From c98010058908cc009663097106f07f48323554a9 Mon Sep 17 00:00:00 2001 From: Bruno Postle Date: Fri, 17 Jul 2026 16:33:02 +0100 Subject: [PATCH] =?UTF-8?q?kpu:=20Schedule=20B=20A/B=20DONE=20=E2=80=94=20?= =?UTF-8?q?graduated=20grain=20ramp=20falsified=20(negative)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit harbor-house 3M (500k/grain x3 + 1.5M polish, workers 4, ~22h): the in-run grain anneal reached 1.26e-08 / 23 fails (canonical byte-for-byte), losing decisively to both the direct --no-leaf-sharing baseline (5.14e-06 / 15) and yaa's single-hard-transition warm chain (4.19e-06 / 15) — ~400x worse fitness, +8 fails. Each grain step spikes the fail count as its unfolded leaves acquire independent shape fails (phase-end 19->21->27, final de-share 27->36); the per-phase budget re-polishes a partially-materialised state the next step materialises further, so coarse-grain gains do not carry forward. The polish phase started from a deeper hole (36) than the warm chain's single clean transition and 1.5M evals recovered only to 23. The sharing-phase topology skeleton is best cashed in once, at full grain — not annealed. Machinery retained (search_annealed, --anneal-grain, unfold above=, seed_pop, max_share override): correct, tested, honest, reusable. Default finish stays §15's single-transition unfold+polish. DESIGN §16 records the verdict; closes homemaker-py-kpu. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_01M8566xAxTnwtJTkpXjYNZm --- .beads/issues.jsonl | 34 +++++++++++++++++----------------- DESIGN.md | 36 +++++++++++++++++++++++++++++++----- 2 files changed, 48 insertions(+), 22 deletions(-) diff --git a/.beads/issues.jsonl b/.beads/issues.jsonl index b5f2e17..9fc7ca2 100644 --- a/.beads/issues.jsonl +++ b/.beads/issues.jsonl @@ -52,7 +52,7 @@ {"id":"homemaker-py-nyb","title":"High-locality topology operators (mutation + subtree crossover)","description":"DESIGN.md §5, §7 Phase 2, §8.4. Mutation moves: divide/undivide leaf, swap children, rotate cut, retype leaf, per-floor delta edits, storey add/delete (cf. Urb Mutate.pm — but geometry sliding belongs to the inner loop, not the operator set). Crossover: area-matched subtree exchange (a subtree = a contiguous region, so crossover is meaningful — Crossover.pm). Operators must be high-locality: small genome change =\u003e small phenotype change, so warm-started inner loops stay cheap.","acceptance_criteria":"Each operator produces valid genomes (oracle scores them without error); locality measured (mean fitness/geometry perturbation per operator)","status":"closed","priority":2,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-06-11T23:37:27Z","created_by":"Bruno Postle","updated_at":"2026-06-12T13:07:37Z","started_at":"2026-06-12T12:54:23Z","closed_at":"2026-06-12T13:07:37Z","close_reason":"operators.py lands: 7 mutations + area-matched crossover, valid-by-construction via genome.encode repair. 115/115 oracle-valid children; locality measured: geom-pert 0.07-0.33 per op, fitness-pert 0.68-0.99 (0.5^n cliff flags raw moves — warm restart + penalty reshaping confirmed load-bearing). Also fixed dom._link stale below-links on structural mutation.","dependencies":[{"issue_id":"homemaker-py-nyb","depends_on_id":"homemaker-py-k2g","type":"blocks","created_at":"2026-06-12T00:39:36Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":1,"comment_count":0} {"id":"homemaker-py-k2g","title":"Topology genome: base-floor tree + per-floor deltas + type assignment","description":"DESIGN.md §5.2, §7 Phase 2. Genome = base-floor slicing topology (primary) + per-leaf type assignment + per-floor divide/undivide deltas (Below-inheritance as regulariser; cut owned by lowest storey where its path is divided — §10). Must round-trip to/from dom.py Node trees so the oracle and inner loop consume it directly. Includes storey count and per-floor type overrides.","acceptance_criteria":"Genome \u003c-\u003e .dom round-trip on all 35 corpus files preserves fitness; multi-storey wall stacking preserved","status":"closed","priority":2,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-06-11T23:37:26Z","created_by":"Bruno Postle","updated_at":"2026-06-12T12:52:34Z","started_at":"2026-06-12T10:55:21Z","closed_at":"2026-06-12T12:52:34Z","close_reason":"genome.py encode/decode lands. 35/35 oracle fitness parity after round-trip (flag-on); genome fixed-point + owned-projection tests. Dead-field discovery: corpus upper storeys carry drifted dead divisions (97) and rotations (187) — canonicalised by decode, validated fitness-neutral.","dependency_count":0,"dependent_count":1,"comment_count":0} {"id":"homemaker-py-d0s","title":"Experiment: inner-loop optimiser bake-off at equal oracle budgets","description":"DESIGN.md §7 Phase 1, §8.3. DOF is only ~rooms-1 (6–7 on corpus). Compare Nelder-Mead vs CMA-ES vs batched multi-start pattern search at equal oracle-call budgets, measuring fitness gained per oracle call and wall-clock (batch-friendliness matters — §4.6). Measure, don't commit blind.","acceptance_criteria":"Table of fitness-per-budget across \u003e=3 candidates; one optimiser chosen and recorded in DESIGN.md","status":"closed","priority":2,"issue_type":"task","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-06-11T23:36:59Z","created_by":"Bruno Postle","updated_at":"2026-06-13T08:48:13Z","started_at":"2026-06-12T21:22:15Z","closed_at":"2026-06-13T08:48:13Z","close_reason":"Bake-off complete: CMA-ES confirmed as Phase 1/2 optimiser. NM wins quality per eval but sequential architecture incompatible with batching (§4.6). Compass stalls on narrow valleys. Results in DESIGN.md §8.3 and experiments/bakeoff_innerloop.*","dependencies":[{"issue_id":"homemaker-py-d0s","depends_on_id":"homemaker-py-1p0","type":"blocks","created_at":"2026-06-12T00:39:35Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":0,"comment_count":0} -{"id":"homemaker-py-kpu","title":"Schedule B: in-run leaf-sharing annealing (ramp grain down, unfold at each step)","description":"Spun out of homemaker-py-yaa, whose investigation is complete. yaa proved Schedule A (two-phase warm-start) works ONLY when shared leaves are unfolded at the sharing-\u003eno-sharing transition: naive warm-start stalls at 8.66e-08/70 fails, but unfold-then-de-share reaches 4.19e-06/15 fails — matching the direct --no-leaf-sharing baseline. operators.unfold_shared_leaves() is built, tested, and proven.\n\nSchedule B is the in-run variant: instead of a manual two-phase chain, anneal leaf_share_factor down within a single driver run (e.g. 4-\u003e3-\u003e2-\u003eoff) at eval thresholds. At each grain transition: (1) rebuild the cached (dir,sharing) evaluator at the new grain, (2) UNFOLD shared leaves that drop below the new grain so the population stays materialised (reuse operators.unfold_shared_leaves), (3) re-evaluate the whole population under the new evaluator, (4) resume local search. Gradual grain ramp = graduated non-convexity: avoids a single fitness cliff, keeps gross topology fixed on the smaller effective problem early, polishes per-room size/proportion/width late.