D-WXS-2a half A: the KILL fires — row-major vs Morton closed, no code change

.claude/board/STATUS_BOARD.md

@@ -12,7 +12,7 @@ with a control that can lose and a stay-silent twin.
 | D-WXS-1 | field manifest v1 — (facet, pair, byte) → (variable, level, unit, floor id), a committed data artifact, ClassView-side | 0 | **SHIPPED 2026-08-13** — `data/field_manifest_v1.tsv` (22 rows = F0 5 pairs + F1/F2 3 pairs each, reserved slots emit NO row) + `manifest.rs`, 13/13; collision guard **disable-verified** (removed → only `colliding_entries_are_rejected` fails, BOTH stay-silent twins stay green). Bar B0's end-to-end half (mutating an entry changes written bytes) DEFERRED — the bake does not exist yet | slot purity §2; bar B0 |
 | D-WXS-1a | variable census as a committed re-runnable probe (17 surface + 91 upper-air + 14 static = 122 fields; 92,044 six-hourly steps) | 0 | **SHIPPED 2026-08-13** — `era5_variable_census.py` + `.json`; `--selftest` PASS on all 10 constants, orchestrator-rerun independently; guard disable-verified (one constant broken → exit 1, correct message) | ends the chat-only-figure defect for the census |
 | D-WXS-2 | key codec `(lat,lon) ↔ NodeGuid` — HEEL 16° tile / HIP within-tile / TWIG dormant; ragged tiles; lon-wrap range-SET | 1 | **SHIPPED 2026-08-13** — `key.rs`, 5/5 green; exhaustive 1,038,240-cell round-trip + collision-free; both bar-B1 halves **disable-verified** by the orchestrator (zeroing the HIP lat byte kills 3 tests incl. collision + ragged; removing the seam split kills the wrap twin while the non-wrap twin stays green) | a 16° box becomes a HEEL-prefix scan; bar B1 |
-| D-WXS-2a | **NEW — row-major vs Morton, pre-registered comparison.** The shipped key assigns one WHOLE byte per axis; OGAR's cascade doctrine specifies the axis bytes **nibble-interleaved** (Morton). §1.2 deviated from the canon it cites and did not say so — now recorded as plan §1.3a. The prefix-scan claim holds under both; what differs is neighbour locality (`lat ± 1` is 1440 cells away under row-major) and how many ranges a non-tile-aligned box needs | 1 | **Bar PRE-REGISTERED 2026-08-13** (plan §1.3b, committed before the run). Half A (pure key-space: range count + neighbour locality; arms SHIPPED/MORTON/CONTROL-BAD) is runnable NOW. Half B (the ζ stencil) is gated on D-WXS-9 → D-WXS-0 | gates any downstream assumption of Morton locality — measured against the ζ stencil (D-WXS-9), metric stated before the run |
+| D-WXS-2a | **NEW — row-major vs Morton, pre-registered comparison.