docs: three-flavours-of-256 framing + review qualifications (X-Trans, A8-gated seek, integer-path replay, [H] dependence, landing-map scope)

Operator refinement of the §8 table-family paragraph — three flavours
of 256:
1. CAM-PQ = 6x256^2 compressed to 6x256: latent per-subspace 256^2
   centroid-distance families, shipped as per-query 6x256 f32 ADC rows
   (6 KB, cam_pq.rs:76-84), never materialized.
2. bgz17 = the explicit 256^2: one materialized 256^2 u16 table + kxk
   u8 compose per palette; the 388 KB SpoDistanceMatrices benchmark is
   this flavour on 3 S/P/O planes (3x128 KB = 384 KB).
3. V3 facet = explicit 6x256^2 as ADDRESS, codec-agnostic: 6x(u8:u8)
   rails = six coordinate pairs into 256^2 spaces = 96 bits;
   classid -> ClassView switches which codec's 256^2 family each rail
   indexes.

Plus five external-review qualifications across knowledge doc / plan /
blackboard:
- X-Trans corrected to "repeating but less-regular 6x6 tile" (was
  "non-repeating").
- (a) reworded to phase-side seekability — CABAC's serial context chain
  remains; bitstream-level seek requires A8's framed/context-reset
  regions.
- (f) bit-exact trajectory replay scoped to the proven CPU/wasm integer
  path; float/GPU stages outside the claim.
- (g) period-17 "friendlier to Jirak" graded [H, probe-gated] — the
  permutation self-test proves bijectivity, not decorrelation; needs a
  correlation-spectrum probe vs a PRNG baseline.
- Landing map aligned with the delivered diff (no §5 probe-queue
  changes were needed).

Knowledge-doc suite green (117/117).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K3RyLEbuNSHxxB3NTTrGki

Author: claude

.claude/blackboard.md

@@ -749,18 +749,23 @@ default is x86-64-v3 (avx2) so ndarray_simd activates on avx512 builds only.
   division of labor stated as a rule.
 - **§8 96-bit facet carving:** CAM-PQ 48b + helix `ResidueEdge` 24b +
   turbovec 24b = 96 bit = the V3 12-byte content-blind payload identity;
-  `Signed360` (48b) is the out-of-row alternate carving. Table families
-  disambiguated (post-review correction): 388KB benchmark = palette
-  `SpoDistanceMatrices` 3×(256² u16 = 128KB); CAM-PQ ADC = per-query
-  6×256 f32 = 6KB (`cam_pq.rs:76-84`), no fixed 256² footprint; bgz17 =
-  one 256² u16 + k×k u8 compose per palette.
+  `Signed360` (48b) is the out-of-row alternate carving. Three flavours
+  of 256 (post-review correction + operator refinement): CAM-PQ =
+  6×256² compressed to per-query 6×256 f32 ADC rows (6KB,
+  `cam_pq.rs:76-84`); bgz17 = the explicit materialized 256² u16 (+ k×k
+  u8 compose; 388KB benchmark = 3 S/P/O planes × 128KB); V3 facet =
+  explicit 6×256² as codec-agnostic ADDRESS (6×(u8:u8) rails = 96 bit;
+  classid→ClassView switches which codec's 256² family each rail
+  indexes).
 - **§9 kernel-shape rule:** VNNI/AMX for matmul-shaped ops, LUT/texture
   for lookup-shaped ops — turbovec NativeLut measured **11.4×** faster
   than the VPDPBUSD GEMM polyfill (n=20k/dim=512/4-bit, FINDING). ITU
   claim scoped to compute kernels only (not CABAC/conformance/ECM count).
 - **§10 replayable-tile synergies:** 4×4 Morton tile as the shared
-  object between H.268 (anti-CABAC seek, seekable grain, C6-scoped
-  native tiling) and cognitive shaders (RNG-free exploration, replayable
-  thinking, anti-confabulation, cache-native 192B working set) — all
-  nine consequences stay **probe-gated** (D-MTS-1..3, PHASE-1/PERT-RHO/
-  PYR-1, WHP-1..4, L4 doc-lock); no kill condition weakened.
+  object between H.268 (phase-side seekability — entropy-level seek
+  still A8-gated; seekable grain; C6-scoped native tiling) and cognitive
+  shaders (RNG-free exploration, replayable thinking on the CPU/wasm
+  integer path, anti-confabulation [H, needs correlation-spectrum
+  probe], cache-native 192B working set) — all nine consequences stay
+  **probe-gated** (D-MTS-1..3, PHASE-1/PERT-RHO/PYR-1, WHP-1..4, L4
+  doc-lock); no kill condition weakened.