\n\nDriver hooks needed (driver.py): the evaluator is cached per (dir, sharing) at fitness.py:415 and driver caches one per worker; the ramp must rebuild it and re-score the pop at each threshold. Modest change. Compare head-to-head vs (a) direct baseline 5.14e-06 and (b) the manual unfold warm-chain 4.19e-06 from yaa — does a graduated ramp beat a single hard unfold transition?\n\nWants the circulation-aware unfold from homemaker-py-8iv once available.","notes":"IMPLEMENTED (2026-07-16): driver.search_annealed + CLI --anneal-grain. Grain ramp lowers evaluator leaf_share_max (new max_share override) per phase; unfold_shared_leaves(above=cap) materialises only leaves exceeding the new cap; search(seed_pop=) carries the whole population across each step; de-share polish finish (honest, canonical byte-for-byte verified). 8iv settled the primitive: reuse plain grid unfold (slice lost). Tests: 258 pass (+4). DESIGN §16. Head-to-head A/B RUNNING on harbor-house (3M total: 500k/grain x3 + 1.5M polish, seed 0, pop 16) vs baseline 5.14e-06 and warm-chain 4.19e-06 — verdict pending.","status":"in_progress","priority":3,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-15T06:48:11Z","created_by":"Bruno Postle","updated_at":"2026-07-16T07:35:57Z","started_at":"2026-07-16T06:47:57Z","dependencies":[{"issue_id":"homemaker-py-kpu","depends_on_id":"homemaker-py-8iv","type":"blocks","created_at":"2026-07-15T07:48:30Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-kpu","depends_on_id":"homemaker-py-yaa","type":"blocks","created_at":"2026-07-15T07:48:28Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":2,"dependent_count":0,"comment_count":0} +{"id":"homemaker-py-kpu","title":"Schedule B: in-run leaf-sharing annealing (ramp grain down, unfold at each step)","description":"Spun out of homemaker-py-yaa, whose investigation is complete. yaa proved Schedule A (two-phase warm-start) works ONLY when shared leaves are unfolded at the sharing-\u003eno-sharing transition: naive warm-start stalls at 8.66e-08/70 fails, but unfold-then-de-share reaches 4.19e-06/15 fails — matching the direct --no-leaf-sharing baseline. operators.unfold_shared_leaves() is built, tested, and proven.\n\nSchedule B is the in-run variant: instead of a manual two-phase chain, anneal leaf_share_factor down within a single driver run (e.g. 4-\u003e3-\u003e2-\u003eoff) at eval thresholds. At each grain transition: (1) rebuild the cached (dir,sharing) evaluator at the new grain, (2) UNFOLD shared leaves that drop below the new grain so the population stays materialised (reuse operators.unfold_shared_leaves), (3) re-evaluate the whole population under the new evaluator, (4) resume local search. Gradual grain ramp = graduated non-convexity: avoids a single fitness cliff, keeps gross topology fixed on the smaller effective problem early, polishes per-room size/proportion/width late.\n\nDriver hooks needed (driver.py): the evaluator is cached per (dir, sharing) at fitness.py:415 and driver caches one per worker; the ramp must rebuild it and re-score the pop at each threshold. Modest change. Compare head-to-head vs (a) direct baseline 5.14e-06 and (b) the manual unfold warm-chain 4.19e-06 from yaa — does a graduated ramp beat a single hard unfold transition?\n\nWants the circulation-aware unfold from homemaker-py-8iv once available.","notes":"A/B DONE — NEGATIVE (2026-07-17). harbor-house 3M (500k/grain x3 + 1.5M polish, workers 4, ~22h): Schedule B = 1.26e-08 / 23 fails (canonical byte-for-byte). Loses decisively to both targets: direct baseline 5.14e-06/15 and warm-chain 4.19e-06/15 (~400x worse, +8 fails). Graduated ramp FALSIFIED: each grain step spikes fails (phase-end 19-\u003e21-\u003e27, final unfold 27-\u003e36); per-phase budget re-polishes partially-materialised states that the next step materialises further, so coarse-grain gains don't carry forward; polish started from a deeper hole (36) than the warm chain's single clean transition and only reached 23. The sharing-phase topology skeleton (yaa) is best cashed in ONCE at full grain, not annealed. Machinery retained (search_annealed, --anneal-grain, unfold above=, seed_pop, max_share override) — correct/tested/honest — but §15 single-transition finish stays the default. DESIGN §16 updated. Closing.","status":"closed","priority":3,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-15T06:48:11Z","created_by":"Bruno Postle","updated_at":"2026-07-17T15:33:01Z","started_at":"2026-07-16T06:47:57Z","closed_at":"2026-07-17T15:33:01Z","close_reason":"Closed","dependencies":[{"issue_id":"homemaker-py-kpu","depends_on_id":"homemaker-py-8iv","type":"blocks","created_at":"2026-07-15T07:48:30Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-kpu","depends_on_id":"homemaker-py-yaa","type":"blocks","created_at":"2026-07-15T07:48:28Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":2,"dependent_count":0,"comment_count":0} {"id":"homemaker-py-8iv","title":"unfold: route circulation to interior children (access/adjacency fails)","description":"Follow-up to homemaker-py-yaa. operators.unfold_shared_leaves() materialises a share=k leaf into a BALANCED binary subtree of k equal-target children. This closes the count deficit (all critical missing-room fails) but the balanced split creates interior children with no direct edge onto a corridor, so it introduces access/adjacency polish fails. Measured on harbor-house evolved-3M.dom: after unfold, 0 critical but 59 total fails, of which ~14 access + ~12 adjacency are attributable to the unrouted interior rooms (18 size / 2 width / 2 proportion are the k*target-\u003eper-leaf sizing mismatch, a separate concern).\n\nIdea: make the unfold subdivision circulation-aware instead of purely balanced — bias each cut so every new child retains an edge onto the shared leaf's original access boundary (or onto a sibling circulation leaf), mirroring the adjacency-aware constructive seeder (§11.6). Options: (a) orient/order the k-leaf subtree so children fan off the corridor side rather than nesting inward; (b) reserve a thin circulation spine within the unfolded block; (c) let a few post-unfold local-search evals fix it (cheaper, but that is exactly what the warm-start already does). Compare final endpoint with/without circulation-aware unfold against the 3M direct baseline (5.14e-06).\n\nRelates to the in-run annealing driver change (Schedule B, option B in yaa): if annealing rebuilds+re-evaluates the population at each grain transition, the unfold used there wants the same circulation-aware subdivision.","notes":"A/B VERDICT (seed 0, budget 150k, 4 workers, warm-start no-sharing polish from\nevolved-3M.dom): GRID WINS DECISIVELY. Circulation-aware slice LOSES.\n slice: 41 fails, fitness 3.52e-14\n grid : 25 fails, fitness 2.36e-09 (~5 orders better, 16 fewer fails)\nGrid led at EVERY milestone and the gap widened, not a near-tie:\n ~12k evals slice 67 / grid 49; ~36k slice 56 / grid 39;\n ~85k slice 46 / grid 30; ~130k slice 41 / grid 25.