Diagrammatic Execution Models for Functional Languages
summary
Diagrammatic semantics represents programs as graphs where a token moves around a program diagram and rewrites visited parts; it implements concrete strategies including models call-by-value and call-by-need and distinguishes left-to-right vs right-to-left evaluation. A working web visualizer and visual debugger with breakpoints on nodes shows step counts, node counts, forward/backward stepping and reveals surprises (e.g. merge sort appearing slower on small inputs). Extensions produce idealized TensorFlow / spreadsheet-style dataflow with clocked steps and work as a usable prototype.
AI-written summary · about
chapters (9)
- 0:00Start
- 36:52Language with Dataflow
- 38:18Dataflow programming
- 43:06Example: alternating signal
- 43:25FIR filter
- 44:22Example: average 3
- 44:40Neural network?
- 45:16Related Work
- 46:56Diagrammatic Semantics
related talks
49mFunctional Design Explained
47mSemi-Automated Direction-Driven Functional Conversion
20mOn the paradigm of functional programming
33mAnalyzing Functional Programs
32mExperience Report: Graph Rewriting with Lexical Effect Handlers
21mDemo: A functional EDSL for mathematics visualization that compiles to JavaScrip...
01h 03mCrash Course on Notation in Programming Language Theory (Part 2)
34mFinite State Machines for Functional Software Machinery