When are too many questions in a single post too many questions?

Question

Working through the close votes review queue, I just failed an audit for the question Lie group cohomology: a jungle of conditions, which I voted to close as In need of Focus whereas apparently the correct thing to do would have been to vote to leave open.

This leaves me somewhat puzzled. After all, the post in question asks, in order:

1. "Similarly, we look for central extensions of $G$ by $A$, where $A$ is an abelian topological group. What classifies these?"
2. "Moreover, in the case of $H^2_{\mathrm{meas}}(G, A)$, what are the $σ$ algebras with respect to which measurability is imposed?"
3. "Which classifies Lie group extensions of $G$ by $A$, where $A$ is assumed to be an abelian Lie group out of these two? Are there other definitions of Lie group cohomology? Which definition classifies precisely what?"
4. "Another important question: for Lie groups $G$, does any of the two definitions above relate to singular or deRham cohomology?"
5. "Last, but not least, is there a possibility of extending these definitions (all of them, or any of them) to the case where $A$ is not abelian?"

At the risk of stating the obvious, that's more than one question, even if generously interpreted.

Now, the post in question is also rather well received; at the time of writing, it has 7 upvotes and an accepted answer (and I agree that the question is good, besides this issue). I would then like to ask what the correct stance / course of action here is: Am I missing something? Does the popularity of the question override concerns about site policy?

(Note that my concern here is not with "how to avoid failing audits like this" but rather with how to behave around questions like this in general.)

Answer 1

I agree with you. I would have voted in the same way. That sounds like it was a bad audit. Unfortunately, bad audits are a fact of life (sigh).

The usual rule on Stack Exchange is "one question per post".

There are several reasons for this rule. We want to build an archive of knowledge that will contribute to knowledge on the Internet, in the form of questions and answers. We also want to keep a reasonable load on answerers, so we need questions to be reasonably scoped; as a rough rule of thumb, in most cases we hope questions can be answered in a few paragraphs. When a post contains multiple questions, it's often asking too much of answerers. Also, our experience suggests that it can lead to bad results when there are multiple questions per post. Perhaps someone writes an answer that answers some of the questions but not others, and that answer is upvoted and accepted. Now the question is treated by our system as answered (e.g., for purposes of search), even though some of the questions haven't been answered (making it harder for those who search for unanswered questions to find). Also, what if someone else is also wondering about one of those questions, finds the page on this site by Internet search, and then discovers that the one of the questions they had wasn't answered by any of the answers? That's a lousy experience.

See Can I ask only one question per post? and Multi-Question Questions: How Best to Handle Them? for more details and reasoning.

That said, the close vote queue is managed by the community -- i.e., by people. It reflects the opinion of people who participate in reviewing questions in the queue. Sometimes people have different values or opinions. I have shared my opinion, but others here might well have a different view about what is best for the site, and vote accordingly. And sometimes people make mistakes or are unaware of site guidelines.

Answer 2

Although I've no expertise in the topic, the question looks to me as if it can be summarised as "here is a thing I understand in the finite dimensional case, and my question is how it extends to the general case. Here is a list of specific things that I don't find obvious, which I hope a good answer would resolve."

It is perhaps unfortunate for the OP that they decided to delimit this list with question marks rather than bullet points or semicolons etc., but that doesn't really change the fact that it's ultimately one question. Although I could be wrong, it feels to me that attempting to split it into multiple questions would just be busy work, resulting in a bunch of less clear questions whose answers wouldn't be as helpful.

Note that both the SE guidelines and the math.SE ones say quite explicitly that you can have closely related sub-questions in the same post. They prohibit multiple unrelated questions in the same post, but sufficiently closely related sub-questions have never been against policy.

It would be unproductive to turn a good principle (one question per post) into a bureaucratic box ticking rule (one question mark per post). Most users can't possibly know about such a one question mark rule - especially since it's not even official policy - and it leads otherwise good questions to be closed for no reason. I think we should regard it as incumbent on close voters to apply the rules in a principled way, and not on question askers to spend hours reading meta before they post anything.