Merge remote-tracking branch 'origin/main' into claude/wonderful-hawking-lodtql

Author: claude

.claude/board/AGENT_LOG.md

@@ -10,6 +10,17 @@
 **Measured (AVX-512+VNNI host, no AMX tiles; n=20k dim=512 k=10 4-bit):** native LUT-ADC 76 µs/q (recall 0.785) ; polyfill GEMM 867 µs/q (recall 0.764) ; scalar 6 267 µs/q. **polyfill 11.4× slower than native** → TurboQuant deliberately trades the matmul away (LUT gather, not dot), so AMX accelerates the op it removed. Native LUT stays the production kernel; polyfill retained as AMX-ready baseline. Placement verdict: index → spine (lance-graph), kernel-math → ndarray (already owns clam/cam_pq/cascade/amx_matmul). The promising synergy is a Belichtungsmesser σ-gate on the LUT scan, NOT AMX.
 
 **Verification:** `cargo build --lib -p turbovec` (fork-wired) green; `cargo test -p turbovec --features ndarray-simd search_polyfill` 2/2 green; `cargo test --manifest-path crates/lance-graph-turbovec/Cargo.toml` green; benchmark ran. Pre-existing upstream turbovec dead-code warning (`avx2_block_epilogue`) silenced minimally. Commits: one per repo on the branch.
+## 2026-06-13 — SoaEnvelope binding for canonical NodeRow (the canon-as-substrate keystone)
+
+**bardioc cross-session.** Closes punchlist item §7.2 of the 2026-06-13 SoA migration diff resolution doc — the canonical row layout is now bound to the envelope ABI. New `NodeRowPacket<'a>` wrapper in `canonical_node.rs` zero-copy-views a `&[NodeRow]` (each row `#[repr(C, align(64))]` at 512 bytes) as a row-strided LE byte packet through `SoaEnvelope`. Three-column descriptor table (`NODE_ROW_COLUMNS`): key (16 × u8 at offset 0), edges (16 × u8 at offset 16), value (480 × u8 at offset 32) — sums to `NODE_ROW_STRIDE = 512`. Internal structure within each slot stays canon-described (`NodeGuid` for the key, `EdgeBlock` for the edges, registry `ClassView` for the value carve-out) — the envelope contract is at the row-stride level, not the field-decomposition level. `NodeRowColumn` enum exports the column ordinals as `pub enum { Key=0, Edges=1, Value=2 }` for type-safe `column_le` access. `as_le_bytes()` is unsafe-free at the API but uses `core::slice::from_raw_parts` internally with a documented SAFETY note (NodeRow `#[repr(C)]` + locked size + canon-LE field accessors). +9 tests covering column-table layout, empty-packet verification, single-row zero-copy (pointer equality), multi-row byte length, `row_le`/`column_le` LE byte ranges, canon-LE key end-to-end, and `LAYOUT_VERSION` parity. `cargo test -p lance-graph-contract --lib`: **603/603 green** (+9); `cargo clippy -p lance-graph-contract --all-targets -- -D warnings`: clean. **No public-API drift in existing code** — `NodeRowPacket`, `NodeRowColumn`, `NODE_ROW_COLUMNS`, `NODE_ROW_STRIDE` are pure additions. This is the keystone the BindSpace dissolution sequence S1-S4 has been blocked behind: Lance's columnar I/O can now read the canonical row packet directly. Next step: MailboxSoA migrating from its column-major `[T; N]` layout to a row-strided `[NodeRow; N]` backing store that impls `SoaEnvelope` through this wrapper.
+
+## 2026-06-13 — SoA migration diff resolution doc (catch-up audit + post-#490 supersession map)
+