\nSlice never crossed. The thin-slab geometric debt (proportion/long/width) from\nforcing all k rooms onto one corridor wall costs MORE than the access routing\nsaves: local search re-routes access via topology moves (level_retype,\nplace_missing, level_fix) faster than it can widen thin slices (which it can't,\nwithout topology change — k equal slices of a compact leaf are intrinsically\nthin). Grid's squarer children are the better warm-start; yaa already showed grid\nreaches 4.19e-06.\n\nCONCLUSION: circulation-aware slicing is the WRONG trade. Retain the grid unfold.\nThe 8iv hypothesis (route access at unfold time) is falsified for the warm-start\nregime — access is better left to local search on a squarer seed. n=1 but the\ngap is large and monotone across the whole 150k-eval trajectory.","status":"closed","priority":3,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-12T15:38:34Z","created_by":"Bruno Postle","updated_at":"2026-07-16T06:35:46Z","started_at":"2026-07-15T13:46:49Z","closed_at":"2026-07-16T06:35:46Z","close_reason":"Investigated and falsified. Circulation-aware unfold (slice shared leaves perpendicular to their access edge so every child touches the corridor) was implemented + unit-tested, but the warm-start A/B (evolved-3M seed, 150k-eval no-sharing polish) shows it LOSES decisively to the existing balanced grid: slice 41 fails/3.5e-14 vs grid 25 fails/2.4e-09, grid leading monotonically at every milestone. Forcing k rooms onto one wall makes intrinsically thin slices whose geometric debt (proportion/long/width) local search cannot pay down without topology change, whereas grid's squarer children let local search re-route access cheaply via level_retype/place_missing/level_fix. Conclusion: retain grid unfold; access is better left to local search on a squarer seed. Code reverted (operators.py, test_operators.py back to grid). Findings in issue notes; A/B traces in examples/harbor-house/ab-8iv-*.","dependencies":[{"issue_id":"homemaker-py-8iv","depends_on_id":"homemaker-py-yaa","type":"blocks","created_at":"2026-07-12T16:48:16Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":1,"comment_count":0} {"id":"homemaker-py-yaa","title":"Investigate leaf-sharing annealing: shared early, materialise-and-de-share later","description":"Follow-up to homemaker-py-3l6. Leaf sharing is a fitness-evaluation knob over an identical genome representation (fitness.py:415; driver caches one evaluator per (dir, sharing)), and leaf_share_factor is a *grain* (0/1=off, N\u003e=2=share at grain N). That makes a coarse-to-fine / continuation schedule feasible: keep sharing on early to fix gross topology (level connectivity, adjacencies, massing) on a smaller effective problem, then reduce sharing to polish per-room size/proportion/width.\n\nTwo schedules to evaluate:\n A. Two-phase warm-start (no code): run sharing to convergence, feed the output .dom as seed to a --no-leaf-sharing run.\n B. In-run annealing (modest driver change): ramp leaf_share_factor down (e.g. 4-\u003e3-\u003e2-\u003eoff) at eval thresholds, rebuilding the cached evaluator and re-evaluating the population at each transition. Gradual grain ramp avoids a single fitness cliff (graduated non-convexity).\n\nKEY IDEA (Bruno): at the sharing-\u003eno-sharing transition, do not rely on place_missing/divide to rediscover the missing rooms. Instead PROGRAMMATICALLY SUBDIVIDE each shared leaf into the correct number of distinct spaces as an explicit 'unfold' operation. This directly pays down the materialisation deficit that otherwise makes a sharing-run seed start deep in the fail hole (evolved-3M.dom was missing ~12 rooms). The unfold turns a shared leaf of code X (share=k) into k sibling leaves of code X splitting its footprint, so the de-shared genome already satisfies the per-room count before local search resumes. This is also option 3 in 3l6 (materialise shared leaves on write) but applied mid-search at the phase change.\n\nRisk to characterise: sharing re-centres size targets on k*target, so the sharing-optimal massing (fewer, larger rooms) is geometrically different from the per-leaf optimum; the transferable value may be the adjacency/topology skeleton, not the sizing. The unfold subdivision needs to produce children with sensible individual proportions/widths, not just area.\n\nBaseline for comparison (harbor-house, init.dom, 3M, niced 1-worker warm chain): direct --no-leaf-sharing reached canonical 4.19e-06, 15 fails, 0 critical, still climbing. A head-to-head warm-start-from-sharing run (seed evolved-3M.dom) is running now (evolved-warmshare.dom) to measure whether the sharing topology, once forced honest, catches the direct route.","notes":"CONCLUSIVE (warm chain complete, ~2.67M total evals): evolved-unfold.dom = 4.19e-06, 15 fails, 0 critical (breakdown: 6 level, 5 size, 1 width). This MATCHES the direct --no-leaf-sharing baseline (evolved-3M-nols-2 4.19e-06/15 fails; -nols-3 5.14e-06/15 fails).\n\nVERDICT for yaa:\n- Schedule A NAIVE (warm-start from raw sharing seed): FAILS — stalls at 8.66e-08, 70 fails; place_missing/divide cannot dig out the ~15-room count deficit.\n- Schedule A + UNFOLD (operators.unfold_shared_leaves at the transition): CATCHES the direct route (4.19e-06, 15 fails). Bruno's key idea validated: the sharing-phase adjacency/topology skeleton is transferable; the materialisation (count) deficit — not the k*target sizing mismatch — was the sole blocker. Unfold pays it down so de-share local search resumes from a competitive genome.\n\nREMAINING: Schedule B (in-run annealing: ramp leaf_share_factor 4-\u003e3-\u003e2-\u003eoff mid-run, rebuilding+re-evaluating the population and unfolding at each grain step) is the still-unimplemented driver change. Now well-motivated: the unfold primitive it needs is built and proven. Recommend spinning Schedule B into its own implementation issue and closing yaa as the investigation it was scoped as. Circulation-routing refinement to unfold tracked in 8iv.","status":"closed","priority":3,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-05T16:33:35Z","created_by":"Bruno Postle","updated_at":"2026-07-15T13:37:07Z","started_at":"2026-07-12T14:52:21Z","closed_at":"2026-07-15T13:37:07Z","close_reason":"Closed","dependencies":[{"issue_id":"homemaker-py-yaa","depends_on_id":"homemaker-py-3l6","type":"blocks","created_at":"2026-07-05T17:33:56Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":2,"comment_count":0} {"id":"homemaker-py-b3v","title":"9o5 veto hatch: interchange:false to suppress harmful auto-derived classes (harbor-house 8-code chain)","description":"Deferred escape hatch from the 9o5 spec (§2 'escape hatch' / §7.5 open item), now JUSTIFIED by the xi7 validation run. Auto-derivation chains harbor-house into a transitive 8-code class {da1,ef1,k1,la1,m,me1,n,ws1} spanning a 6x size range (Meeting 10 m2 .. Dining/Neighbourhood 