Reproducible tests

• Use mpmath with mp.dps set to reach numerical goals (< 1e-16 where claimed).
• Include unit tests that assert numeric invariants and convergence properties.

Recommended test types

• Deterministic numeric tests (mpmath)
• Statistical comparisons: t-tests, Cohen's d (example helper functions in axioms.py)
• Visual inspections: error plots, residual histograms

Case studies (examples)

• Physical: muon lifetime time-dilation checks vs Z = T(v/c)
• Discrete: prime counting approximations vs PNT and Z5D methods from notebook

Limitations & labeling

• All unverified claims must be labelled UNVERIFIED.
• Document assumptions (e.g., selection of c, functional forms for A).

How to contribute validation results

• Add notebooks in /notebooks/
• Push reproducible CI-friendly tests to tests/ using pytest and fixtures

---