** The shipped key assigns one WHOLE byte per axis; OGAR's cascade doctrine specifies the axis bytes **nibble-interleaved** (Morton). §1.2 deviated from the canon it cites and did not say so — now recorded as plan §1.3a. The prefix-scan claim holds under both; what differs is neighbour locality (`lat ± 1` is 1440 cells away under row-major) and how many ranges a non-tile-aligned box needs | 1 | **CLOSED by KILL 2026-08-13** — half A run, `examples/layout_probe.rs` + `.json`. **primary FAIL** (MORTON won ONE metric, not both: neighbour locality 16 vs 32 = 2x better, range count 212.50 vs 140.00 = ~1.5x worse). **control PASS** (CONTROL-BAD 3100 ranges / 15862 neighbour — the metrics do measure locality). **stay-silent twin PASS** (tile-aligned = exactly 1 range under BOTH, `[1,1,1,1]`). Pre-registered consequence applied: **no code change, no migration**. §1.3b's word "harmless" corrected in §1.3c — MORTON wins on exactly the metric half B would care about, so the honest reading is *no unambiguous win in key space* | gates any downstream assumption of Morton locality — measured against the ζ stencil (D-WXS-9), metric stated before the run |
 | D-WXS-3 | shared canonical floor calibration (global 0.4–99.6 pct, frozen per epoch, stamped in dataset metadata) | 1 | **SHIPPED 2026-08-13** — `floor.rs`; bar B2 **disable-verified** (widening the "narrow" control floor kills only the control, twin stays green); version-stamp mismatch detected, ±½-bucket round-trip asserted. **AMENDED same day** (`E-A-TOTAL-FUNCTION-THAT-CANNOT-REFUSE-IS-A-CORRUPTION-PATH-1`): `saturation_of` folded non-finite input into the metric — `quantize` sends `NaN`/`-inf`→0 and `+inf`→255, all **rim** buckets, so an all-`NaN` population scored **1.0** ("fully saturated") where the truth is "no data at all". Now returns `SaturationScore {fraction, finite, non_finite}` — reported, never folded and never silently dropped. `calibrate` checked **CLEAN** (already filters `is_finite`) | bar B2 — this is the bar's own INSTRUMENT, so the defect would have corrupted a measurement, not a value |
 | D-WXS-3b | **NEW — the L4 lane (pack/unpack ONE 16-byte facet).** The plan gave the lane a worker in §6.2 but **no D-id in §4's ladder** — it jumped D-WXS-3 → D-WXS-4. Added here as the pack/unpack half the bake will call | 1 | **SHIPPED 2026-08-13** — `lane.rs`, 33/33 crate-wide; 4 disables verified by the orchestrator (lo/hi swap → the swap test; hard-coded slot → 3 tests incl. manifest-load-bearing; version guard bypassed → the version test; unmapped slots emitting values → the reserved-slot test). The lane names no ERA5 variable in its own source — the caller's closure owns that **AMENDED same day (codex P1, PR #948):** `pack_facet` accepted non-finite readings; `quantize` maps them to valid-looking buckets, so a missing ARCO-ERA5 chunk (all-`NaN` — **valid store semantics**, and five W1 variables 404 at the arc's own fixture timestep) would have been stored as plausible low-bucket measurements. Now `LaneError::NonFiniteValue`, covering `±inf` too since they land on the rim. Disable-verified | precursor to bar B3; §2.6 slot purity as code |
 | D-WXS-4 | the bake: one timestep → 1,038,240 NodeRows → ONE Lance version | 1 | Queued — **blocked behind D-WXS-0** (must refuse to write without a minted classid) | bar B3; the missing path |

