Explore a 20-minute conference talk from OOPSLA2 2023 that delves into improving type error messages for constraint-based type inference systems. Learn about a novel approach using subtyping constraints to capture data flow, which aims to classify and explain type errors in a more programmer-friendly manner. Discover how the proposed algorithm illustrates type errors as faulty data flows, presenting them as sequences of relevant program locations. Understand the application of these concepts to languages without subtyping, including their integration with Hindley-Milner type inference. Examine the results of a user study evaluating the effectiveness of flow-based messages compared to traditional implementations. Gain insights into the challenges of creating meaningful type error messages and the potential benefits of this data flow-oriented approach for improving programmer understanding and error localization.
Getting into the Flow: Towards Better Type Error Messages for Constraint-Based Type Inference
ACM SIGPLAN via YouTube
Overview
Syllabus
[OOPSLA23] Getting into the Flow: Towards Better Type Error Messages for Constraint-Based ...
Taught by
ACM SIGPLAN