60 m2) — semantically nonsensical (Meeting\u003c-\u003eDining\u003c-\u003eKitchen\u003c-\u003eMechanical are not interchangeable). xi7 A/B (3 seeds, budget 2500) shows superpose ON HURTS: OFF wins 2/3 on collapsed score, and ON ADDS fails in both losses (38v33, 48v43), i.e. the collapse re-typing perturbs feasibility. ACTION: add a per-space 'interchange: false' opt-out in patterns.config that removes a code from class derivation (programme.interchangeable / derive_interchange_classes honour the flag). Lets the architect veto a misgroup without disabling superposition globally. NOTE: superpose default stays OFF regardless (xi7 verdict null/negative overall), so this only matters if/when superpose is used on real configs. Lower priority.","status":"open","priority":3,"issue_type":"task","owner":"bruno@postle.net","created_at":"2026-06-30T07:23:03Z","created_by":"Bruno Postle","updated_at":"2026-06-30T07:23:03Z","dependencies":[{"issue_id":"homemaker-py-b3v","depends_on_id":"homemaker-py-xi7","type":"related","created_at":"2026-06-30T08:24:17Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} @@ -76,25 +76,25 @@ {"id":"homemaker-py-erc.6","title":"Experiment: inner-loop slack-expansion objective term","description":"Inner-loop counterpart to plot-fill construction. If Diagnostic B shows the inner loop has room to expand leaves into slack but no objective gradient to do so (the scalar rewards hitting target area but not exceeding it where slack exists), add a term/incentive so the ratio optimiser pushes leaf boundaries out to consume neighbouring slack and satisfy size, rather than parking at target.\n\nCONDITIONAL on Diagnostic B: build this only if B localizes the gap to the inner loop (room to expand, no gradient); if B shows construction targets too-small dims, prefer the plot-fill construction sibling. Must preserve the §5.4 inner-loop cliff / §4.9 lexicographic protection — the term sits where it cannot displace the fail-count ordering. A/B vs §12.2 baseline, seeds 0/1/2, 20000 evals, staged, default-OFF. Record DESIGN.md §13.6.","notes":"DEPRIORITISED by Diagnostic B (§13.2). B shows the inner loop CANNOT repair undersize: the slack is depth-driven maldistribution baked into the frozen topology, and the equal-offset ratio DOF cannot shrink a 14x leaf to feed a starved one without trading into shape fails (0.5^n cliff). Wrong DOF and wrong direction — the blocker is slicing POSITION, not a missing expansion reward. Fix belongs upstream in construction/topology (erc.4 re-scoped, erc.3). Keep as a low-priority follow-up only if a depth-balanced construction still leaves a residual size gradient the inner loop could pick up.","status":"closed","priority":4,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-06-22T23:16:24Z","created_by":"Bruno Postle","updated_at":"2026-06-28T13:22:22Z","closed_at":"2026-06-28T13:22:22Z","close_reason":"wont-fix (DESIGN §13.7): Diag B (§13.2) showed the inner loop cannot repair undersize (wrong DOF — slicing position, frozen-topology ratios). Superseded by depth-balanced construction (erc.4). Condition unmet.","dependencies":[{"issue_id":"homemaker-py-erc.6","depends_on_id":"homemaker-py-erc","type":"parent-child","created_at":"2026-06-23T00:16:23Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-erc.6","depends_on_id":"homemaker-py-erc.2","type":"blocks","created_at":"2026-06-23T00:16:47Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":0,"comment_count":0} {"id":"homemaker-py-erc.5","title":"Experiment: compactness-aware cuts (minimize leaf perimeter/area)","description":"Attacks the #1 factor, crinkliness (346) — a per-leaf perimeter/area property DISTINCT from proportion (aspect ratio). Proportion-aware seeding (leu.2) sizes splits but does not bias toward balanced, square-ish subdivision. Add a KD-tree-style 'keep both children compact' cut rule (prefer the cut orientation/position that minimises summed child perimeter/area) in construction.\n\nCONDITIONAL on Diagnostic A: if A shows per-leaf shape-fail is FLAT across densities (floor intrinsic to slicing density), better cuts at the same leaf count will not pay → this should be closed wont-fix in favour of leaf-sharing. Only build if A shows shape-fail RISES with density. A/B vs §12.2 baseline, seeds 0/1/2, 20000 evals, staged, default-OFF. Record DESIGN.md §13.5.","notes":"DEPRIORITISED by erc.1 verdict (§13.1): per-leaf shape-fail flat vs slicing density and cuts already squarest (_size_divisions_from_targets picks squarest rotation) yet still ~1.8 fails/leaf =\u003e little compactness headroom at fixed leaf count. Floor is intrinsic to leaf COUNT, not cut quality. Revisit only if leaf-sharing (erc.3) underdelivers.","status":"closed","priority":4,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-06-22T23:16:21Z","created_by":"Bruno Postle","updated_at":"2026-06-28T13:22:17Z","closed_at":"2026-06-28T13:22:17Z","close_reason":"wont-fix (DESIGN §13.7): Diag A (§13.1) showed the floor is intrinsic to leaf COUNT not cut quality; revisit condition was 'only if leaf-sharing underdelivers' but leaf-sharing OVER-delivered (−32…−39%, §13.3). Condition unmet.","dependencies":[{"issue_id":"homemaker-py-erc.5","depends_on_id":"homemaker-py-erc","type":"parent-child","created_at":"2026-06-23T00:16:21Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-erc.5","depends_on_id":"homemaker-py-erc.1","type":"blocks","created_at":"2026-06-23T00:16:43Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":0,"comment_count":0} {"id":"homemaker-py-2g5","title":"Rebuild occlusion/daylight/sun subsystem in Python (post-Phase-5, after optimisation fully native)","description":"DESIGN.md §6 port scope — a whole subsystem, not a term. quality_daylight (Leaf.pm:281-296) needs Urb::Misc::Sun + Urb::Field::Occlusion (+CIESky); quality_uncrinkliness also takes the occlusion object. Indoor spaces return 1 for daylight; cost is outdoor spaces + crinkliness. Port Sun_horizontal (262980-minute normalisation) and the occlusion wall set from Dom-\u003eWalls.","acceptance_criteria":"Daylight and crinkliness factors match Perl (float tolerance) across the corpus, including multi-storey cases","notes":"Re-scoped 2026-06-12: occlusion disabled in the Urb oracle instead of ported (see homemaker-py-gp2). Native fitness ships with simple crinkliness (illumination factor = 1, in homemaker-py-gnw). This issue is now the eventual Python occlusion rebuild, only after optimisation works entirely in Python. Restores outdoor-daylight and shaded-wall selection pressure.