicitly demands metatagging. It works well for MO.

The issue with demaning CWification is that the whole point of a Q&A site is to attract answers, and the SO model is to use reputation as both some sort of incentive for answering and metric for how we are doing. Having such a class of question means that they are questions whose answers we don't really value. If so, why not close them?

For MO, this isn't really a problem, because the whole place has a rather collegial feel, and the reputation incentive functions mainly as a gate that non-professionals find hard to get through. The real incentive is real reputation, the opinion of your peers. This works because mathematicians meet each other at conferences, serve on tenure review boards, &c.