andrew-johnson-4 / perplexity

A notational semantic for documenting neural networks through diagrams

Hi, I'm Andrew Johnson

All things considered... copy and paste is probably the best option.

2024 Roadmap (Working towards a verified kernel language)

• ✓ 1.0.0 Bootstrap Compiler 🥳🎉🎁
• ✓ 1.6.0 Production Compiler
• ✓ 1.10.0 Ad-Hoc Specialization and Hygienic Macros
• 100% compiled code in compiler / somewhat stable compiler / somewhat stable language standard
  • ✓ cli.lm
  • ✓ tokenize.lm
  • ✓ parse.lm
  • ✓ preprocess.lm
  • ✓ typecheck.lm
  • ✓ assemble.lm
  • codegen.lm
  • fragment.lm
  • ✓ utility.lm
• publish language standard v1.0

Upcoming Planned Features

• actually useful errors messages with code snippets and source location (mostly working on strict compiler)
• arbitrary sized return values
• parametric polymorphism

Request for Comment

If you feel like contributing to either LM or LSTS it would be enormously helpful to have anyone simply read a bit and ask a question. The projects are starting to grow to a size where it is hard to understand where a curious learner might get stuck. Different people have different backgrounds so it takes all shapes and sizes. Everyone is welcome, though I fear that the learning curves are still very steep.