Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 21 additions & 0 deletions research/p104-odd-unit-per-grading/NOTICE
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
Apache-2.0. DreamingOfClouds with Claude Opus 5.5 (Anthropic), via Claude Code. The outside-session lanes
DECZERO, COURIER, YYMIX, XBIRTH and WEBS are DreamingOfClouds'. Parts of the analysis (the Clone lemma,
the b-family, M45 as the B-port contract) were Fable-assisted and reproduced independently by our lanes.

data/base-7546-word.json.gz is the #7546 word by ToikY (bytes of expected/role902-WORD.json.gz at #7546 head
99438c43df7aac1e5dbcc43c1578fcd886c119ec, SHA256 pinned in words.py); credit ToikY. The unpaid words in Part III
are our own edits of it, shipped as edit lists (data/edits-*.json.gz, data/window8-targets.json.gz).

data/W6805.json.gz is CubeJerry's #6805 word, used as data only by cuts.py; credit CubeJerry.
No code from that PR is run. The feeder6805 edit list (data/edits-feeder6805.json.gz) is our own (an outside-session lane, Fable-assisted), and data/mix14-bprime-R.json is our own lane data.

data/lines120.json (Gram and F2^9 labels of the 120 E8 lines, the first 104 being the P104 ports) is our derived
data, checked against the repository file research/complex-frame-pair-decoders/inputs/pr717/P104_HIT_flip_3a.txt
(PR968 decoder, pinned in SOURCES.json) through data/iso968.json, as in #7583; credit the PR968 authors.
data/half18_frames.json and data/ports_B.json are our own lane data.

Credit Rohan Arun for the #7113 signed caller and its B pair, Clear1nsight, ToikY and the other contributors of the
P104 price and replay lineage for the conventions, and SYLforge, FanPUYen and aliromman for the frontier PRs
referenced in the README.

All checkers in this package are our own code; no code is imported from the repository or any other package.
302 changes: 302 additions & 0 deletions research/p104-odd-unit-per-grading/PROOF.md

Large diffs are not rendered by default.

347 changes: 347 additions & 0 deletions research/p104-odd-unit-per-grading/README.md

Large diffs are not rendered by default.

