Changelog
The authoritative project changelog is the SESSION_LOG.md file at the
repository root, which contains an append-only record of every working
session.
For each version release, this page summarizes the headline changes; for
the full per-session history, see SESSION_LOG.md.
v5.27.0 — 2026-06-18 — Tier-1 production-complete
Headline: All 13 Tier-1 production-readiness items complete. PyPI wheel built and smoke-tested. Dual AGPL + commercial license adopted.
Major additions:
9 truly-independent primitives (down from 11 via PAPER_1521/1522 LANDMARK reduction proving D_BSFG and K_MEX are derivative)
Standard Model 12-fermion spectrum COMPLETE (m_u + m_d added)
Neutrino oscillation splittings Δm²_21 + Δm²_31 wired
Bayesian model comparison: ΔBIC = 94.1 (decisive UQFF preference over SM+ΛCDM on parameter-count grounds)
5-band uncertainty classification of all schema-tagged closures
Forward-predictions catalog with 42 falsifiable predictions
Provenance audit answering Q3: SSq=0.57 is truly independent
Production deliverables:
LICENSE+LICENSE-AGPL-3.0.txt+COMMERCIAL.md+CITATION.cff+NOTICEPREDICTION_LABELS.md+STATISTICAL_HYGIENE.md+PROVENANCE_AUDIT.md+COVERAGE.md+INPUT_DOMAINS.mdpyproject.toml+MANIFEST.in+ PyPI wheel + sdist.github/workflows/ci.yml+release.ymlSphinx documentation scaffolding (
docs/)
Calculator state:
2.66 MB / 48,405 lines / 34 public surfaces / 794 paradox keys
854 + 3 new = 857 fidelity gate tests passing, 0 failures
263 schema-tagged closures + 530 legacy_freeform
128 EXACT structural identities (residual < 1e-10)
1,795 whitepapers
Earlier history
Mining cycles tier-1 through tier-37 (sessions 2026-06-07 through
2026-06-17) drained 1,795 whitepapers, wired 794 paradox closures,
extended C++ reference to 368 functions, established the 9-sector UQFF
Lagrangian, closed all 8 Clay Millennium prize problems, and brought the
fidelity gate from 0 to 854 passing tests. See SESSION_LOG.md for the
full sequence.