\nReframed 2026-06-17: orthogonal to epic homemaker-py-c4c. This is fitness FIDELITY (restoring daylight + shaded-wall selection pressure to match Perl), not search CAPABILITY — it changes what 'good' means, not the search's ability to find good. It will NOT improve final designs in the sense currently sought. Stays P4, deferred until the topology-search-quality epic lands and optimisation is fully native.","status":"open","priority":4,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-06-11T23:38:25Z","created_by":"Bruno Postle","updated_at":"2026-06-17T19:14:48Z","dependency_count":0,"dependent_count":0,"comment_count":0} -{"_type":"memory","key":"deceptive-valleys-in-topology-search-when-every-single","value":"Deceptive valleys in topology search: when every single-step mutation from a target state passes through a high-fail intermediary (e.g. level_fix displaces a room into 5+ new fails), a compound operator that atomically applies two coordinated changes can escape. Design compound operators to land on the low-fail state directly, bypassing the deceptive gradient. Programme-house example: level_compound_fix atomically moves the level-constrained room AND re-inserts the displaced room adjacent to C in one step (operators.py, 2026-06-14)."} -{"_type":"memory","key":"homemaker-py-3l6-fix-leaf-sharing-evolve-runs","value":"homemaker-py-3l6 fix: leaf-sharing evolve runs now auto-finish before write via driver.polish_finish — unfold_shared_leaves() then a warm-started leaf_sharing=False polish search (--polish-budget, default budget//2). Makes the written .dom honest under canonical homemaker-fitness (internal==canonical when leaf_sharing off). Interrupt path forces polish_budget=0 (unfold+rescore only). This is yaa's unfold-then-polish, made automatic; Schedule B annealing is still kpu."} -{"_type":"memory","key":"multi-storey-staircase-consistency-when-dividing-or-retyping","value":"Multi-storey staircase consistency: when dividing or retyping a circulation (C) leaf at one level, the same structural change should be propagated to the matching leaf on ALL other storeys so the stair core path is maintained. The optimizer cannot fix staircase disruptions through trial-and-error geometry alone — it requires a synchronized multi-level operator that applies the same topology change to every storey simultaneously."} -{"_type":"memory","key":"unfold-strategy-for-shared-leaves-homemaker-py-8iv","value":"Unfold strategy for shared leaves (homemaker-py-8iv, resolved 2026-07-16): use the BALANCED GRID (operators._grow_balanced/_size_subtree_equal), NOT circulation-aware slicing. Slicing a shared leaf perpendicular to its access edge so every child touches the corridor was implemented + A/B-tested and LOST decisively (150k-eval warm-start polish from evolved-3M: slice 41 fails/3.5e-14 vs grid 25 fails/2.4e-09, grid ahead at every milestone). Reason: k rooms all touching one wall are intrinsically thin slices; that geometric debt (proportion/long/width) is unfixable without topology change, while the grid's squarer children let local search re-route access cheaply via level_retype/place_missing/level_fix. Lesson: at the sharing-\u003eno-sharing transition, prioritise squarer children and leave access to local search; do not reintroduce slicing in Schedule B (kpu)."} -{"_type":"memory","key":"9o5-multi-use-leaves-is-path-a-superposition","value":"9o5 multi-use leaves is path (a) — superposition as SEARCH RELAXATION that COLLAPSES to specific usage at the end, NOT path (b) loose-fit/no-collapse. Bruno's intent: codes with SIMILAR leaf requirements form an interchangeable equivalence class; during evolution the solver doesn't commit which leaf serves which specific usage (smoother landscape, no fighting over exact leaf usage); at the end the layout is CONDENSED to specific usages by brute-forcing the in-class assignment (3 interchangeable usages over 3 leaves = 3! = 6 combinations to check, pick best). 'Derive automatically' compatibility = requirement-similarity grouping. This reverses the issue's stated 'path b preferred' note."} {"_type":"memory","key":"homemaker-py-pythonpath-set-pythonpath-home-bruno-src","value":"homemaker-layout PYTHONPATH: package installed as 'homemaker-layout' via pip install -e . so 'import homemaker_layout' works from anywhere without PYTHONPATH. For running tests use 'python -m pytest' from project root /home/bruno/src/homemaker-layout (pyproject.toml adds src/ automatically). Never try pip show homemaker — that's the old homemaker-addon conflict."} +{"_type":"memory","key":"multi-storey-staircase-consistency-when-dividing-or-retyping","value":"Multi-storey staircase consistency: when dividing or retyping a circulation (C) leaf at one level, the same structural change should be propagated to the matching leaf on ALL other storeys so the stair core path is maintained. The optimizer cannot fix staircase disruptions through trial-and-error geometry alone — it requires a synchronized multi-level operator that applies the same topology change to every storey simultaneously."} +{"_type":"memory","key":"programme-house-optimisation-result-2026-06-14-15","value":"Programme-house optimisation result (2026-06-14/15): best achievable is 1 fail (l1 wrong level, score ~0.005). 0 fails is geometrically impossible: l1 (min 27m²) must occupy ll (~23m²) at level 0, which eliminates the t3-adj-C provider; dividing ll into lll(l1)+llr(C) gives llr proportion ~6:1 (fails). Python memetic optimizer achieves 1 fail in 50k evals vs Perl optimiser's 2-3 fails. Winning topology: TWO C nodes at level 0 — ll(C) for t3-adj-C via geometric contact, rl(C) for staircase via tree-sibling adjacency to rrr(O). Best .dom: scratch/from-warmstart-fixed.dom and scratch/from-compound3-fixed.dom."} +{"_type":"memory","key":"proportion-aware-constructive-seeding-leu-2-12-2","value":"Proportion-aware constructive seeding (leu.2/§12.2): sizing seed cuts from target AREAS only regresses (thin slivers wreck aspect); you must ALSO pick each cut's rotation for child squareness. It is a convergence ACCELERATOR via a deeper local optimum around the constructed topology: wins where that topology is roughly right and budget is scarce (harbor -13%, maple -10% at 20k evals) but DELAYS small programmes where the seed must be restructured by undivide (programme-house regresses at fixed budget, yet reaches the floor given budget - speed, not asymptote). Default-on. Also: n_storeys must honour storey_minimum, not just level: keys (programme-house storey_minimum:2, all rooms level:0 - was seeded 1 storey short; cq1)."} +{"_type":"memory","key":"experiment-harness-gotcha-the-leaf-sharing-relaxed-objective","value":"Experiment harness gotcha: the leaf-sharing RELAXED objective (§13.3) is injected ONLY by monkeypatching fitness.load_config in the parent process (run_staged_search.py / probe scripts). This is parent-process-only and does NOT propagate into ProcessPoolExecutor workers (n_workers\u003e1), which re-import fitness fresh and score under the STRICT on-disk patterns.config -\u003e r.n_fails MISMATCH (worker strict vs parent relaxed re-score). ALL §13.x floor runs were therefore SERIAL. Any future PARALLEL leaf-sharing experiment will silently mis-score until leaf_sharing lives on disk/CLI (tracked: homemaker-py-x3b). The parallel driver itself is correct; both paths score via load_config(programme_dir)."