30 changes: 30 additions & 0 deletions research/p104-odd-unit-per-grading/SOURCES.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,30 @@
{
"schema_version": 1,
"base_commit": "e54523517d0f03e58352e647c403a5af3e525a57",
"files": [
{
"path": "docs/research/agent-task-catalog.json",
"usage": "provenance",
"git_blob": "3404fbdd5e7529a0674eef26114b5071e4586cec",
"sha256": "ead6cdc9d0af6d58e8a49744d8ec8c7c15e13d2ba93c276c434bcf9b51b93a81"
},
{
"path": "research/complex-frame-pair-decoders/inputs/pr717/P104_HIT_flip_3a.txt",
"usage": "input",
"git_blob": "458ddb13dcf73ed031ffb96d9df149c994c7e4ee",
"sha256": "8e0178db02c14e1b29058cdbdbd50aac22f2b55a1ba21dc82f417610c9932526"
},
{
"path": "upstream/build/sections/05-layers.tex",
"usage": "provenance",
"git_blob": "5ce8b3c83c455f221bb84858a97230b0285644aa",
"sha256": "20cfc8d12099d4e4ed9e3e7eeb6162edd9b6e0e076f24693364cd24377252594"
},
{
"path": "upstream/build/sections/06-transforms.tex",
"usage": "provenance",
"git_blob": "de073e9292d58573f7065a3cd4f214e0d05f3444",
"sha256": "924316e4bc905066281d8f9e8746a49d8829aeb393688608deb1e61c3c87853c"
}
]
}
135 changes: 135 additions & 0 deletions research/p104-odd-unit-per-grading/TASK.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,135 @@
{
"schema_version": 1,
"task_id": "p104-side-correction",
"priority_revision": 8,
"base_commit": "e54523517d0f03e58352e647c403a5af3e525a57",
"original_card": {
"id": "p104-side-correction",
"question": "Can joint helper bases and a legal causal split improve the complete P104 productive/paid invoice?",
"starting_points": [
"docs/research/current-priorities.md",
"docs/research/research-family-status.md",
"certificates/selected-result.json",
"docs/research/pr-intake-evidence/pr5599-completion/PROOF-REVIEW.md",
"research/p104-helper-subspace-splits/README.md",
"research/dyadic-helper-channel-capacity/README.md"
],
"methods": [
"Construct and independently check one source-bound instance of the stated first experiment.",
"Seek an exact counterexample or target-specific complete-budget obstruction to that same proposed change."
],
"acceptance_gate": "An exact arbitrary-input identity, source/target frame availability and restoration at the chosen cut, plus the complete channel+correction+idle/hosting bill or an explicit favorable bound. The native channel alone is not a supplier. A scoped local result may be retained without a complete supplier. Expanding it into further construction rounds requires the family-level budget and obligation gate in research-family-status.md; an optimistic passing relaxation alone is not acceptance.",
"first_step": {
"command": [
"python3",
"-B",
"scripts/research_selected_baseline.py"
],
"expected": "PASS_RESEARCH_BASELINE_BINDINGS_ONLY",
"inspect": "Read revision 8, the selected review and live exact-source handoffs. Then follow this card's bounded first_experiment.",
"scope": "Source-bound orientation only, not a supplier replay or theorem proof."
},
"first_experiment": "Start with the selected #5599 emitted word and review pending #5718 distinct input/output bases plus #5717/#5719 coverage before claiming work. Pick one uncovered source word, helper group and cut region. Prove the arbitrary-input relation and legal entry/exit corrections, then compare full target and paid profiles. Reject if completion, owned lifetimes or consumer costs repay the gain. A static pairing rank alone is not a physical split.",
"bottleneck_link": "Judge the actual current supplier contract and complete signed target/paid bill; frozen-profile ceilings require an applicability proof.",
"allowed_variations": [
"One exact source head, helper/cut region or quantified proposition and changed hypothesis; check live handoffs first. Different seeds are not different scopes."
],
"scope_change_triggers": [
"Changing source head, helper/cut domain, operation alphabet or caller boundary without recording the new scope.",
"Treating an archived proposal or local numerical pass as complete supplier admission."
],
"overlap_terms": [
"P104",
"helper basis",
