Coding 2025-11-28
I'm awake! Somehow...
Anyway, I got some roots chosen for the purposes of grammaticalizing later, but also providing other words. Now, let's think about the bare minimum functionality required for representing semantic drift.
First decision I want to lock in: it keys off of the post-sound-change pronunciation, because it makes more sense to me to say "when it's pronounced like this, it means this" than to say "first we change the meaning, then we change the sound". Next, key off the short gloss. Then, have enough data for the full replacement.
Now, if either the pronunciation or the short gloss aren't found, then something has gone wrong, and it needs to bail out with an error.
Anyway, that format is interesting, because it's actually just a lexicon. And something very similar could be implemented for loanwords.
Okay, just bashed out some prototype code that looks really ugly, but I don't want to think about improving it right now. Tomorrow, I'll check whether it even works.
Anyway, that's everything I wanted done, so it's time to wrap up.
Good night.