Files

T

Robert Allan JamesandClaude Sonnet 5 1e2dc7fbf0 Close punch-list item #4 : amd64 dict_hash non-determinism does not reproduce

Ran the amd64 kernel acceptance leg 5 times back to back as the punch
list's action item asked. dict_hash was byte-identical across every run
for every VM (Artemis, both Hermes instances, Hera/MAMA_INIT), and
matched aarch64/riscv64 exactly each time. Not jitter, and not
stable-but-different either.

The underlying mechanism the hypothesis pointed at is still real and
unchanged (capsule_dict_hash_hook() still folds execution_heat into the
hash; amd64 still runs its timer in RELATIVE mode under this
hypervisor) — but PARITY:MAMA_INIT and the child-VM PARITY:BIRTH lines
all print before the heartbeat starts, and heat only decays on
heartbeat ticks, so there's no window for the timer's non-determinism to
reach execution_heat before any of these hashes get computed. Most
likely the original 2026-07-24 observation was a one-off (loaded host
machine, coincidental timing), not a real gap.

No code change. No amendment to CLAUDE.md's acceptance criteria needed —
"identical dict_hash across all three architectures" holds up under
repeated testing. Closure note added to the punch list; the 5 verification
runs' logs and DoE CSVs are kept as the supporting evidence.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

2026-08-02 08:05:28 -04:00

architecture

Initial commit — LithosAnanke kernel

2026-08-01 07:49:56 -04:00

README.md

docs/working/ — Living Documentation

Source material for the formal documentation tier. Design notes, DoE logs, architecture decisions, experiment reports, draft specs — everything not yet promoted to docs/formal/. See docs/CLAUDE.md for the two-tier model.

Layout

Directory	Contents
`architecture/`	System design: physics engine, HAL, heartbeat, pipelining, word-ACL, adaptive systems, getting-started, plus dated design/finding notes for Artemis, Hermes, and the K-PUSH dictionary-shadowing bug. (`03-architecture/`, `getting-started/`, `heartbeat_csv_export.md`)
`experiments/`	DoE designs and run data. (`02-experiments/`, `campaigns/` = former top-level `experiments/`, including the `l8_attractor_map`, `l8_validation`, `doe_2x7`, `shape_validation`, and `window_scaling` campaigns)
`specifications/`	Protocol/format specs — currently `HOLA_PROTOCOL.adoc`.
`hardware/`	Board bring-up, platform/perf notes — `PGO_GUIDE.adoc`, `PLATFORM_BUILD_GUIDE.adoc`, plus `performance-profiling/` and `platform-integration/` subtrees.
`legal/`	Patent-adjacent material — HOLD. See `legal/README.md`. Nothing moved here without Bob's instruction.
`papers/`	Academic / citable writing: claim tables, anti-claims, null hypothesis, reproducibility, ontology (markdown), SSRN companion (`SSRN_companion/`), `FINAL_REPORT/`, published SSRN PDF (`published/`).
`scratch/`	Uncategorized working material: governance/meta docs (`governance/`, `ci-cd/`, `testing-quality/`), the full `docs/src/` developer-reference tree (`src/`), asset holding area (`assets/`).
`archive/`	Historical / superseded / obsolete records: quality audits, operations, research drafts, session logs, the former `docs/archive/` tree (`legacy/`), artifacts.

Phase 2 reorganization notes (2026-06-16)

Moves used git mv (history preserved). Each moved .md/.adoc/.tex file carries a one-line provenance header noting its original path.
Reorg used directory-level consolidation rather than splitting every mixed-tag directory file-by-file, to keep companion data (logs, CSVs, randomization matrices) with the reports they belong to. Per-file status tags remain authoritative in docs/INDEX.md.
docs/src/ landed under scratch/src/ as a holding area — it is mixed developer reference that still needs semantic re-filing into architecture/, hardware/, and specifications/.
Repo-root convention files (README.md, ROADMAP.md, ONTOLOGY.md, CONTRIBUTING.md, CODE_OF_CONDUCT.md, NOTICE.md) and the docs landing files (docs/CLAUDE.md, docs/INDEX.md, docs/README.md) were left in place.
DELETE candidates: see docs/DELETE-CANDIDATES.md (nothing deleted).