"helper split",
"5718",
"dyadic capacity"
],
"existing_checks": [
{
"scope": "PR #4133: literal-slot defect",
"evidence": "research/p104-literal-slot-defect/README.md",
"remaining": "Independently validate the stated scope or change its producer/consumer premise and pay the complete boundary.",
"status": "author-checked",
"source_commit": "88666f1dee93fbd85b31fd5d6f91a219e111b5f4",
"selected_record_sha256": "bb2eaa58647b353f88d0aec2eafb9d30df1be53589231b23e12ceef3e4bad129"
},
{
"scope": "PR #4136: uniform commutation gap",
"evidence": "research/p104-uniform-commutation-gap/README.md",
"remaining": "Independently validate the stated scope or change its producer/consumer premise and pay the complete boundary.",
"status": "author-checked",
"source_commit": "281ce601a8be9db470de630759179c23e1dfaa88",
"selected_record_sha256": "bb2eaa58647b353f88d0aec2eafb9d30df1be53589231b23e12ceef3e4bad129"
},
{
"scope": "PR #4140: uniform encoding gap",
"evidence": "research/p104-uniform-encoding-gap/README.md",
"remaining": "Independently validate the stated scope or change its producer/consumer premise and pay the complete boundary.",
"status": "author-checked",
"source_commit": "de25c6525aa866962309df5d107b32d55e6e5628",
"selected_record_sha256": "bb2eaa58647b353f88d0aec2eafb9d30df1be53589231b23e12ceef3e4bad129"
},
{
"scope": "PR #4153: source-chart shortcut exhausts its optimistic allowance",
"evidence": "research/p104-source-chart-shortcut/README.md",
"remaining": "Independently validate the stated scope or change its producer/consumer premise and pay the complete boundary.",
"status": "author-checked",
"source_commit": "9123241fea53df77e51ffe30d9aca1625d4bf3a8",
"selected_record_sha256": "bb2eaa58647b353f88d0aec2eafb9d30df1be53589231b23e12ceef3e4bad129"
},
{
"scope": "PR #4154: fiber-tag seam mismatch",
"evidence": "research/p104-fiber-tag-seam/README.md",
"remaining": "Independently validate the stated scope or change its producer/consumer premise and pay the complete boundary.",
"status": "author-checked",
"source_commit": "4ffcdaee5d6255a9cea1fe53a5424435068089d0",
"selected_record_sha256": "bb2eaa58647b353f88d0aec2eafb9d30df1be53589231b23e12ceef3e4bad129"
},
{
"scope": "Selected PR5599: complete helper-subspace construction and independent caller replay; five framework assumptions remain.",
"evidence": "docs/research/pr-intake-evidence/pr5599-completion/PROOF-REVIEW.md",
"remaining": "The next experiment must identify uncovered source-specific work. Successor archive intake does not transfer this admission.",
"status": "reviewed",
"source_commit": "09dd0965711ca7eeb169057a420abab1c9052137",
"selected_record_sha256": "2a0220e866417774606e0f991e41b9c1d26b1f43b97857238cf478dbc7d330b5"
}
]
},
"scope_status": "changed",
"actual_scope": "One day of P104 research (Nov 14) on the decoder wall and the t = -1 park, organised around one cost, the odd unit paid per grading sigma_p: (I) proved lemmas with exact checks - o(N) = rank Odd(N) for any block/ring/coupling, the Cayley form (t = -1 iff the sigma-even part is singular), per-grading ledgers, t_eff = ab/(cd-ab) and M45 as the t = -1/2 B-port contract, the no-cloning family (Rank, Lock, Clone, Theorem W, NT, AS/VA, lift law, Freeze, shared lift <= 8, Commutation, XRECALL, Zone, owned additivity, wiring, the web-general Cut Lemma), decoder structure (UP+DOWN forced, G/2 = 13 proj, -12 I composition, joint rank >= 8k), decoder-wall pieces since #7583 (Lemma A(acyc), Lemma B mod-4, HALF18 rank-15/13 certificates, early-set cap), and local re-solve laws (Theorem F, LIN, TR, 2B, Descent, RT); (II) every route closed that day as a scoped negative result with its reason and numbers; (III) the words built that did not pay, replayed exactly against #7546 with their prices.",
"scope_change_reason": "The card asks for a constructive helper-basis or causal-split improvement. This package instead records lower bounds, scoped negative results and unpaid words on P104 so that others can see which decoder, t = -1, contract, multi-call and local-edit routes are closed and why; it makes no kappa claim and ships no new frontier word.",
