Scoping worldbuilding

Science fiction and fantasy stories demand worldbuilding, but a world is an enormous thing where it's only a slight exaggeration to say everything affects everything else—so where the hell do I start?

Well, I could start anywhere. Wherever convenient, wherever's fun. If I have an idea already, start there and fan out. That's always how I worked with fanfiction. Although, in that context I had a lot of existing inspirational material to start from. But I expect it would also work outside that context.

How to know when's enough to start outlining or writing the story? I have no idea, but maybe it's useful to work through how worldbuilding ties into the outlining/writing process:

1. The world includes the magic systems. Hard magic systems affect how the characters can solve problems, which affects how I have to set up the plot. Soft magic systems tie into the characters, settings, vibe, and theme. Both affect the kinds of settings and characters and events that will make sense in the world.
2. The world includes the setting(s). That is most of what the world is, besides the magic system. Note I take setting to also include flora and fauna, organizations, institutions, and so on.
3. The world tie into the characters. The characters interact with the people, places, institutions, and systems of their world. That interaction changes both parties.
4. The world ties into the plot/story, big time. The characters behave according to how they have shaped/been shaped by the world. The settings may matter to the plot. The magic system matters to the plot for obvious reasons.
5. The world ideally ties into the theme. Characters, settings, and magic systems all offer ways of touching or commenting on the theme.

So, if I need to sort out anything first, it's the magic system. But that's a huge project on its own, because I need to specify a theoretically enormous number of interactions in the world if I want it to work in a consistent, reasonable way. There's too much stuff in the world to extend the magic system to cover everything and specify it down to a law of physics. How do I scope that?

Maybe the end goal is "if I give it to Munchkin George, he won't be able to break in less than three days." That is, Munchkin George a specific person who is happy to look at the system and try to break it. Munchkin George could also theoretically be a group. So: Do my best, wait a day, and if on rereading the rules, I can't see an obvious way to break it, send it to Munchkin George. (This requires finding a Munchkin George who likes breaking magic systems. That's a separate problem to solve.) Repeat until Munchkin George stops breaking things.

That's assuming that a non-broken magic system is important for the project. I'd really like to rely on non-broken magic systems. I think it gives the world a subtly unified character when one can rely on the magic system not breaking down on edge cases—even if the reader is never going to notice the edge cases. Blindsight (hard science fiction) and Worth the Candle (~hard fantasy) in their world and magic systems feel meticulously constructed. They include not just the first-order effects of their magic systems but second-order ones and probably third-order as well—that's what makes their worlds feel like they do. They can afford to have those effects because none of them is going to blow up the world—that's the value of a non-broken magic system.

But this seems like a lot of work to put in. Who sends their magic system out for review before even starting to write their story? It's a lot more coordination and overhead and increases the bus factor of my writing process.

Anyway, there's a bigger problem with this: I've been assuming I shouldn't start writing until I first create the world. That means if I'm going to practice writing every day, and I more or less do, I can't practice it with a story where I'm still trying to put together the world. I generally only have room for one story in my head at a time; I can try to stretch it out but one of them's going to get the shaft and both will suffer.

The alternative is to write before I figure out any of the details. Then I'm either anchoring myself to whatever I come up with on the fly while trying to tell a good story, or risking that I'll have to rewrite the whole damn thing once I figure out the world.

I have a feeling I'm overthinking it. Sketch ideas and write in parallel. If I create any inconsistencies, write them down to fix later and write as if I already fixed them. If I have anything I want to backpropagate into the story, write it down. Maybe this approach even works.

It feels unprincipled and sloppy. But the current situation is that when I write science fiction or fantasy I don't write from any kind of worldbuilding, so sloppy's what I have already. This approach is if anything less sloppy.

I suppose it's worth a try.