} +{"_type":"memory","key":"experiment-seeding-pitfall-run-search-scaled-py-s","value":"Experiment seeding pitfall: run_search_scaled.py's default PH_SEED (c964…dom) is a FINISHED programme-house design — passing it warm-starts and floors at ~3 fails, NOT a blank-slate topology search. For blank-slate runs comparable to §11.5/§11.6 baselines, seed from examples/programme-house/init.dom (a bare undivided plot; driver bootstrap auto-triggers only on bare plots). Bit the 6zy sweep — first pass used c964 and falsely showed 3-fail floor across the whole grid."} +{"_type":"memory","key":"warm-x0-initialization-bug-pattern-when-a-topology","value":"warm_x0 initialization bug pattern: when a topology operator explicitly sets division ratios on a newly-created node (e.g. compound_fix sets node.division=[0.25,0.25] for t3), parent.ratios has no entry for that node (it was a leaf). warm_x0 defaults it to 0.5, corrupting the inner loop's starting point and making the operator invisible to lex comparison. Fix: only propagate child ratios for nodes where the parent node was NOT already divided; stale hidden nodes revealed by structural mutations (swap flipping b.below) must NOT contribute their pre-writeback values. See driver.py lines 259-267 (fixed 2026-06-14)."} +{"_type":"memory","key":"unfold-strategy-for-shared-leaves-homemaker-py-8iv","value":"Unfold strategy for shared leaves (homemaker-py-8iv, resolved 2026-07-16): use the BALANCED GRID (operators._grow_balanced/_size_subtree_equal), NOT circulation-aware slicing. Slicing a shared leaf perpendicular to its access edge so every child touches the corridor was implemented + A/B-tested and LOST decisively (150k-eval warm-start polish from evolved-3M: slice 41 fails/3.5e-14 vs grid 25 fails/2.4e-09, grid ahead at every milestone). Reason: k rooms all touching one wall are intrinsically thin slices; that geometric debt (proportion/long/width) is unfixable without topology change, while the grid's squarer children let local search re-route access cheaply via level_retype/place_missing/level_fix. Lesson: at the sharing-\u003eno-sharing transition, prioritise squarer children and leave access to local search; do not reintroduce slicing in Schedule B (kpu)."} +{"_type":"memory","key":"ld2-13-6-interior-o-seed-diagnostic-all","value":"ld2/§13.6 interior-O seed diagnostic: ALL crinkliness fails in the constructed bal+share seed are UNDER-exposed (crink\u003c0.62, landlocked rooms with no facade + no uncovered-O neighbour) — zero over-exposed sliver fails. So the erc crinkliness residual is genuine under-daylighting, validating the interior light-well premise. Default outside_divisor=6 was too sparse (null: harbor 147-\u003e142, crinkliness even rose). odiv=3 is the seed-optimal joint setting: harbor seed fails 147-\u003e129 (-18), maple 219-\u003e206 (-14), landlocked fails drop, at cost of more leaves (harbor +4, maple +8). Because it ADDS leaves it carries the §13.4 wash-out risk; A/B to convergence pending."} +{"_type":"memory","key":"never-use-corpus-filenames-candidate-001-dom-candidate","value":"Never use corpus filenames (candidate-001.dom, candidate-002.dom, generated.dom, init.dom, etc.) as --output targets when running experiments. These are test fixtures. Always write experimental outputs to scratch/ or a timestamped path. Lesson from 2026-06-14: warm-start runs overwrote candidate-001/002.dom and broke graph tests."} +{"_type":"memory","key":"user-preference-bruno-this-is-a-fedora-system","value":"User preference (Bruno): this is a Fedora system — NEVER install Python packages via pip without asking first; always ask whether to install the rpm via dnf (e.g. python3-cma) before considering pip. Applies to any dependency additions."} {"_type":"memory","key":"correction-to-urb-fitness-bug-memory-bruno-2026","value":"CORRECTION to urb-fitness-bug memory (Bruno, 2026-06-12): 'C' is NOT a 'covered' type — Is_Covered is a geometric predicate (indoor space above). Urb's generic types are canonically UPPERCASE: C=circulation, O=outside, S=sahn (get_space_types qw/C O S/; corpus is 100% uppercase, never 'c'/'o' leaves). The mixed-case designs that fired the latent ratio_type first-match bug were created by homemaker's own operator type pool emitting lowercase 'c'/'o' — fixed: driver/operators now emit uppercase generics only, and class checks use t[0].lower() in 'cos'. The Urb class-sum patch stays as defensive hardening (zero impact on canonical designs). Native port (3y7/gnw): treat type classes case-insensitively, generics canonically uppercase."} +{"_type":"memory","key":"deceptive-valleys-in-topology-search-when-every-single","value":"Deceptive valleys in topology search: when every single-step mutation from a target state passes through a high-fail intermediary (e.g. level_fix displaces a room into 5+ new fails), a compound operator that atomically applies two coordinated changes can escape. Design compound operators to land on the low-fail state directly, bypassing the deceptive gradient. Programme-house example: level_compound_fix atomically moves the level-constrained room AND re-inserts the displaced room adjacent to C in one step (operators.py, 2026-06-14)."} +{"_type":"memory","key":"island-model-psk-14-is-a-null-priming","value":"Island model (psk, §14) is a NULL: priming a population from N converged independent elites + crossover-heavy migration does not beat best-of-N at equal total budget (maple island 124 vs control 116). The child_probe instrument shows WHY: area-matched crossover across independently-converged elites almost never synthesizes (1-3 of ~64 children beat the better parent, max drop 2-5) because the slicing encoding is non-canonical (9gp), so splices are disruptive not combinatorial. Search-machinery null #3 after graded-objective and niching/restarts; residual stays geometry/shape-bound."} +{"_type":"memory","key":"9o5-multi-use-leaves-is-path-a-superposition","value":"9o5 multi-use leaves is path (a) — superposition as SEARCH RELAXATION that COLLAPSES to specific usage at the end, NOT path (b) loose-fit/no-collapse. Bruno's intent: codes with SIMILAR leaf requirements form an interchangeable equivalence class; during evolution the solver doesn't commit which leaf serves which specific usage (smoother landscape, no fighting over exact leaf usage); at the end the layout is CONDENSED to specific usages by brute-forcing the in-class assignment (3 interchangeable usages over 3 leaves = 3! = 6 combinations to check, pick best). 'Derive automatically' compatibility = requirement-similarity grouping. This reverses the issue's stated 'path b preferred' note."