- Wednesday: I made the type annotations in the Ink port a little less sloppy. I also figured out how to get the conlang stuff into a more inspectable form without immediately starting to yak-shave.
- Thursday: I tried to make the code follow the structure of the C# version a little better.
- Friday: I got the rest of the way through those updates, and started porting tests again.
- Saturday: I realized that the Ink VM is too low-level to unilaterally implement the semantics tweak I want for it. Between this entry and the next, I realized that I need to get into compiler hacking for that.
- Sunday: With some efforts on the side in the days before, I got the glosses I've written put into Sphinx.
- Monday: I started pulling from my other documentation efforts into the Sphinx project. Ended up thinking about how to arbitrarily divide these fictional words into specific categories. In my defense, I don't think the specific issue I'm worried about is settled for natural languages (where applicable).
Next week, I'm going to try to reach another failing test in the Ink port. "Approximately alphabetically" might not be a great order to port the tests in, but it's good for being able to compare later. Also, I'm going to just commit to a position on the conlang stuff.