anyway it's a lot of work and, while it would benefit the world, we think we can do more to help the world in the political domain rather than by spending our time building things.
Conversation
thank you for asking.
2
3
1
1
Oh neat! I'm currently making a dependently typed data description language at - the intension is to describe existing binary data formats, but that is a big challenge, hehe. Trying to chip away at it. I really like the sound of this!
1
5
Using it as a basis for other tooling sounds really cool. :) :)
1
1
> we got bogged down in the type checker.
haaaa! The struggle! I'm implementing it directly in Rust, but it often feels cumbersome and slow. I do wonder if I would have been better served making more prototypes.
2
1
yeah the challenge is basically that most of the people who understand dependent type checkers are mathematicians rather than programmers. so all the explanations of it are, um, abstract.
3
4
Agreed. Fortunately we understand unification, but there's really no on-ramp for anyone from a computer science background to learn about it easily.
1
1
What should I be on the lookout for to avoid this fate? And what is the distinction that you missed - was it between first order unification and higher order unification?
The role that unification plays, primarily, e.g. the whole subject of implicits.
2
We think we understand what you're getting at, but would love to hear you explain it.
2