.claude/plans/weather-soa-bake-v1.md

@@ -261,6 +261,75 @@ whether a difference exists **in key space at all**. If half A kills, half B is
 moot; if half A confirms, half B still has to run before any migration. Stated
 now so a green half A is not later read as a mandate.
 
+### §1.3c `D-WXS-2a` half A — RUN 2026-08-13. **The KILL fires. Closed, no code change.**
+
+Probe: `crates/weather-poc/examples/layout_probe.rs` (+ `--json` output committed
+beside it). Bar: §1.3b, committed **before** this run (`d83b4d3e`). Selftest run
+first, green — including the check the brief demanded be *worked out rather than
+assumed*: a full longitude row under SHIPPED is **23 ranges** (one per longitude
+tile), not 1.
+
+#### Measured
+
+| arm | rc median (non-tile-aligned) | neighbour-locality median |
+|---|---|---|
+| **SHIPPED** | **140.00** | 32.00 |
+| **MORTON** | 212.50 | **16.00** |
+| CONTROL-BAD | 3100.00 | 15862.00 |
+
+Box set: 20 boxes — 4 tile-aligned, 8 non-tile-aligned, 4 seam-crossing,
+4 pole-adjacent. Neighbour sample: deterministic stride 677, n = 1534.
+
+#### The three verdicts, as pre-registered
+
+- **primary — FAIL.** The bar required MORTON to beat SHIPPED on **both**
+  metrics. It won one and lost one: **2× better** on neighbour locality
+  (16 vs 32) and **~1.5× worse** on range count (212.50 vs 140.00).
+- **control — PASS.** CONTROL-BAD is far worse on both (3100 ranges;
+  neighbour distance 15862). The metrics therefore measure locality rather
+  than nothing — without this the primary result would be unreadable.
+- **stay-silent twin — PASS.** Every tile-aligned box is **exactly one range
+  under both** arms (`[1,1,1,1]` each). §1.2's load-bearing claim — *a 16° box
+  is a HEEL-prefix scan* — holds, and shows **no difference where the plan
+  claims none**.
+
+#### Consequence (the pre-registered one, unchanged)
+
+**`D-WXS-2a` closes. No code change. §1.3a's layout stands as shipped.** The
+migration is not opened.
+
+#### ⚠ A correction to §1.3b's own wording, not a softening of the verdict
+
+§1.3b phrased the KILL as *"the deviation is **harmless for this workload**"*.
+The data shows that phrasing was imprecise, and it is corrected here rather
+than left to read as more than was measured:
+
+**MORTON wins, 2×, on exactly the metric half B would care about.** Neighbour
+locality is what a ζ stencil spends; range count is what a box scan spends. So
+the honest statement is **"no unambiguous win in key space"**, not "harmless".
+
+The verdict is untouched: the bar said *both*, it got *one*, and one is a FAIL.
+Re-reading a split as a win is the failure this arc has a rule against
+(*"re-normalising a metric until it stops showing the confound is how a
+confound becomes a finding"*). What changes is only my own summary sentence,
+which claimed more than the measurement.
+
+**The prior this leaves for half B**, if the ζ stencil is ever measured under
+both layouts (gated on `D-WXS-9` → `D-WXS-0`): MORTON is expected to win there,
+by roughly 2×, because that is what the locality metric already shows. That is
+a **prior, not a result** — and a 2× locality win still would not by itself
+justify a migration, because the same measurement shows it costs ~1.5× on
+range count.
+
+#### Why SHIPPED's numbers look the way they do (arithmetic, checked)
+
+SHIPPED orders `[lat_tile, lon_tile, lat_hip, lon_hip]`, so within one tile the
+order is `lat_hip`-major / `lon_hip`-minor. A `lon ± 1` neighbour is therefore
+**1** apart and a `lat ± 1` neighbour is **64** apart; the median over the
+4-neighbourhood is the median of `[1, 1, 64, 64]` = **32** — exactly the
+measured value. The two-level blocked order §1.3a identified (rather than the
+flat "row-major" I first wrote) is what keeps its range count competitive.
+
 ### §1.4 classid — a mint decision, NOT taken here
 
 `0x0F = Geo` already exists in the OGAR domain table; free domains are `0x03–0x06`