.claude/knowledge/pr-x12-h268-morton-wgpu-synergies.md

@@ -120,9 +120,11 @@ any float-path GPU bit-exactness claim.
 The Pythagorean comma is the residue of a stack of pure fifths that never
 closes back onto the octave; a piano tuner's real-world dodge (equal
 temperament) trades exactness for closure. Fujifilm's X-Trans sensor
-generalizes the same move spatially: its non-repeating 6×6 pixel arrangement
-is deliberately incommensurate with common demosaic/moiré periods, so the
-anti-aliasing filter can be thinned or dropped. Both are the same design
+generalizes the same move spatially: its 6×6 colour-filter tile — repeating,
+but far less regular than Bayer's 2×2 (a larger period with a quasi-random
+arrangement inside the tile) — is deliberately less commensurate with common
+demosaic/moiré periods, so the anti-aliasing filter can be thinned or
+dropped. Both are the same design
 pattern: **a generator that does not resonate with the sampling lattice
 avoids the periodic beat pattern (the comma) that a resonant generator
 produces.**
@@ -211,26 +213,33 @@ row — it is the **out-of-row / alternate-carving variant**, selected
 instead of (not in addition to) the `ResidueEdge` + turbovec pairing when
 full-sphere signed precision is needed.
 
-**Table-family clarification (corrected 2026-07-16 post-review — do not
-conflate three distinct structures):**
-- The measured **388 KB `SpoDistanceMatrices` benchmark** (§1 above) is the
-  **palette** structure: **3 S/P/O planes × one 256×256 u16 distance table
-  (128 KB each) = 384 KB** (`palette_distance.rs:145-158` — `DistanceMatrix
-  { data: Vec<u16>, k }`, one per subject/predicate/object plane).
-- **CAM-PQ has no fixed 256×256 distance-table footprint.** Its 6-subspace
-  structure is the per-vector 48-bit **code** (the §8 lane table above);
-  its ADC distance tables are **per-query** `6 × 256` f32 = **6 KB**,
-  recomputed per query and L1-resident (`cam_pq.rs:76-84`).
-- **bgz17's palette layer** is one 256×256 u16 distance table **plus** a
-  k×k u8 compose table, per palette.
-The first and third share the "256×256 dense u16 LUT, texture-isomorphic"
-shape (§3 row 2 above); none of the three footprints should be added to or
-substituted for another. (An earlier draft of this paragraph attributed
-the 384 KB to "6 × 64 KB CAM-PQ tables" — wrong on both the arithmetic, a
-256×256 u16 table is 128 KB, and the attribution.)
-This carving is a `ClassView` projection over content-blind bytes — see
-lance-graph `le-contract.md §3` for the canonical 6×(u8:u8) / 4×(u8:u8:u8)
-/ 3×(u8:u8:u8:u8) readings the same 12-byte register supports.
+**Three flavours of 256 (table-family clarification, corrected + refined
+2026-07-16 post-review — operator framing):**
+- **CAM-PQ — 6×256² compressed to 6×256.** The latent structure is six
+  per-subspace 256×256 centroid-distance families, but the shipped path
+  never materializes them: ADC precomputes a **per-query 6×256 f32
+  row-set** = **6 KB**, L1-resident, recomputed per query
+  (`cam_pq.rs:76-84`) — the compressed, asymmetric projection of the
+  latent 6×256². (A symmetric code↔code variant would materialize the
+  full six tables; the shipped codec does not.)
+- **bgz17 — the explicit 256².** One materialized dense 256×256 u16
+  distance table per palette, **plus** a k×k u8 compose table. The
+  measured **388 KB `SpoDistanceMatrices` benchmark** (§1 above) is this
+  flavour on the three S/P/O planes: **3 × (256² u16 = 128 KB) = 384 KB**
+  (`palette_distance.rs:145-158`).
+- **V3 facet — explicit 6×256² as ADDRESS, codec-agnostic.** The 12-byte
+  content-blind payload read as `6×(u8:u8)` rails = six coordinate pairs
+  into 256×256 spaces = **96 bits of pure address**. The payload does not
+  know whose tables its rails index — **`classid → ClassView` selects
+  which codec's 256² family** (CAM-PQ subspace, bgz17 palette, helix
+  residue) interprets each rail. See lance-graph `le-contract.md §3` for
+  the canonical 6×(u8:u8) / 4×(u8:u8:u8) / 3×(u8:u8:u8:u8) readings of
+  the same register.
+The first two are table *footprints* (compressed vs materialized); the
+third is the *addressing shape* they are consumed through. None of the
+footprints should be added to or substituted for another. (An earlier
+draft attributed the 384 KB to "6 × 64 KB CAM-PQ tables" — wrong on both
+the arithmetic, a 256² u16 table is 128 KB, and the attribution.)
 
 ## 9. The kernel-shape rule (engine follows operation shape)
 
@@ -277,10 +286,15 @@ function of its address via the bijective coprime walk (§7), and the only
 bytes actually stored are magnitudes. Same object, two consumers.
 
