The Calysto Scheme Project
summary
Calysto Scheme implements full Scheme (including call/cc) by writing an interpreter in continuation-passing style and mechanically transforming it into Python; the CPS source is ~6,000 lines, the CPS→data→register stages expand and then shrink, and the final Python runtime works and is integrated into Jupyter. It adds a nondeterministic choose operator for automatic backtracking, Python interop (numpy, matplotlib, TensorFlow), a simpler macro system, and notebook magics for parallel and inline plotting.
AI-written summary · about
chapters (11)
- 0:00Introduction to Calisto
- 1:22Origins in robotics
- 4:58The case for a new scheme
- 7:18Implementation methodology
- 12:58Development and debugging
- 15:03Macro system and features
- 17:11Machine learning applications
- 19:53Python integration
- 23:38Jupiter and parallel execution
- 25:15Educational use cases
- 27:51Q&A and concluding thoughts
related talks
31mAdding Python support to Scala tooling
25mPython as a Hackable Language for Interactive Data Science
36mA lightweight approach for accessing Python modules from Gambit Scheme
36mParens for Python
31mSupercharge your Python library using AST parsing
01h 00mKeynote: Python now and in the future
23mHow cocotb made Python-powered verification mainstream -- and what comes next
33mSPy (Static Python) lang: fast as C, Pythonic as Python