} {"_type":"memory","key":"run-to-run-reproducibility-in-homemaker-layout-serial","value":"Run-to-run reproducibility in homemaker-layout: serial search (workers=1) is byte-for-byte deterministic; parallel (workers\u003e1) is now deterministic too AFTER fixing driver._run_batch to admit futures in submission order (was as_completed/completion order, bug xcy). Reproducibility holds only for a FIXED worker count — serial vs parallel differ because children-per-iteration is 1 vs n_workers (different batch granularity), which is expected, not a bug. The constructive seeder was NEVER nondeterministic: _assign_adjacency_aware has unique idx tiebreaks; comparing topologies with Python builtin hash() of the signature STRING is invalid (PYTHONHASHSEED salts str hashing per process) — use a stable hash (sha1) or genome.signature equality."} {"_type":"memory","key":"strategy-decision-2026-06-12-bruno-occlusion-daylight","value":"Strategy decision 2026-06-12 (Bruno): occlusion/daylight is ORTHOGONAL to building a scalable optimiser. Disable it in Urb (env flag, homemaker-py-gp2) rather than port it; native fitness uses simple crinkliness (illumination factor = 1); rebuild occlusion in Python only after optimisation is fully native (homemaker-py-2g5, now P4). Consequence: all scores change when the flag flips — re-baseline corpus/.score, DESIGN \\$4.5 gains, gate bars at one clean boundary AFTER homemaker-py-1p0 closes; Phase-2 urb-evolve benchmark must run with the same flag."} -{"_type":"memory","key":"never-use-corpus-filenames-candidate-001-dom-candidate","value":"Never use corpus filenames (candidate-001.dom, candidate-002.dom, generated.dom, init.dom, etc.) as --output targets when running experiments. These are test fixtures. Always write experimental outputs to scratch/ or a timestamped path. Lesson from 2026-06-14: warm-start runs overwrote candidate-001/002.dom and broke graph tests."} -{"_type":"memory","key":"urb-oracle-nondeterminism-urb-fitness-pl-output-varies","value":"Urb oracle nondeterminism: urb-fitness.pl output varies run-to-run from Perl hash-order randomisation — .fails line ORDER shuffles (compare sorted, use oracle.Score.fail_lines) and the score float can flip by ~1 ULP (compare with math.isclose rel_tol=1e-12, never ==). Not a batching artifact; affects single runs too. Matters for the Phase 3 native-fitness parity gate (homemaker-py-uxz)."} -{"_type":"memory","key":"experiment-harness-gotcha-the-leaf-sharing-relaxed-objective","value":"Experiment harness gotcha: the leaf-sharing RELAXED objective (§13.3) is injected ONLY by monkeypatching fitness.load_config in the parent process (run_staged_search.py / probe scripts). This is parent-process-only and does NOT propagate into ProcessPoolExecutor workers (n_workers\u003e1), which re-import fitness fresh and score under the STRICT on-disk patterns.config -\u003e r.n_fails MISMATCH (worker strict vs parent relaxed re-score). ALL §13.x floor runs were therefore SERIAL. Any future PARALLEL leaf-sharing experiment will silently mis-score until leaf_sharing lives on disk/CLI (tracked: homemaker-py-x3b). The parallel driver itself is correct; both paths score via load_config(programme_dir)."} -{"_type":"memory","key":"programme-house-optimisation-result-2026-06-14-15","value":"Programme-house optimisation result (2026-06-14/15): best achievable is 1 fail (l1 wrong level, score ~0.005). 0 fails is geometrically impossible: l1 (min 27m²) must occupy ll (~23m²) at level 0, which eliminates the t3-adj-C provider; dividing ll into lll(l1)+llr(C) gives llr proportion ~6:1 (fails). Python memetic optimizer achieves 1 fail in 50k evals vs Perl optimiser's 2-3 fails. Winning topology: TWO C nodes at level 0 — ll(C) for t3-adj-C via geometric contact, rl(C) for staircase via tree-sibling adjacency to rrr(O). Best .dom: scratch/from-warmstart-fixed.dom and scratch/from-compound3-fixed.dom."} -{"_type":"memory","key":"adjacency-in-binary-slicing-tree-is-structural-not","value":"Adjacency in binary slicing tree is structural, not geometric: the inner-loop NM cannot fix topological adjacency failures. Two paths exist: (1) tree-sibling adjacency — a node is adjacent to its sibling in the tree; (2) cross-zone geometric adjacency — leaves from different subtrees that happen to share a boundary. Staircase/adjacency fails require a topology mutation that changes which nodes are siblings or which zones touch. This was proved empirically on programme-house: staircase fail from rot=0 layout could not be fixed by NM but was fixed by level_retype creating a two-C topology (2026-06-14/15)."} -{"_type":"memory","key":"experiment-seeding-pitfall-run-search-scaled-py-s","value":"Experiment seeding pitfall: run_search_scaled.py's default PH_SEED (c964…dom) is a FINISHED programme-house design — passing it warm-starts and floors at ~3 fails, NOT a blank-slate topology search. For blank-slate runs comparable to §11.5/§11.6 baselines, seed from examples/programme-house/init.dom (a bare undivided plot; driver bootstrap auto-triggers only on bare plots). Bit the 6zy sweep — first pass used c964 and falsely showed 3-fail floor across the whole grid."} -{"_type":"memory","key":"island-model-psk-14-is-a-null-priming","value":"Island model (psk, §14) is a NULL: priming a population from N converged independent elites + crossover-heavy migration does not beat best-of-N at equal total budget (maple island 124 vs control 116). The child_probe instrument shows WHY: area-matched crossover across independently-converged elites almost never synthesizes (1-3 of ~64 children beat the better parent, max drop 2-5) because the slicing encoding is non-canonical (9gp), so splices are disruptive not combinatorial. Search-machinery null #3 after graded-objective and niching/restarts; residual stays geometry/shape-bound."} -{"_type":"memory","key":"proportion-aware-constructive-seeding-leu-2-12-2","value":"Proportion-aware constructive seeding (leu.2/§12.2): sizing seed cuts from target AREAS only regresses (thin slivers wreck aspect); you must ALSO pick each cut's rotation for child squareness. It is a convergence ACCELERATOR via a deeper local optimum around the constructed topology: wins where that topology is roughly right and budget is scarce (harbor -13%, maple -10% at 20k evals) but DELAYS small programmes where the seed must be restructured by undivide (programme-house regresses at fixed budget, yet reaches the floor given budget - speed, not asymptote). Default-on. Also: n_storeys must honour storey_minimum, not just level: keys (programme-house storey_minimum:2, all rooms level:0 - was seeded 1 storey short; cq1)."} {"_type":"memory","key":"cli-tool-style-prefer-python-m-homemaker-module","value":"CLI tool style: prefer python -m homemaker.module --parameters pattern, installable via pip install -e . with pyproject.toml entry_points. Not standalone bin/ scripts."} -{"_type":"memory","key":"ld2-13-6-interior-o-seed-diagnostic-all","value":"ld2/§13.6 interior-O seed diagnostic: ALL crinkliness fails in the constructed bal+share seed are UNDER-exposed (crink\u003c0.62, landlocked rooms with no facade + no uncovered-O neighbour) — zero over-exposed sliver fails. So the erc crinkliness residual is genuine under-daylighting, validating the interior light-well premise. Default outside_divisor=6 was too sparse (null: harbor 147-\u003e142, crinkliness even rose). odiv=3 is the seed-optimal joint setting: harbor seed fails 147-\u003e129 (-18), maple 219-\u003e206 (-14), landlocked fails drop, at cost of more leaves (harbor +4, maple +8). Because it ADDS leaves it carries the §13.4 wash-out risk; A/B to convergence pending."