[Matarener]

Requested by Mark · project thread

Summary

This is an independent scientific review and source-bound replay of #5827 (Ali Romman, κ 0.00312476680) and #5907 (DreamingOfClouds, κ 0.003125282), the two latest conditional P104 words. Both are recorded in the intake ledger as partially_reviewed, with "independent scientific review and source-bound replay remain open". This PR adds no new κ.

Result: no error found. The independent layers reproduce both authors' claims, including the exact raw crossings (.00312476681507, .00312476681508) and (.00312528909884, .00312528909885).

Layer	Result
Own exact replay of the #5811, #5827 and #5907 words (no package imports)	PASS: all 799 endpoint rows, strict MOVE and helper nesting, owned storage 29/31/13. Deltas exactly as declared: −H1−H4+H5 (seven events at 13815); paid −4H1+H4+H6+H7, targets −3H1+H3+H6+H7 and five boundary changes for #5907
Own invoice recount from the physical caller events and from the closed formula	equals both complete_cost entries: 66,870 / 13,444 and 66,852 / 13,444, surplus 760
Dense all-column check with absolute phase (PR5548 interpreter)	PASS on all 3,163 operators of both words: 2,918 and 2,921 distinct signed maps plus 208 adapters. Phase-only and sign faults on the new map 3→712 are rejected
Own 80-digit decimal funding engine	both crossings equal the author intervals; tested savings positive; tail ratios α, γ, δ hold; crossing unique on [0, .0034]

Reproducibility defect in #5907: research/p104-helper468-rung-removal/verify.py fails on Linux (Python 3.12.3, zlib 1.3) at Entire regenerated WORD.json.gz, byte for byte. The regenerated and shipped containers differ only in gzip header byte 9 (OS code 0x13 versus 0x03), and the payloads are equal. With that comparison made on the decompressed payload (as CONTRIBUTING.md asks), the complete #5907 chain passes, including its committed RESULT.json. #5827's own verifier passes unchanged.

Package

research/p104-unread-rung-independent-audit/ (README, PROOF, verify.py, FULL-CHECK.json with all 3,163 matrix digests, RESULT.json, SOURCES.json, TASK.json, VERIFY.json). Base: canonical main 19fdb086ea96af9a9f87936a2b888e10af4ed810. Only new files are added; no existing file is changed.

python3.12 -B research/p104-unread-rung-independent-audit/verify.py                  # ~2.5 min, subset of layer 3
python3.12 -B research/p104-unread-rung-independent-audit/verify.py --full --jobs 4   # all 3,163 maps

Checks run: the default verifier (PASS, 2.5 min); --full --jobs 4 (PASS); scripts/research_package.py --check-only and the registered run; scripts/check_research_commit.py --check-modes.

Scope

The finite local words and their complete invoices are covered. The five inherited framework assumptions remain. The physical caller is recounted but not re-derived. The compiler threshold and the 54+4 outer conditions are checked only by the authors' verifiers, which pass. This is an agent check, not human acceptance. The selected record is unchanged.

Credit: Ali Romman, DreamingOfClouds, Mason Hale, Chris R (PR5548 dense interpreter), unpaidmoon and the P104 lineage. Prepared for Matarener with Claude (Anthropic) via Claude Code.

🤖 Generated with Claude Code

https://claude.ai/code/session_011VhYm9Jh5Lo7JQJaBFt7FU