Explore a 32-minute conference talk from GPCE 2023 that introduces techniques to bridge the gap between abstract typing rules and multi-phased type checker implementations. Delve into a monadic interface for phased name resolution and the use of applicative functors to map abstract syntax trees onto monadic computations representing typing constraints. Examine how these approaches reduce discrepancies between type checker implementations and typing rules, demonstrated through implementations for Mini-ML with Damas-Hindley-Milner type inference and LM, a toy module language with complex import resolution. Gain insights into stable name resolution, scope graphs, and phasing in type checker design, presented by researchers from Delft University of Technology.
Overview
Syllabus
[GPCE23] A Monadic Framework for Name Resolution in Multi-Phased Type Checkers
Taught by
ACM SIGPLAN