Lithe-POC VS konna

For the month of March, my actual goal is to consolidate all my skills. Just like for the past several years, I want to break into making money through reinforcement learning in particular. I want to get good at nurturing my agents and using them. I've set a rule for myself that the games I will be training them on should be interactive. It is just such obvious loser behavior to be doing everything from the command line. That having said, until I learned how to do reactive programming properly just recently meeting that goal seemed impossible. Programming is the kind of activity where you can sit down and grind away at it and make progress, so the weak and stupid can do it too, but to achieve true excellence rather than a mess, you need the right tools and techniques.

The second was called Konna. AFAIK it’s Finnish for “frog”, but sources seem to disagree? I don’t speak Finnish, I got the word from a Finnish video game. My third and current language is called Peridot. I’m pretty proud of this name, although it’s less searchable than the previous ones. The origin is pretty simple, I was just looking around at gemstones and thought peridot looked neat.

Continuing work on Konna. I recently finished implementing dependent pattern matching, a pretty big feature! I figured now is the time to do some refactoring and rewriting - the elaborator is the first thing on my list. Once all this maintenance work is done I'll be implementing features like implicit arguments, overloading, and pattern matching on code values.

Github repo: https://github.com/eashanhatti/konna

Continuing work on my functional language Konna. The structured editor has progressed a whole lot recently - the most glaring bugs have been fixed and you can work with the entire language in it. The language itself is going well too, I'm currently thinking through:

Working on a rewrite of Konna (formerly Clamn). After using Rust for a little over a year in the original implementation, I decided I wasn't enjoying it. I'm using Haskell for the rewrite - I'd always wanted to write a big project in Haskell anyway haha. Definitely enjoying the higher-level conveniences it offers. The rewrite has been underway for about two weeks now, and so far I've got basic dependent types and partial evaluation implemented.

Continuing work on my functional systems language Clamn. I'd taken the last few weeks to fix a bunch of performance issues, but now it's finally back to implementing features: record types. I've got dependent types in my lang, which means I can get a bunch of more exciting features for free by implementing records, ADTs for instance.