Future Directions for Session Types
summary
Panelists argue concrete next steps: integrate failure handling plus timeouts/congestion into the theory, design a missing module system for session types, and push implementations into mainstream languages (a Rust EDSL and a Servo experiment exist). Mechanization is essential but costly—one infrastructure took ~1 year to build and mechanization adds 200–300% effort—so bootstrap libraries (connectivity-graph, logical-relations mechanizations, benchmarks) first. Practical directions include runtime monitoring, protocol inference from traces, and automated test generation from global types.
AI-written summary · about
chapters (9)
- 0:00Welcome
- 2:48Panelists' background
- 5:39Future directions
- 15:26Session types' future?
- 22:10Session types mainstream
- 32:13Mechanization in research
- 58:48Funding strategies
- 1:06:03Testing session types
- 1:15:24Closing remarks
related talks
01h 03mPanel: Next 50 Years of POPL
40mSystems Thinking for Developers
32mMulticompatibility for Multiparty-Session Composition
33mPanel Session - Modularity Futures: OSGi and Java 9+
01h 03mThe Language of the System
43mLearning to Live with Errors
46mThoughtful Software Engineering
31mToken Engineering Panel