crates/weather-poc/examples/layout_probe.json

@@ -0,0 +1,60 @@
+{
+  "arms": ["shipped", "morton", "control_bad"],
+  "sample_stride": 677,
+  "box_set": [
+    {"kind": "tile_aligned", "lat_lo": 128, "lat_hi": 192, "lon_lo": 0, "lon_hi": 64},
+    {"kind": "tile_aligned", "lat_lo": 320, "lat_hi": 384, "lon_lo": 640, "lon_hi": 704},
+    {"kind": "tile_aligned", "lat_lo": 512, "lat_hi": 576, "lon_lo": 1344, "lon_hi": 1408},
+    {"kind": "tile_aligned", "lat_lo": 192, "lat_hi": 256, "lon_lo": 960, "lon_hi": 1024},
+    {"kind": "non_tile_aligned", "lat_lo": 10, "lat_hi": 37, "lon_lo": 5, "lon_hi": 53},
+    {"kind": "non_tile_aligned", "lat_lo": 100, "lat_hi": 233, "lon_lo": 200, "lon_hi": 410},
+    {"kind": "non_tile_aligned", "lat_lo": 50, "lat_hi": 521, "lon_lo": 33, "lon_hi": 999},
+    {"kind": "non_tile_aligned", "lat_lo": 400, "lat_hi": 405, "lon_lo": 700, "lon_hi": 1300},
+    {"kind": "non_tile_aligned", "lat_lo": 20, "lat_hi": 680, "lon_lo": 900, "lon_hi": 905},
+    {"kind": "non_tile_aligned", "lat_lo": 65, "lat_hi": 129, "lon_lo": 63, "lon_hi": 129},
+    {"kind": "non_tile_aligned", "lat_lo": 200, "lat_hi": 205, "lon_lo": 1, "lon_hi": 2},
+    {"kind": "non_tile_aligned", "lat_lo": 300, "lat_hi": 450, "lon_lo": 1100, "lon_hi": 1150},
+    {"kind": "seam_crossing", "lat_lo": 100, "lat_hi": 200, "lon_lo": 1400, "lon_hi": 40},
+    {"kind": "seam_crossing", "lat_lo": 300, "lat_hi": 350, "lon_lo": 1430, "lon_hi": 10},
+    {"kind": "seam_crossing", "lat_lo": 500, "lat_hi": 600, "lon_lo": 1439, "lon_hi": 1},
+    {"kind": "seam_crossing", "lat_lo": 50, "lat_hi": 70, "lon_lo": 1300, "lon_hi": 100},
+    {"kind": "pole_adjacent", "lat_lo": 0, "lat_hi": 30, "lon_lo": 100, "lon_hi": 200},
+    {"kind": "pole_adjacent", "lat_lo": 700, "lat_hi": 721, "lon_lo": 500, "lon_hi": 600},
+    {"kind": "pole_adjacent", "lat_lo": 0, "lat_hi": 1, "lon_lo": 0, "lon_hi": 1440},
+    {"kind": "pole_adjacent", "lat_lo": 715, "lat_hi": 721, "lon_lo": 900, "lon_hi": 950}
+  ],
+  "results": {
+    "shipped": {
+      "by_kind": {
+        "tile_aligned": {"range_counts": [1, 1, 1, 1], "median": 1.0000, "max": 1},
+        "non_tile_aligned": {"range_counts": [27, 269, 963, 19, 660, 130, 5, 150], "median": 140.0000, "max": 963},
+        "seam_crossing": {"range_counts": [200, 99, 198, 45], "median": 148.5000, "max": 200},
+        "pole_adjacent": {"range_counts": [61, 42, 23, 6], "median": 32.5000, "max": 61}
+      },
+      "neighbour_locality_median": 32.0000
+    },
+    "morton": {
+      "by_kind": {
+        "tile_aligned": {"range_counts": [1, 1, 1, 1], "median": 1.0000, "max": 1},
+        "non_tile_aligned": {"range_counts": [57, 289, 1053, 47, 660, 136, 5, 295], "median": 212.5000, "max": 1053},
+        "seam_crossing": {"range_counts": [208, 100, 200, 68], "median": 150.0000, "max": 208},
+        "pole_adjacent": {"range_counts": [65, 46, 90, 13], "median": 55.5000, "max": 90}
+      },
+      "neighbour_locality_median": 16.0000
+    },
+    "control_bad": {
+      "by_kind": {
+        "tile_aligned": {"range_counts": [4096, 4096, 4096, 4096], "median": 4096.0000, "max": 4096},
+        "non_tile_aligned": {"range_counts": [1296, 13440, 61824, 3000, 320, 4224, 5, 3200], "median": 3100.0000, "max": 61824},
+        "seam_crossing": {"range_counts": [5120, 1000, 128, 4800], "median": 2900.0000, "max": 5120},
+        "pole_adjacent": {"range_counts": [3000, 2100, 1440, 300], "median": 1770.0000, "max": 3000}
+      },
+      "neighbour_locality_median": 15862.0000
+    }
+  },
+  "verdicts": [
+    {"name": "primary", "pass": false, "detail": "morton_rc_median=212.50 vs shipped_rc_median=140.00 (non-tile-aligned); morton_nb_median=16.00 vs shipped_nb_median=32.00"},
+    {"name": "control", "pass": true, "detail": "control_rc_median=3100.00 vs shipped=140.00/morton=212.50 (non-tile-aligned); control_nb_median=15862.00 vs shipped=32.00/morton=16.00"},
+    {"name": "stay-silent twin", "pass": true, "detail": "shipped_tile_aligned_counts=[1, 1, 1, 1]; morton_tile_aligned_counts=[1, 1, 1, 1] (every entry must be 1)"}
+  ]
+}