} +{"_type":"memory","key":"homemaker-py-3l6-fix-leaf-sharing-evolve-runs","value":"homemaker-py-3l6 fix: leaf-sharing evolve runs now auto-finish before write via driver.polish_finish — unfold_shared_leaves() then a warm-started leaf_sharing=False polish search (--polish-budget, default budget//2). Makes the written .dom honest under canonical homemaker-fitness (internal==canonical when leaf_sharing off). Interrupt path forces polish_budget=0 (unfold+rescore only). This is yaa's unfold-then-polish, made automatic; Schedule B annealing is still kpu."} {"_type":"memory","key":"urb-fitness-bug-found-fixed-2026-06-12","value":"Urb fitness bug found+fixed 2026-06-12 (patch in /home/bruno/src/urb, uncommitted): ProgrammeDriven.pm ratio_o/ratio_type grepped case-insensitively over the ratios hash and took the FIRST key — nondeterministic (x4.5 score swings) for designs with mixed-case type classes (both 'c' circulation and 'C' covered). Fixed to SUM the class (matches Is_Circulation//Is_Outside semantics); 35/35 corpus scores unchanged. CRITICAL for homemaker-py-3y7/gnw: the native port must implement class-SUM ratios. Building.pm has the same unpatched pattern (site-driven path, not used by our oracle). Also: the memetic search reward-hacked this bug before the fix — search results predating it are noise artifacts."} -{"_type":"memory","key":"user-preference-bruno-this-is-a-fedora-system","value":"User preference (Bruno): this is a Fedora system — NEVER install Python packages via pip without asking first; always ask whether to install the rpm via dnf (e.g. python3-cma) before considering pip. Applies to any dependency additions."} -{"_type":"memory","key":"warm-x0-initialization-bug-pattern-when-a-topology","value":"warm_x0 initialization bug pattern: when a topology operator explicitly sets division ratios on a newly-created node (e.g. compound_fix sets node.division=[0.25,0.25] for t3), parent.ratios has no entry for that node (it was a leaf). warm_x0 defaults it to 0.5, corrupting the inner loop's starting point and making the operator invisible to lex comparison. Fix: only propagate child ratios for nodes where the parent node was NOT already divided; stale hidden nodes revealed by structural mutations (swap flipping b.below) must NOT contribute their pre-writeback values. See driver.py lines 259-267 (fixed 2026-06-14)."} +{"_type":"memory","key":"adjacency-in-binary-slicing-tree-is-structural-not","value":"Adjacency in binary slicing tree is structural, not geometric: the inner-loop NM cannot fix topological adjacency failures. Two paths exist: (1) tree-sibling adjacency — a node is adjacent to its sibling in the tree; (2) cross-zone geometric adjacency — leaves from different subtrees that happen to share a boundary. Staircase/adjacency fails require a topology mutation that changes which nodes are siblings or which zones touch. This was proved empirically on programme-house: staircase fail from rot=0 layout could not be fixed by NM but was fixed by level_retype creating a two-C topology (2026-06-14/15)."} +{"_type":"memory","key":"urb-oracle-nondeterminism-urb-fitness-pl-output-varies","value":"Urb oracle nondeterminism: urb-fitness.pl output varies run-to-run from Perl hash-order randomisation — .fails line ORDER shuffles (compare sorted, use oracle.Score.fail_lines) and the score float can flip by ~1 ULP (compare with math.isclose rel_tol=1e-12, never ==). Not a batching artifact; affects single runs too. Matters for the Phase 3 native-fitness parity gate (homemaker-py-uxz)."} diff --git a/DESIGN.md b/DESIGN.md index a04a4d1..67ca32e 100644 --- a/DESIGN.md +++ b/DESIGN.md @@ -2253,7 +2253,7 @@ honesty — the point of the bug — is restored. Tests: `driver.polish_finish` (unfold+rescore stitching, polish-search accounting, no-best noop); 254 pass. -## 16. In-run leaf-share grain annealing — Schedule B (`homemaker-py-kpu`) — IN PROGRESS +## 16. In-run leaf-share grain annealing — Schedule B (`homemaker-py-kpu`) — DONE (negative) **Premise.** §15's finish crosses the sharing→off objective cliff in a *single* hard transition (unfold every shared leaf at once, then polish). Schedule B (yaa's @@ -2309,7 +2309,33 @@ is meaningless here; the head-to-head below runs at the baselines' budget.) Test seeding; `search_annealed` phase stitching / honest finish / degenerate-ladder fallback; 258 pass. -**Head-to-head (RUNNING).** harbor-house, `init.dom`, seed 0, pop 16, child 80, grain -`4,3,2`, budget 1.5M (500k/phase) + polish 1.5M = **3M total**, matched to the yaa -baselines. Targets: (a) direct `--no-leaf-sharing` 5.14e-06; (b) manual unfold -warm-chain 4.19e-06. Verdict pending run completion. +**Head-to-head (DONE — NEGATIVE).** harbor-house, `init.dom`, seed 0, pop 16, child +80, grain `4,3,2`, budget 1.5M (500k/phase) + polish 1.5M = **3M total** (workers 4, +~22h), matched to the yaa baselines. Result: **1.26e-08, 23 fails** (canonical +`homemaker-fitness` byte-for-byte). Both targets **beat it decisively**: + +| route | fitness | fails | +|-------|---------|-------| +| direct `--no-leaf-sharing` | 5.14e-06 | 15 | +| yaa warm-chain (single hard unfold) | 4.19e-06 | 15 | +| **Schedule B (graduated 4→3→2→off)** | **1.26e-08** | **23** | + +~400× worse fitness, +8 fails. **Verdict: the graduated grain ramp loses to the +single hard sharing→off transition (§15).** The trajectory shows why — each grain +step spikes the fail count as its unfolded leaves acquire independent shape fails +(phase-end fails 19 → 21 → 27, then the final de-share unfold 27 → 36), and the +per-phase budget re-polishes a partially-materialised state that the *next* step +materialises further, so the coarse-grain gains (19 fails at grain 4) do not carry +forward. Splitting the budget across three intermediate materialisations left the +polish phase starting from a deeper hole (36 fails) than the warm chain's single +clean transition, and 1.5M polish evals recovered only to 23 — short of the 15 both +baselines reach. Graduated non-convexity is **falsified** for this materialisation +cliff: the transferable value is the sharing-phase topology skeleton (yaa), and it is +best cashed in **once**, at full grain, not annealed. (Caveat: this run used +`workers=4` vs the baselines' `workers=1`; the ~400×/+8-fail gap is far larger than +worker-count trajectory noise, so the direction is robust.) + +The machinery is retained (`search_annealed`, `--anneal-grain`, `unfold_shared_leaves( +above=)`, `search(seed_pop=)`, the `max_share` evaluator override) — it is correct, +tested, and honest, and the `seed_pop` / grain-cap primitives are reusable — but the +default finish stays §15's single-transition unfold+polish. Tests: 258 pass.