D-RMAP-DIR-1: compiled membership closure vs a 256-slot focus over dir-sim (lab only)

.claude/board/INTEGRATION_PLANS.md

@@ -1,3 +1,7 @@
+## 2026-10-20 — d-rmap-dir-1-membership-closure-focus-v1 — compiled membership closure vs a 256-slot focus over dir-sim → `.claude/plans/d-rmap-dir-1-membership-closure-focus-v1.md`
+
+**Status:** D-RMAP-DIR-1 MEASURED, lab only (`crates/dir-rmap-probe`); D-RMAP-DIR-2 PROPOSED, not authorized. Sharing compiled closure rows across users (B+) is where the acceleration comes from (thousands of times over dir-sim's per-call scan); a 256-slot focus (F) adds at most 2× and only for small foci reused about 1,000 times. Entry `2026-10-20-d-rmap-dir-1-membership-closure-focus.md`.
+
 ## 2026-10-20 — thought-git-basics-v1 — the basics of thinking with a git API → `.claude/plans/thought-git-basics-v1.md`
 
 **Status:** D-TGB-0..6 PROPOSED, lab, test-first; nothing built. From the ideal state: a slot is taken only at the Rubicon crossing (`Planning → CognitiveWork`) and freed at `Commit` / `Plan` / `Prune`, so no stale branch can exist; squash, prune and GC are kanban outcomes; every thought is replayable from its receipt. `Commit` lands one sparse batch through `LanceCycleWriter`, which gives `KanbanColumn::Commit` its declared-but-missing calcify step. Lance 13 supplies `Ref::Version(branch, version)`, branches, tags, `delta()`, `read_version_transaction`; MemWAL is an optional arm (D-WAL, D-LNC-7). Parent: `thought-git-api-v1.md` correction (#1628).

.claude/board/STATUS_BOARD.md

@@ -54,6 +54,8 @@ These PRs merged with no board rows. Every row below is a PROPOSAL (documents an
 | D-RPR-3 | MeTTa-facing parity of a chain's answers | BLOCKED on D-ASM-4 (not authorized) | — |
 | D-MPS-8 | native join over a versioned relationship table: R and S laid out by `y` slot, leapfrog intersection as one mask-risc `And`, per-side generations and reseal, per-thought filter planes (`mork-pathmap-schnellverschluss-v1.md` §7b; `crates/mork-pathmap-probe/src/join.rs`) | MEASURED, probe only; not admitted (D-RMAP-1 BIND membrane review) | 8 join tests + native arms in the differential suite vs the oracle; 7 guards disable-verified; cyclic bodies open |
 | D-RMAP-2 | context-declaration reuse and GraphBLAS/BF16 tile probe (`d-rmap-2-context-tile-algebra-probe-v1.md`) | LAB ONLY, not admitted to production; PARTIAL: §5a MEASURED on the polyfill tier and on the AMX tile tier (§5a-AMX); `VDPBF16PS`-only tier and GraphBLAS arm not run | arms 1–5 in the plan; 297× (D-GTM-0j) is not evidence |
+| D-RMAP-DIR-1 | dir-sim membership: A (View) vs D (crepe oracle) vs B (bidirectional CSR) vs B+ (compiled closure rows + roles by ordinal) vs F (256-slot focus over B+) (`d-rmap-dir-1-membership-closure-focus-v1.md`; `crates/dir-rmap-probe`, CI `dir-rmap-probe.yml`) | MEASURED, lab only; nothing admitted, no production file changed | 5 tests: every arm = A = D on every user/group/chain; 7 guards disable-verified red, 1 shown redundant; B removes A's per-call full scan (W2 ~4,000×), B+ adds 2–3× (groups) and 12–15× (roles), F ≤ 2× only for ≤ 64 focus groups after ~1,000 reuses, none near 256. **Round 2 (§9):** 11 tests; F per chain, refuses incomplete answers; projection through down rows moves F's break-even to 2–100 reuses; F 2–5× warm / ~2× cold while the question masks fit L1, its gain growing with group count (2.9× at 1.5k groups → 5.5× at 11k); slower than B+ once ~4,300 conjunction masks exceed L1; rebuilding per question never pays; not promoted |
+| D-RMAP-DIR-2 | B built in `Snapshot::build` from the lanes; B+ closure as a lazily compiled, version-witnessed cache behind the existing `View` methods; F not proposed | PROPOSED; not authorized; waits on `ISS-ADMISSION-PREP-BINDING` (hot path reads rows at a user's direct-group values) | dir-sim tests + D-RMAP-DIR-1 oracle as the equivalence gate |
 | D-SRAM-0 | inventory Soufflé AST/ast2ram/RAM passes against D-ASM/D-RA/D-TGA owners | Queued | evidence per mapped pass; no code imported |
 | D-SRAM-1 | bound AST → RAM-like schedule for pure conjunctive selection | Queued | same answers and read sets on hand-built fixtures; unsupported ops typed-refused |
 | D-SRAM-2 | cold optimizer rewrites (SCC/strata, magic sets, ordering, dedup) | Queued (research until D-ASM AST placement and D-RA-2); lab spike measured 2026-10-20, `entries/2026-10-20-crepe-rule-optimizer-spike.md`, not authorization | optimized == unoptimized == oracle; rewrite off removes only the plan shape; disabling a semantic guard fails |