"bottleneck_link": "The P104 invoice's owned decoder mass (56 per plane) and the t = -1 park are the two fixed costs every proposed redesign has hit; this package shows they are one cost (the odd unit per grading) and that every construction tested pays about one owned unit per port per call, against a break-even near 0.214.",
"budget_minutes": 240,
"stop_rule": "Every proved claim has a proof in PROOF.md; every finite fact quoted as [C] is re-run by a standard-library checker with positive and negative controls; lane-only numbers are tagged [C-lane] or [C-param]; open items stated as open; validator PASS.",
"outcome": "negative-result",
"evidence": "README.md, PROOF.md, odd_unit.py, e8laws.py, walls.py, onerank.py, cuts.py, words.py, data/, VERIFY.json.",
"author": "DreamingOfClouds",
"ai_assistance": "Claude Opus 5.5 (Anthropic), via Claude Code; some analysis Fable-assisted, reproduced independently (see README credits).",
"open_questions": [
{
"question": "Can one owned ascent serve several t = -1 ports (is M >= n at t = -1)?",
"why_it_matters": "Up-set cuts allow sharing; if sharing is impossible every one-plane t = -1 design on P104 has M >= 104 (S <= 712) and the t = -1 route is closed in the exact model.",
"first_step": "Decide the F2^3, n = 4 case (labels 1, 2, 4, 7) exactly for the 13 cut-feasible climb-3 owned sets.",
"blocked_by": "Nothing."
},
{
"question": "Is there an early set with contamination rank d = 1 and 577-600 cells?",
"why_it_matters": "It is the only window the early-set cap leaves open for the Theorem-M lever at the LP-perfect U layer.",
"first_step": "Use the proportional-class structure forced at d = 1 (bicliques T x S) to bound |E| below 577.",
"blocked_by": "Nothing."
},
{
"question": "Is C >= 52 on P104 (the h = 9 wall), and do >= 6 replacement rows always cost >= 28?",
"why_it_matters": "These are the remaining decoder-side savings (at most 4 owned units per plane).",
"first_step": "Either build an acyclic decoder of cost <= 51 (Lemma A) or find a certificate that uses intermediate flags; for six rows prove the union-rigidity lemma of #7583.",
"blocked_by": "Nothing."
},
{
"question": "Can a rank >= 9 free-lift decoder network (zeroing + frame memory) cost below 56?",
"why_it_matters": "It is the only decoder freedom found that is not a relabel of the row model.",
"first_step": "Search rank-9 lift networks on E7/E6 toys with exact pricing.",
"blocked_by": "Nothing."
}
]
}
71 changes: 71 additions & 0 deletions research/p104-odd-unit-per-grading/VERIFY.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,71 @@
{
"schema_version": 1,
"status": "ready",
"python": "3.12",
"requirements": [],
"inputs": [
"research/complex-frame-pair-decoders/inputs/pr717/P104_HIT_flip_3a.txt"
],
"commands": [
{
"argv": [
"{python}",
"-I",
"-B",
"{package}/odd_unit.py"
],
"timeout_seconds": 60,
"checks": "Part I.1-I.2, exact over Q at generic rational points: I1 o(N) = rank Odd(N) on 40/40 random coupled 2-4x4 matrices in both gradings; I2 dim ker(R+I) = dim ker E on 40/40 (17 with singular E) plus the three park cases; I3 per-port det R = 1, det(R+I) = -4cd/(ab-cd), t_eff = ab/(cd-ab), t_eff = -1 iff cd = 0 on 300 points, HOST t_eff = alpha*beta; I4 coupled singular pencil (det E_a = 0, det E_0 = 1, o_a = o_b = 2); I5 isotropic kernel o_a = o_b = 0, monomial q_a q_b o = 1,1; I6 q-free coupling keeps m_a = 2; I7 M45 = U+L-I, H = LU, o = 2, R^2 = -I, M^2 = 2J, M^4 = -4I, det 2, diag(1,1/2)M45 = HOST(-1,1/2), t_eff = -1/2, o(H M45^-1) = 1; I8 dressing invariance on 200 cases, control q-odd dressing changes tr R on 165/200; I9 b-family t = -1/(1+b), gauge invariance, window-scalar pole. Prints PASS_ODD_UNIT_PER_GRADING."
},
{
"argv": [
"{python}",
"-I",
"-B",
"{package}/e8laws.py"
],
"timeout_seconds": 120,
