Target date: mid february 2024

Wanted:

• python-3.12.2 with torch and numba
• Operation Research packages refresh
• a big clean-up in packages that still don't have a windows/3.12 wheel
• python-3.13.0a4 : to bring "interpreters" (Eric Snow) and JIT (Brandt Bucher),

postponed to next:

• spyder-5.5.1 (too fresh)
• jupyterlab-4.1 (too fresh)

Focus:

• python-3.13.0a4 (multi-interpreters, JIT, as free threads seems more for a6-a7 at best hope)
• reduce maintenance burden on:
  • operation research
  • IA packages

Exploratory:

• llm, phy-2, https://webllm.mlc.ai/
• onnx
• qt6
• in search of an Operational Research toolset that is NOT Operational Research to maintain:
  • big iron motors: ortools from google, (big, really, but a big name too)
  • motors that are simple to support (rust style): clarabel
  • wrappers that don't force into un-maintained packages : qpsolvers, pyomo,
  • NOT wrappers that force into un-maintained packs: osqp & cvxpy
  • multi-objective optimizations tool

Direction:

• speed: python-3.13 with multi-interpreters, free-threads, small jit:
  • multi-interpreter scout tester is here : Tony Baloney(@anthonypjshaw
  • 'jit' option speed is branch 'justin' in https://github.com/faster-cpython/benchmarking-public
  • 'free-thread' 39 steps towards testable branch is there PEP 703 -- Making the Global Interpreter Lock Optional in CPython python/cpython#108219
  • keeping an eye on mojo: https://docs.modular.com/mojo/changelog.html
• how to evolve python to manage this is here: https://discuss.python.org/t/what-is-the-long-term-vision-for-a-parallel-python-programming-model/39190
• Rust packages are trendy, and mainly "universal" binaries.