Can Logic Programming Be Liberated from Predicates and Backtracking? (Lightning talk)
summary
Claims a systematic method that transforms Prolog predicates into Curry functions and a tool that compiles those into Haskell or Go, with a Curry package, Docker image and web interface. Benchmarks against SWI‑Prolog show some programs keep the same steps but many drop from infinite or factorial search to finite or much smaller search spaces; some cases run orders of magnitude faster. A fair search via goroutines yields completeness without large overhead, and the transformation is semantics‑preserving in the worst case.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 0:43Logic programming foundations
- 2:07Functional logic with Curry
- 3:32Transforming to Curry
- 4:56Curry vs Prolog benchmarks
- 6:23Avoiding backtracking
- 7:45Fair search advantages
- 8:49Fair search benchmarks
- 9:51Parallelism via scheduling
- 10:12Advantages over Prolog
- 11:17Teaching and future
- 12:01Q&A: Compiling Curry to Prolog
- 13:10Q&A: Prolog vs Curry expressiveness
- 14:56Q&A: Prolog to Curry limits
- 16:21Q&A: Type systems and extensions
- 17:43Q&A: Probabilistic programming
- 18:45Q&A: Prolog vs functional mindset
- 20:17Q&A: Fair search complexity
- 20:58Q&A