 **H.268 consequences:**
-- **(a) Anti-CABAC random access** — no serial phase state to carry, so a
-  decoder can seek directly into a tile without replaying a bitstream
-  prefix; strengthens the C4 path once A8 (region-addressable stream
-  framing) lands.
+- **(a) Phase-side seekability (the anti-CABAC direction)** — the phase
+  generator carries no serial state, so the *phase side* of any tile is
+  reconstructible at any address with no prefix replay. This is NOT by
+  itself CABAC random access: the entropy-coded magnitudes still carry
+  CABAC's serial context chain, so bitstream-level seeking additionally
+  requires independently framed / context-reset regions — which is
+  precisely what A8 (region-addressable stream framing) provides. The
+  consequence strengthens the C4 path **only once A8 lands**; until then
+  it is phase-side seekability only.
 - **(b) Seekable grain** — unlike AV1's seeded film-grain synthesis
   (decoder must carry PRNG/seed bookkeeping), the integer walk regenerates
   identically from the address alone and survives a WGSL port per the C9
@@ -301,13 +315,21 @@ bytes actually stored are magnitudes. Same object, two consumers.
   loop (composes with `E-NOBODY-WAITS-1`).
 - **(f) Replayable thinking = auditable cognition** — combined with
   temporal-stream replayability (`E-MARKOV-TEMPORAL-STREAM-1`), a full
-  trajectory including exploration noise re-runs bit-exactly;
-  counterfactual replay therefore stores **zero** exploration state.
-- **(g) Anti-confabulation = anti-moiré in concept space** — a coprime
-  probe schedule is decorrelated from the palette lattice by construction;
-  a *known* period-17 dependence structure is friendlier to
-  `I-NOISE-FLOOR-JIRAK`'s weak-dependence analysis than an *unknown* PRNG
-  correlation structure would be.
+  trajectory including exploration noise re-runs bit-exactly **on the
+  proven CPU/wasm integer path** (the pinned-order five-backend contract);
+  float and GPU stages stay outside the bit-exactness claim per §3 row 9
+  (the integer-only GPU caveat) unless a deterministic-backend probe is
+  added. Within that scope, counterfactual replay stores **zero**
+  exploration state.
+- **(g) Anti-confabulation = anti-moiré in concept space [H, probe-gated]**
+  — a coprime probe schedule is decorrelated from the palette lattice by
+  construction, and a *known* period-17 dependence structure is plausibly
+  friendlier to `I-NOISE-FLOOR-JIRAK`'s weak-dependence analysis than an
+  *unknown* PRNG correlation structure. The "friendlier" half is an
+  **unverified inference**: the period-17 permutation self-test proves
+  bijectivity, NOT decorrelation. Promotion to FINDING requires a measured
+  dependence probe — the walk's correlation spectrum against the palette
+  lattice vs a PRNG baseline, judged under Jirak weak-dependence rates.
 - **(h) Exact phase-side unbinding** — sign is recomputable per address
   with no cleanup codebook needed; a cleanup codebook is only needed for
   magnitudes. The two-algebra rule (sign = XOR, magnitude = `vsa_bundle`,

.claude/plans/H268-comma-96bit-replayable-addendum-v1.md

@@ -47,14 +47,18 @@
   indexes which table).
 - Budget constraint: `Signed360` (6B) does NOT fit alongside both other
   lanes; it is the out-of-row/alternate-carving variant.
-- Table-family clarification **(CORRECTED post Opus-review — the original
-  bullet here claimed "six 256×256 CAM-PQ tables, 6×64KB=384KB", wrong on
-  arithmetic AND attribution)**: the 388KB `SpoDistanceMatrices` benchmark
-  is the **palette** structure — 3 S/P/O planes × one 256² u16 table
-  (128KB each) = 384KB (`palette_distance.rs:145-158`); CAM-PQ has NO
-  fixed 256² footprint — its ADC tables are per-query 6×256 f32 = 6KB
-  (`cam_pq.rs:76-84`); bgz17's palette layer is one 256² u16 + k×k u8
-  compose per palette.
+- Table-family clarification **(CORRECTED post Opus-review, REFINED per
+  operator — the original bullet claimed "six 256×256 CAM-PQ tables,
+  6×64KB=384KB", wrong on arithmetic AND attribution)**. Three flavours
+  of 256: (1) **CAM-PQ = 6×256² compressed to 6×256** — latent
+  per-subspace 256² families, shipped as per-query 6×256 f32 ADC rows =
+  6KB (`cam_pq.rs:76-84`), never materialized; (2) **bgz17 = the explicit
+  256²** — one materialized 256² u16 table + k×k u8 compose per palette;
+  the 388KB `SpoDistanceMatrices` benchmark is this flavour on 3 S/P/O
+  planes = 3×128KB = 384KB (`palette_distance.rs:145-158`); (3) **V3
+  facet = explicit 6×256² as ADDRESS** — 6×(u8:u8) rails = six coordinate
+  pairs into 256² spaces = 96 bits, codec-agnostic; `classid → ClassView`
+  selects which codec's 256² family interprets each rail.
 