+**bardioc cross-session.** Operator directive: *"the biggest goal would be to catch up on all SoA bindspace migration plans and resolve the diff."* Surveyed the SoA / BindSpace / identity plan family (9 plans + 4 board files + 5 code files + canon doc-locks) and produced a single resolution doc at `.claude/plans/soa-migration-diff-resolution-2026-06-13.md` that names every plan-vs-shipped diff post-#487/#489/#490. **Headline drifts named:** (1) `identity-architecture-exists-vs-needs-v1.md §N1`'s UUIDv8 layout fully superseded by OGAR/CLAUDE.md P0 canon — the namespace/entity_type/kind/niblepath_prefix/shape_hash/RFC ceremony framing did NOT ship; canon's classid·HEEL·HIP·TWIG·family·identity (no ceremony) won (PR #489/#490). (2) `bindspace-singleton-to-mailbox-soa-v1.md`'s `CollapseGateEmission` / `MailboxSoA::emit()` / Baton-as-type was retired in PR #487 tombstone; `last_emission_cycle → last_active_cycle` rename per #477 supersession. (3) `unified-soa-convergence-v1.md §4.2` stack pins drifted (`lance =7.0.0` / `lancedb =0.30.0`); 2026-05-29 addendum partially addressed. (4) `polyglot-container-query-membrane-v1.md` ratified research-only — self-describing-key convergence dissolved the membrane question. **D-MBX-A2 status:** still queued, still the gating gap; MailboxSoA<N> has no Hamming columns. **S2-S4 status:** unshipped; `driver.rs:56` still has `pub(crate) bindspace: Arc<BindSpace>`, both `bin/serve.rs:29` + `bin/grpc.rs:29` still call `BindSpace::zeros(4096)`. **SoaEnvelope status:** trait shipped (#477), zero real implementors — only `TestEnvelope` in tests; MailboxSoA does NOT impl it. **Staunen/Wisdom-as-entropy×energy-substrate-state correction with §6.1 alternative framings (Bugwelle / aerodynamics / event horizon / Friston FEP)** — added per operator relay 2026-06-13: (a) Staunen as the entropy *Bugwelle* (bow wave) of thinking-in-progress — Staunen is not static, it's the leading-edge entropy disturbance GENERATED by cognitive motion; (b) Aerodynamic / shock-wave analogy — as cognitive velocity through the substrate increases, the Bugwelle steepens, "sonic boom" = the *aha* breakthrough where entropy collapses abruptly; (c) Event-horizon / inertia — Wisdom's (low entropy × high energy) corner has gravitational properties, novelty needs escape velocity to break out of the well, the canon's "reserve don't reclaim" at classid==0/family==0 keeps the bootstrap basin always-escapable; (d) Friston Free Energy Principle as the scientific anchor — high FE = Staunen, FE-minimisation in progress = Confusion/Chaos quadrant, minimised FE = Wisdom; `consume_firing(row)` IS active inference (energy ≥ threshold ⇒ fire, in-place mark, integrate prediction error). The four framings stack: Bugwelle = shape; aerodynamics = velocity scaling; event-horizon = inertia/why Wisdom resists; FEP = drive function. Same underlying substrate dynamics, four vocabularies for reach. (handover §8 + operator image relays + operator framing relay 2026-06-13). Canonical DIKW = Data → Information → Knowledge → Wisdom, bridged by Processing → Cognition → Judgment; Wisdom IS the canonical DIKW apex rung. The operator's precise framing: **Staunen = high entropy × low energy** = "needs entropy work" marker = cognitive pressure + emerging insight (not yet crystallised). **Wisdom = low entropy × high energy** = crystalline knowledge with supporting plasticity + integrated insights, the substrate has invested heavily and locked it in. Diagonal opposites on the entropy×energy plane, NOT two ends of one axis. The other two quadrants: **Confusion / Chaos** (high entropy × high energy = in-progress climb state, substrate has invested energy but entropy hasn't yet collapsed) and **Boredom / Inert** (low entropy × low energy = ordered but not energised). Substrate column map: Energy = `MailboxSoA.energy: [f32; N]` (signed spatio-temporal accumulator); Plasticity = `MailboxSoA.plasticity_counter: [u8; N]` (saturating Hebbian counter = long-term investment); Entropy proxy = classid-prefix-resolved codebook hit-rate × local edge-neighbourhood density. Two-algebra rule maps onto the plane: entropy axis = signed side (`vsa_bind`), energy axis = magnitude side (`vsa_bundle`). The canon's 3×4 uniform cascade (HEEL · HIP · TWIG = three u16 tiers) shape-matches DIKW's three transitions + four layers — not coincidentally. NOT YET corrected in lance-graph CLAUDE.md (line ~120 still says "Magnitude = Contradiction depth from Staunen × Wisdom qualia") — flagged as `TD-CLAUDE-MD-STAUNEN-MISNAME` for a separate maintenance pass with three specific edits identified (line ~120 rewrite citing entropy×energy markers, §11.5 rephrasing, new DIKW-anchor sub-section under "The Click" mapping cascade tiers onto DIKW transitions + the entropy×energy quadrant diagram). **LE-contract violations still on the books:** `engine_bridge.rs` f32→i4 qualia re-encode, `Vsa16kF32` persisted as cross-boundary in singleton, DTO-as-owned-Vec sites — all dissolve at S2/S4. Errata stubs prepended to 4 affected plans (bindspace-singleton-to-mailbox-soa, identity-architecture-exists-vs-needs, unified-soa-convergence, polyglot-container-query-membrane) pointing at the resolution doc. Resolved punchlist §7 lists 9 follow-up PRs in priority order. Docs-only PR; no code touched.
+
+## 2026-06-13 — #489 canonicalised: wire-in + self-describing Display + retire Phase-A wrapper
+
+**bardioc cross-session.** Operator pin: *"#489 is canonical."* Audited OGAR/CLAUDE.md P0 against `canonical_node.rs` group-by-group. Key (classid·HEEL·HIP·TWIG·family·identity = 8·4·4·4·6·6 hex) matches exactly; RFC-WAIVED matches ("No UUID ceremony"); 3×4 uniform tiers match (each u16; tier-of-level = `level >> 2`); 16-byte EdgeBlock at fixed offset = row-layout analogue of the zero-fallback ladder (default class's default ClassView reserves it, registry-resolved opt-out for non-default classes, "reserve-don't-reclaim" at row level). **One gap closed:** canon mandates *"every printed GUID is self-describing at sight"* via the dash-groups, wrapper had no `Display`. Added `impl Display for NodeGuid` emitting canonical `{classid:08x}-{heel:04x}-{hip:04x}-{twig:04x}-{family:06x}{identity:06x}` (LE in-memory bytes folded through the accessors so hex print is canon-ordered regardless). +2 Display tests. **Phase-A wrapper retired in the same PR** (operator: *"delete #480 from your mind"*): `identity.rs` deleted (UUIDv8 NodeGuid + RFC ceremony bits + IDENTITY_LAYOUT_VERSION + SHAPE_HASH_BITS/LOCAL_BITS — all canon-incompatible per *"wrappers adapt to the canon, never the reverse"*); `pub use identity::{NodeGuid, IDENTITY_LAYOUT_VERSION}` → `pub use canonical_node::{EdgeBlock, NodeGuid, NodeRow}`; `pub mod identity;` removed. Two stale doc references reworded: `hhtl.rs:192` (`from_packed` now a general HHTL utility, not identity::NodeGuid-specific), `lance-graph-ontology/src/registry.rs:405` (`niblepath_of` now points at the canon's `classid·HEEL·HIP·TWIG` resolution). `cargo test -p lance-graph-contract --lib`: **594/594 green** (−10 retired UUIDv8 tests, +2 Display, +8 wire-in canonical_node tests now visible); `cargo check -p lance-graph-ontology`: clean (5 pre-existing `oxrdf::Subject` deprecation warnings, untouched files); `cargo clippy -p lance-graph-contract --all-targets -- -D warnings`: clean. Anchored on #482 (GUID canon) + #489 (canonical_node) + OGAR/CLAUDE.md P0 (the canon itself, *"wrappers audited against this canon group-by-group — never the reverse"*).
 
