Current study and negative-result overview

Equations in these preserved Markdown records retain their TeX notation where present; this is not a newly typeset proof edition.

CM compiler study artifacts — 26 September 2026

Start with README_FIRST.md and the current paper under 02_CONTRACT_REVISION_20260926/. This unpacked current collection includes the manuscript PDF and TeX, its build/QA and audit evidence, the synthetic v4 primary study, the original EPFL natural admission/correctness gate, the frozen natural v5 secondary timing run, raw rows, analysis, verification scripts, and P14 historical reproducibility material. Prior superseded project ZIP bundles and unrelated papers are outside this current-study archive.

The v4 synthetic primary utility test failed. The v5 natural secondary run has 70/71 complete Python timing cases; one calibration failure is retained. ABC conversion/truth checks completed for all 71. The corrected selected design count is ten. See 05_NATURAL_TIMING_20260926/README.md for the frozen-text erratum and precise limits. The historical P14 interval analysis replays, but original imported-runtime execution identity is not certified.

BUNDLE_MANIFEST_SHA256.csv inventories current unpacked study files using study-relative paths, with SHA-256 and byte counts. Frozen sub-study manifests and run freezes are preserved.

The current manuscript was updated and adversarially audited on 13 October 2026. It remains unpacked. The two earlier manuscript baseline directories are preserved in the parent collection's CM_Compiler_Pre_Semantic_Addition_and_Pre_Audit_2026-10-13.zip; see ../ARCHIVE_INDEX_2026-10-13.md. Earlier ZIPs and frozen sub-study records are preserved.