"You could try to get me to invent a magic system that suits you," Castle suggests. "But I can understand if you'd rather steer clear."

"That's tempting but I wasn't confident enough in your interest in helping me to ask. It seems like it would be a major undertaking."

"It would be an interesting undertaking and I like those," she says. "And I'd come out of it with another magic system, if it worked, which is no small thing."

"This is true. Where would be the best place to start?"

"Well," she says. "You could describe the magic system you already know, and what you've learned about the local ones, and what features you like and dislike about each of the above."

"Sorcery works by observing the setting and target and concentrating on how one wants it changed - and if the harmonics are loopy by mapping those and counting that under 'observing the setting', although you can dispense with that part if you're good enough at the rest. It rewards working on one's home turf. It can do things like growing plants, invisibility, setting things on fire, healing, turning people into frogs, transmuting materials, candying dewdrops, wards, gates, and traps. The local one or ones appear to be made of mud and sticks and glitter, require dramatically more per-spell but less per-location preparatory work and generally be more flexible about the prerequisites for a casting, and focus almost entirely on combat or at least conflict applications. If that was not editorialized enough to say what features I like, do say so."

"I might benefit from more editorializing, yes."

"Sorcery is prettily simple. I do need to look up how to do something I've never done before, it took me weeks to learn to make gates - and the notation for harmonic mapping is a nightmare - but the basic concept is straightforward: know what you're doing and what you're doing it to, and it'll happen, no materials or awkward sacrifices or treating with third-party entities or particular risk of backfire. Unfortunately, it has some major gaps in its possible applications - fairies have been working on sorcery for a long time and new advances are rare and esoteric in recorded memory; it is likely that we know already how to do all of the things the system can do, at least without cheating via harmonics in some way. It is difficult to use in unfamiliar territory, on short notice, in opposing someone who thinks faster, or in opposition to a specialized fairy kind magic; I don't know how it stacks up against mud-and-sticks-and-glitter. It can't do anything about orders except that if you know someone well enough to do mental sorcery on them you can make them forget their orders and wind up in a loop trying to recall what they were, and similarly mental sorcery could delete a name from someone's head."

"So if you wanted to invent a magic system, you'd be looking for something that works for you like chess magic works for me," Castle suggests. "Works, has the flexibility to cover what you want it to do, doesn't backfire on you, doesn't have the same gaps as sorcery. Am I on the right track?"

"Yes. And ideally would dovetail well with sorcery. You affected harmonics - that could be an amazing gamechanger - but you did it as a side effect. Since it's doable with mud-sticks-glitter at all, it would be lovely to work it in intentionally."

"Mm. I see what you mean," she agrees. "Do you have any ideas for a... hmm... a founding aesthetic, a structure, the way I used chess for chess magic? Chess magic is primarily ritual-based, but that isn't necessary to all forms of magic; would you prefer something more fluid and immediate, with less preparation time?"

"I would, although it's less important than some other desiderata."

"Well, I'm not expecting to pull together a whole system in the next hour, so there's time to think about it and consider options and maybe try a few things. Thoughts on aesthetic and structure?"

"My aesthetic mostly revolves around things like trees and lights and colors. I'm not sure if that gets us anywhere. What do you mean by structure?"

"Well, the conceptual structure of chess magic is useful because instead of being limited by whatever spells happen to exist to do the thing I want, I can construct a narrative out of distantly chess-based metaphors and as long as I have the right pieces to fill it and can cover any gaps with things like my focus stone, it works. I'd probably be most comfortable inventing a new system if it had some kind of structured conceptual basis like that. Not necessarily based in a game of any kind."

"I'm not sure I feel an affinity for anything quite that... modular. Is modularity the thing?"

"Modularity is definitely an important aspect."

"I like books but they're closer to freeform than modular..."

"Yes. I can't think of a good way to turn books as a general category into the basis for a magic system."

"Writing... has possibilities," she muses.

"Possibilities that don't take forever? I can write very quickly with fairylights but only in situations where I can already do some sorcery acceptably well."

"I'm sure if Royal were here she'd suggest you use math," says Castle. "Which isn't a bad thought, but you'd need to know the math to get any good out of it, and most people are less eager than Royal to learn math. Hmm. Do you want to come up to my workroom and experiment with affecting harmonics?"

"Okay."