See this question: History of Worldbuilding - The Flying City
Should questions such as this be considered on or off topic?
The question asks about the first date when a concept was put in a story. That's not worldbuilding, but plain literature history. If I have a question about literature history I ask it on the pertinent SE community.
That the OP states that it helps them in worldbuilding, doesn't make it a worldbuilding question, same as asking "how do I conjugate to be in space?" remains a grammar question and not a space exploration question.
To discuss it's relevance we should look at what the help center says is on topic for the site:
Worldbuilding Stack Exchange is a site for designers, writers, artists, gamers and enthusiasts to get help creating imaginary worlds.
[...] When asking questions keep in mind that the goal of the site is to help you build your world, not to tell your story.
[...] If you are looking for discussion, brainstorming, or an overall process rather than specific questions and answers, the Worldbuilding Stack Exchange might not be a good place for your question.
The goal of the site is to help build your world, not to look for recommended readings/movies on the subject. Instead, this is brainstorming ideas for your world, which is not on topic.
Here are some examples of questions that would be relevant to worldbuilding:
To summarize, a question about the history of a subject used in different worlds does not belong on Worldbuilding.SE, instead it belongs on SciFi.SE, more specifically its belongs on SF&F.SE under the history-of tag.
Worldbuilding requires expertise in a variety of fields. People need to know
But what it does not require is knowledge of science fiction and fantasy. However, there is a stack that requires expertise in science fiction and fantasy, where this question would be on-topic. There is no reason to extend our charter to cover something that is already covered elsewhere.
In general, when someone asks a questions that might be on-topic elsewhere but still related to worldbuilding, I think that it should be on-topic here. This is especially true when it is only questionably on-topic elsewhere. Because elsewhere, things may not be on-topic because they are purely speculative, with no real world analog. But here, speculative, even fictional, is our reason for existence.
In this case, we have a question that is for authors but is nothing about making a reasonable, believable world. In fact, it is specifically trying to find reasonable, believable worlds and do it a different way. We are not a one stop shop for anything related to writing. There is already Writers.SE and we are increasingly pushing away valid worldbuilding questions as Too Broad or Primarily Opinion Based. If we want more questions, we'd be better off adjusting the limits there.
This is amenable to a simple bright line rule that puts all these questions on one stack, which is preferable. There is no circumstance under which a question that is not of sufficient focus on that stack should be on-topic here.
Questions about how things were done in previous science fiction works are not on topic on Worldbuilding. The single exception is if the question is not about how it was done but how it could have been done. E.g. how could my story have a realistic light saber? That's on-topic here. But a question about how light sabers work in the Star Wars universe is not.
This rule cuts out this question entirely. But everything it cuts out would be on-topic on SFF.SE. There may be some questions that are Too Broad or opinion-based for either stack.
There Must be Lines in the Sand
If you went over to Stack Overflow and asked this question, formatted for their site. I.E., "Are History of Programming Questions On-Topic?" You would get a clear and resounding "NO!" This, despite the fact that the history of programming is very much a part of programming.
But it also clogs the site with questions that have very low value to the community. On an ultra-creative site like this, we often forget that one of Stack Exchange's mandates is that questions and answers should benefit everybody, not just the OP — and the claim "everybody can benefit from History of X questions" is simply fallacious.
Based on the basic idea that Stack Exchange is a group of (usually) happy volunteers with the goal of helping the community as a whole: both those who are here today and those who may come tomorrow...
If the question's value is basically for the OP alone, it shouldn't have been asked here. We are not here to be anyone's personal research assistant.
If the question has an incredibly narrow application, maybe it shouldn't have been asked here. This usually means the OP hasn't taken the time to re-scope the question to make it applicable to more people than him/herself.
If the question isn't directly contributing to the primary purpose of the site (if you have to start making arguments about why it's on-topic), then it probably shouldn't be asked here. The site really can't be everything to everyone. That's what Quora and Reddit are for and why they're so hard for anyone other than the original OP to use. Those sites are basically set up as individual research projects and whether or not anyone else is benefited is irrelevant. That's not what we do.
Having said that, is the current version of Chasly's question off-topic?
Yes, it is. But not for the reasons we're discussing here. It's too broad. SE's focus is one-specific-question/one-best-answer. A "best answer" can only be one that has a complete (or at least reasonably complete) history of floating cities in literature. A bunch of answers each contributing a piece of the puzzle is unacceptable because it breaks SE's basic mold.
In fact, if it weren't too broad, what Chasly's asking for is pretty good
Chasly's asking for a canonical answer. It just happens to be a HUGE canonical answer. Because it is the nature of people to post something, the question is likely to draw a lot of nearly useless answers: the antithesis of what SE wants.
I think asking about worldbuilding history falls on the other side of the line. Perhaps by only a smidgen, but it's there, nonetheless. The questions will (IMO) always tend to be too broad. Therefore, as a class of question, I vote no.