 ## 2026-06-11 — tombstone commit: emission artifacts removed per PR #477 follow-up
 

.claude/plans/bindspace-singleton-to-mailbox-soa-v1.md

@@ -1,5 +1,7 @@
 # bindspace-singleton-to-mailbox-soa-v1 — dissolve the shared `Arc<BindSpace>` into per-mailbox `MailboxSoA<N>` thoughtspace
 
+> **ERRATA (2026-06-13, post-#490):** D-MBX-A1 columns are shipped (PR #386, mailbox_soa.rs); `last_emission_cycle` is renamed to `last_active_cycle` (PR #477); `CollapseGateEmission` + `MailboxSoA::emit()` are deleted (PR #487); §2.6 DTO inventory predates the TD-RESONANCEDTO-DUP-1 deferral. §5 sequencing (D-MBX-A2 → S1 → S2 → S3 → S4) is still directionally correct; D-MBX-A2 is the current gating gap. Full diff resolution: `soa-migration-diff-resolution-2026-06-13.md`.
+
 > **Status:** CONJECTURE / design (migration spec). NOT yet implemented.
 > **Date:** 2026-05-27.
 > **Owns the answer to:** *"make MailboxSoA the individual, mailbox-owned, ephemeral

.claude/plans/identity-architecture-exists-vs-needs-v1.md

@@ -1,5 +1,7 @@
 # Identity Architecture — What Exists vs What Needs Building (v1)
 
+> **ERRATA (2026-06-13, post-#490):** §N1's proposed UUIDv8 layout (`namespace + entity_type + kind + niblepath_prefix + shape_hash + local + RFC version/variant + layout_version`) is **fully superseded by the operator-pinned canon** (OGAR/CLAUDE.md P0). Canon's `classid · HEEL · HIP · TWIG · family · identity` (8·4·4·4·6·6 hex; 3×4 uniform; RFC-WAIVED; no shape_hash, no layout_version) shipped in `canonical_node.rs` (PR #489) and was canonicalised via lib.rs re-export switch in PR #490; the Phase-A wrapper `identity.rs` was deleted in the same PR (zero in-tree consumers per #480's own body). §N5 `entity_type ↔ NiblePath` bijection: **shipped** in PR #484 (`lance-graph-ontology::registry`). §N3 `SoaEnvelope` impls: still queued. Full diff resolution: `soa-migration-diff-resolution-2026-06-13.md`.
+
 > **Status:** INTEGRATION MAP + PLAN. Grounded by first-hand reads + two parallel
 > cross-repo sweeps (2026-06-09). Companion to
 > `cognitive-write-roundtrip-substrate-v1.md` (the round-trip mechanism).

.claude/plans/polyglot-container-query-membrane-v1.md

@@ -1,5 +1,7 @@
 # Polyglot Container Query Membrane — SurrealQL AST + DataFusion UDF + Cypher over one HHTL address space (v1)
 
+> **ERRATA (2026-06-13, post-#490):** Ratified as **research-only / superseded in spirit**. The "membrane / strategy-registry" framing was superseded in discussion by the self-describing-key convergence (class-in-key makes the cold path already a graph; no membrane needed) and by the GUID-canon PR #482 that crystallised the operator's canon. The verified file:line surface of lance-graph + the surrealdb fork's AST/RecordId/kv-lance remains useful as a grounded inventory but is not a committed direction. Full diff resolution: `soa-migration-diff-resolution-2026-06-13.md`.
+
 > **Status:** RESEARCH MAP + INTEGRATION PLAN. Grounded 2026-06-09 by two parallel
 > repo sweeps (lance-graph + the surrealdb fork) with main-thread spot-verification
 > of every load-bearing claim (one agent claim caught false and corrected, §2.4).