.claude/board/entries/2026-10-20-d-rmap-dir-1-membership-closure-focus.md

@@ -0,0 +1,12 @@
+# D-RMAP-DIR-1: compiled membership closure vs a 256-slot focus over dir-sim (2026-10-20)
+
+**Status:** MEASURED, lab only. **Basis:** `crates/dir-rmap-probe` (`tests/equivalence.rs` 5/5, `examples/bench.rs` at 8,000 and 65,000 users, 3 runs at 8,000). Plan `d-rmap-dir-1-membership-closure-focus-v1.md`.
+
+- **Arms:** A dir-sim `View` walks; D crepe Datalog over fixture ids; B membership and nesting CSR; B+ closure rows per group and chain plus roles by ordinal; F B+ projected onto a 256-slot focus (and F+users, one focus row per user).
+- **Equivalence:** B, B+, D equal A on every user, group and chain of the small fixture; B+ equals D on all 79,229 (8,000) and 611,965 (65,000) memberships before timing. 7 guards disable-verified red; the closure's parent-admit check stayed green and is marked an optimisation.
+- **A's cost is a full membership scan per call:** 0.4 ms at 35,000 rows, 0.7 ms at 263,000.
+- **B:** W2 user → groups 92–124 ns vs 408–466 µs; all-user reconciliation 7.1 s → 0.9 ms (8,000 users).
+- **B+:** 32–49 ns per user's groups, 11–15 ns per user's roles; compile 1.6 ms independent of user count; one nesting pair inserted incrementally in 2–2.5 µs.
+- **F:** ≤ 2× faster evaluation than B+ with ≤ 64 focus groups, break-even ~1,000 evaluations per prepared focus; 1.3× at 128; none near 256 and none cold. The focus saves the OR stage only; question tests dominate past a few hundred questions.
+
+OPEN: hot-path read pattern of B/B+/F (reads at a user's direct-group values) under `ISS-ADMISSION-PREP-BINDING`; membership row patching; ordinal remap after a snapshot rebuild; real directory data; D-RMAP-DIR-2 not authorized.

