📖Turning the IDE inside out with datalog
- authors
- Vilter, Pete
- year
- 2020
- url
- https://petevilter.me/post/datalog-typechecking/
- Writing algorithm on code in datalog allows for better incrementality and traceability → Datalog can be used to express algorithm on code
- for typechecking, it is hard(impossible) to specify error hints (what exactly went wrong) with datalog
- rust is about to use prolog for type checking (http://smallcultfollowing.com/babysteps/blog/2017/01/26/lowering-rust-traits-to-logic/)
- code: https://github.com/vilterp/datalog-ts
- demo: https://datalog-typechecker.netlify.app/
- you can store editor structures (e.g., current cursor position, syntax highlight) in the datalog itself