 ### B3 — Kernel-shape rule (new §9)
 - **Match engine to operation shape:** VNNI/AMX for matmul-shaped stages
@@ -72,23 +76,33 @@
 ### B4 — Replayable-tile synergies: H.268 × cognitive shaders (new §10)
 The object: 4×4 Morton tile (2bit x ⊕ 2bit y), phase address-derived via the
 bijective walk, magnitudes the only stored content. Consequences:
-- H.268: (a) anti-CABAC random access (no serial phase state; strengthens
-  the C4 path once A8 lands); (b) seekable grain (vs AV1 seed bookkeeping;
-  integer walk survives WGSL per C9); (c) conformance = the period-
-  permutation self-test; error localizes to magnitudes; (d) parallelism:
-  16 cells = one SIMD lane group / wgpu workgroup tile — NATIVE to the
-  H.268 scene codec; HEVC-compat lane keeps 8×8/64×64 (C6 correction
-  preserved).
+- H.268: (a) phase-side seekability, the anti-CABAC *direction* (no serial
+  phase state; NOT CABAC random access by itself — entropy-coded magnitudes
+  keep CABAC's serial context chain, so bitstream-level seek additionally
+  requires A8's independently framed/context-reset regions; strengthens the
+  C4 path only once A8 lands) **[qualified post-review]**; (b) seekable
+  grain (vs AV1 seed bookkeeping; integer walk survives WGSL per C9);
+  (c) conformance = the period-permutation self-test; error localizes to
+  magnitudes; (d) parallelism: 16 cells = one SIMD lane group / wgpu
+  workgroup tile — NATIVE to the H.268 scene codec; HEVC-compat lane keeps
+  8×8/64×64 (C6 correction preserved).
 - Cognitive shaders (the bigger half): (e) RNG-free exploration — phase =
   pure function of position; deletes the last shared-mutable-state
   candidate from the thinking loop (composes with E-NOBODY-WAITS-1);
   (f) replayable thinking = auditable cognition — with temporal-stream
   replayability (E-MARKOV-TEMPORAL-STREAM-1), full trajectory incl.
-  exploration noise re-runs bit-exactly; counterfactual replay stores zero
-  exploration state; (g) anti-confabulation = anti-moiré in concept space
-  (coprime probe schedule decorrelated from the palette lattice; known
-  period-17 dependence structure is friendlier to I-NOISE-FLOOR-JIRAK than
-  unknown PRNG correlations); (h) exact phase-side unbinding (sign
+  exploration noise re-runs bit-exactly **on the proven CPU/wasm integer
+  path only** (pinned-order five-backend contract; float/GPU stages stay
+  outside the claim per the integer-only GPU caveat) **[qualified
+  post-review]**; counterfactual replay stores zero exploration state
+  within that scope; (g) anti-confabulation = anti-moiré in concept space
+  **[H, probe-gated — qualified post-review]** (coprime probe schedule
+  decorrelated from the palette lattice by construction; the claim that
+  known period-17 dependence is friendlier to I-NOISE-FLOOR-JIRAK than
+  unknown PRNG correlations is an unverified inference — the permutation
+  self-test proves bijectivity, not decorrelation; promotion needs a
+  measured correlation-spectrum probe vs a PRNG baseline under Jirak
+  rates); (h) exact phase-side unbinding (sign
   recomputable per address; cleanup codebook needed for magnitudes only;
   two-algebra rule intact); (i) the 4×4 tile = cache-native working set
   (16×2B ×6 lanes = 192B = 3 cache lines; L4 substrate is flat Morton SoA
@@ -105,7 +119,7 @@ bijective walk, magnitudes the only stored content. Consequences:
 
 | Repo | Branch (restarted from default; both prior PRs merged) | Files |
 |---|---|---|
-| ndarray | `claude/x265-x266-plans-review-h9osnl` @ origin/master | matrix doc §7-§10 + probe-queue additions; this plan file; blackboard append |
+| ndarray | `claude/x265-x266-plans-review-h9osnl` @ origin/master | matrix doc §7-§10 (no §5 probe-queue changes needed — the standing queue already carries every probe §7-§10 references); this plan file; blackboard append |
 | lance-graph | `claude/x265-x266-plans-review-h9osnl` @ origin/main | EPIPHANIES prepend (E-H268-REPLAYABLE-TILE-1); capstone pointer line; PR_ARC #697 post-merge entry + LATEST_STATE entry (board-hygiene rule) |
 
 Gates: ndarray knowledge-doc suite (117 tests) green; no affirmative stale