.claude/board/entries/2026-10-20-d-rmap-dir-1-round-2-focus-geometry.md

@@ -0,0 +1,23 @@
+# D-RMAP-DIR-1 round 2: when 256-slot focus geometry beats compiled closure (2026-10-20)
+
+**Status:** MEASURED, lab only; nothing promoted. **Basis:** `crates/dir-rmap-probe` (`tests/equivalence.rs` 11/11; `examples/bench.rs`, 4 runs at 8,000 users plus 1 at 65,000). Plan `d-rmap-dir-1-membership-closure-focus-v1.md` §9.
+
+- **Completeness:**
+  - F binds one chain and projects the full B+ closure through down rows, so paths through unfocused groups are represented (tested on two-hop cases for every chain);
+  - it refuses >256 groups, duplicates, a stale B+ or focus, and unfocused groups;
+  - the witness now compares the `Snapshot` by pointer, so two stores both starting at `VersionId(0)` refuse each other;
+  - 8 guards disable-verified red.
+- **Preparation:**
+  - down-row projection costs 1.6–2.7 µs at 64 groups, against 72–82 µs for round 1's per-slot test, and that moves F's break-even from ~1,000 reuses to 2–100;
+  - eager per-user rows cost 84–95 µs and 32 B per user; lazy rows fill at ~16 ns per user;
+  - every arm's hot path makes 0 allocations.
+- **Working set:**
+  - B+'s closure passing L2 cost it only 5–17 % (warm vs hot); it slows with row width (218 → 436 ns from 1.5k to 11k groups);
+  - F stays at 74–79 ns, a 2.9× → 5.5× advantage;
+  - cold: F ~2×, F+users ~4×.
+- **Question shape decides:**
+  - composite membership (one bit): F 1.7–2.9× at every focus size;
+  - conjunctions: 2.4–5.4× up to ~650 questions, but at 256 groups (~4,300 conjunctions, 135 KiB of masks, more than L1d) F's mask form is 1.5–1.7× *slower* than B+.
+- **Rebuilding a focus per question never pays;** question-aware B+ (needed words only) never helps.
+
+OPEN: real directories' question-to-group ratio; other CPUs (small L3); the gather admission (`ISS-ADMISSION-PREP-BINDING`); D-RMAP-DIR-2 not authorized.

.claude/board/entries/README.md

@@ -25,7 +25,7 @@ index row, (3) no duplicate entry id. Checks 1 and 2 are deliberately
 opposite directions; the stranding this convention prevents shows up in
 exactly one of them, never both.
 
-389 entries, 2026-08-06 .. 2026-10-20.
+391 entries, 2026-08-06 .. 2026-10-20.
 
 | date | entry id | finding | file |
 |---|---|---|---|
@@ -39,6 +39,8 @@ exactly one of them, never both.
 | 2026-10-20 | `D-TGA-SLOT-AXIS` | thoughts are 64 slots, and the slot number is CE64's witness pointer | [2026-10-20-d-tga-slot-axis-witness-pointer.md](2026-10-20-d-tga-slot-axis-witness-pointer.md) |
 | 2026-10-20 | `D-RPR` |  | [2026-10-20-d-rpr-consolidation-1624.md](2026-10-20-d-rpr-consolidation-1624.md) |
 | 2026-10-20 | `D-RPR-1` |  | [2026-10-20-d-rpr-1-handoff-lab-demo.md](2026-10-20-d-rpr-1-handoff-lab-demo.md) |
+| 2026-10-20 | `D-RMAP-DIR-1` |  | [2026-10-20-d-rmap-dir-1-round-2-focus-geometry.md](2026-10-20-d-rmap-dir-1-round-2-focus-geometry.md) |
+| 2026-10-20 | `D-RMAP-DIR-1` |  | [2026-10-20-d-rmap-dir-1-membership-closure-focus.md](2026-10-20-d-rmap-dir-1-membership-closure-focus.md) |
 | 2026-10-20 | `D-RMAP-2` |  | [2026-10-20-d-rmap-2-tile-gemm-cell.md](2026-10-20-d-rmap-2-tile-gemm-cell.md) |
 | 2026-10-20 | `D-RMAP-2-AMX` |  | [2026-10-20-d-rmap-2-amx-tier.md](2026-10-20-d-rmap-2-amx-tier.md) |
 | 2026-10-20 | `D-RMAP-2-AMX-RERUN` | §5a-AMX re-measured on a second AMX host | [2026-10-20-d-rmap-2-amx-remeasured.md](2026-10-20-d-rmap-2-amx-remeasured.md) |