"checks": "Exact (Bareiss over Z, stdlib). Data hashes; 120-line E8 Gram (rank 8, G^2 = 30G) and the 104-port block equal to the pinned PR968 Gram on all 104^2 entries. HANDT1: diagonal odd, 5,720 odd cells all G = 0, 4,992 even cells G = +-1, G104^2 = 26 G104 (control 30G fails), G/26 idempotent of trace 8, transposed-twin L rows frame-legal (swapped frames illegal), M_L = G/2, W^T U = 13 I, composition residue -12 on the diagonal and 0 on all 5,720 odd cells, 1,360 cells hit by a single same-index term; XRECALL q.q = 1 (control: even vector inside its perp); max pairwise-orthogonal set and non-orthogonal clique number 8 on 104 and on the 96 A ports; XDEC copies give identity blocks of rank 16 and 24; HALF18 frame certificates rank 15 on E8-120 and 13 on P104 (partitions, zero on even cells, nonzero diagonal), lower bound 8; EARLYCAP rank(2I+C_F) <= d+8 and the |E| cap table, T1CANCEL res = 0 on two E8 matchings; BYPASS rank_K{e_t : t in A} = 95; FACTORSWAP 10 decompositions (8 at 56, 2 at 64); Lemma A sums (104 for the frame Z, control with invalid Z). Prints PASS_E8_LAWS."
},
{
"argv": [
"{python}",
"-I",
"-B",
"{package}/walls.py"
],
"timeout_seconds": 300,
"checks": "Exact toys: Rank lemma (A4 real-token snapshot CLEARs drop rank 20 -> 16, final map singular 16/20; token-free owned carriers invertible; undelivered-carrier, held-copy CLEAR and public-swap controls; 250 random legal programs, 0 rank increases, dirty-CLEAR lossless-iff-in-span 0 violations); Lock/Theorem W (free wiring min D = n at n = 1, 2 in both orders, D = 0 with n owned ascents); Commutation Theorem (0/300 Lock violations with arbitrary local maps, cross-label control 47/300; chart rotation 59/59, 944/944); lift law (PLN variants top-solvable 0/216 A3, 0/360 A4 with D = 1; controls D = 0 exact); Zone lemma (two senders exact with 0 owned, one sender fails, proportional weights singular); Gram lift rank Psi = 6 on A3. Prints PASS_WALL_TOYS."
},
{
"argv": [
"{python}",
"-I",
"-B",
"{package}/onerank.py"
],
"timeout_seconds": 60,
"checks": "Exact: Theorem F (rank[A|b] - rank A in {0,1} on 3,000 random matrices, solution or left witness in each case; two-column control misses by 2); Lemma LIN (formula exact on 947/947 circuits meeting the hypothesis, fails on 528/553 that do not); Lemma TR criterion matches on 400/400; T = I network with meetings M then M^-1 rigid at 4m single-stay drops for m = 1..4 against optimum 0, all repaired by Lemma TR, redundant-stay controls repairable. Prints PASS_ONERANK."
},
{
"argv": [
"{python}",
"-I",
"-B",
"{package}/cuts.py"
],
"timeout_seconds": 240,
"checks": "Vendored W6805 bytes (CubeJerry #6805 word, sha256 pinned, data only); feeder6805 rebuilt from W6805 + edit list (pinned sha256); exact replay of both; Cut Lemma with exact ranks mod 1000000007 and 998244353 (must agree) on all 104 ports: W6805 def(G_s) = 0 on 104/104 with 0 owned crossings (24 owned moves all descend), feeder6805 def(G_s) = 1 on 104/104 with exactly one owned crossing each, lemma holds 104/104; control: feeder with crossings dropped flagged on 104/104, W6805 not flagged; alpha(P104) = 8 and alpha(MIX14 (b') R) = 6 by exhaustive branch and bound. Prints PASS_CUT_LEMMA."
},
{
"argv": [
"{python}",
"-I",
"-B",
"{package}/words.py"
],
"timeout_seconds": 400,
"checks": "Vendored #7546 word bytes (sha256 pinned); unpaid words rebuilt from #7546 + edit lists with pinned canonical sha256 (SEEDBUILD S1, BREAKER8 costly-relocation control ctrl_up, BREAKER8 Lemma RT swap words for mains 6 and 886, WINDOW8 w4_w00); exact Fraction replay and legality of each; S1 and ctrl_up rows equal #7546, 17-call invoice, kappa 0.003645014837609... (S1, x0.99907, +21.3448 D0/call) and 0.003645741005801... (ctrl_up, x0.99927, +16.8184 D0/call), both below #7546's reproduced 0.003648411294790845...; both Lemma RT words: all rows equal #7546, S, D0 and kappa equal #7546's in every digit, no-swap-back control rejected; WINDOW8 all 1517 finals equal T, X = x-y, Y = x+y on 8 ports, T differs from #7546 on 16 rows only, cost model 150.8892 + 16 x 3.1395 = 201.1210 D0/call, no kappa (merged plane); controls (corrupted ADDs, tampered edit list, tampered targets, tampered invoice) rejected. Prints PASS_UNPAID_WORDS."
}
]
}
Loading
Loading