Skip to content
Merged
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

Large diffs are not rendered by default.

68 changes: 68 additions & 0 deletions research/p104-joint-contact-native-caller-audit/PROOF.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# Proof notes for the PR5551 native, caller and funding audit

All statements concern the fixed P104 local word `research/p104-joint-contact/WORD.json.gz` (PR5551) on
canonical main `5e0052ca0aa3036a3afa8b0438927377c8d97d53`, and its complete invoice in
`research/p104-joint-contact/RESULT.json`.

## Layer 1: replay and delta

`replay()` is new code that imports nothing from the source packages. Frames are bases of subspaces of F2^9;
their spans are built as sets. For every event it checks:

- MOVE: the register is in the source frame, its ownership flag is correct, and the target span strictly contains
(public) or is strictly contained in (owned) the source span. The paid rank is the dimension gap.
- ADD, COPY, SCALE: the coefficient is nonzero and the destination (and source) registers are in the event's
frame, so the scalar operation runs in a common chart. COPY writes only into a cleared owned field.
- INIT and CLEAR: owned lifetimes, peak occupancy one.

At the end every public register is in its final frame, no owned field is live, and the 735 + 8 rational rows
equal the identity on public registers plus X added into Y = X + 112 for the 104 live ports.

For the delta, the non-MOVE programs of PR5545 and PR5551 are equal after removing frame labels. Comparing the
spans of the frame labels event by event gives exactly 6 ADD and 15 SCALE chart edits; the six ADD sites are
listed in RESULT.json and match PR5551's construct.py. The paid multisets differ by 2H1 − 3H2 + H3 − H4 + H5,
and every public initial and final span is unchanged, so the helper target profile and the productive targets
are unchanged.

## Layer 2: caller and target binding

The PR5550 checker is used as in PR5567: four literal input lines are replaced in a temporary copy (word path,
helper profile, invoice, local route count from the MOVE count). The positive control runs the patched checker
on the PR5545 word and invoice and reproduces the complete frozen PR5550 output byte for byte, so the patch
preserves the checker. On PR5551 the checker recounts the invoice from the MOVEs and native ranks, requires it to
equal PR5551's `complete_cost`, and checks all productive targets, helper normalizers and ordinary costs.
PR5550 reports `paid_native_maps_replayed: false`; layer 3 supplies that check.

## Layer 3: native maps

For a MOVE from frame A to frame B the native operator is chart(A)⁻¹ chart(B). The canonical compiler
`clifford_exact.lower_child` lowers it to a recipe D·P·X·phase·(rank C gates)·D·P·X·phase. The literal
definition is `adjoint(frame(A)) + frame(B)` from `native_verify.py`. PR5548's dense interpreter evaluates
both circuits exactly, with Gaussian-dyadic integers, on each of the 512 basis inputs. We require

1. exact equality of every column, including absolute phase;
2. squared norm one for every column;
3. exactly 2^rank nonzero entries in every column, where rank is the paid rank of that transition.

The same is done for each distinct helper endpoint pair (chart(initial)⁻¹ chart(final), credited rank
dim final − dim initial) and for the 208 intrinsic main adapters (input C-gate into chart, chart into the output
C-gate and Fourier transform), which must lower at rank zero. Checking a phase fault with an unchanged signed
tableau and a flipped child sign shows that the comparison sees what tableau checks miss.

Condition 3 binds each paid job to an actual native Gaussian of the charged rank. Together with layer 1, which
fixes which maps occur and how often, this checks the native paid-map inventory of the word, not only the
maps new relative to PR5545.

## Layer 4: funding

The moment margin at κ is Σ_targets n·r^(1−κ) − Σ_paid n·r^(1−κ). Each power is computed as exp((1−κ)·ln r)
with 80 significant digits and widened by a relative 10⁻⁷⁰, which is far above the rounding error of the
Decimal ln/exp. Bisection on the grid 10⁻¹⁴ gives the largest grid point with a certified positive margin and
the next with a certified negative margin. They equal PR5551's `raw_root_interval`. The engine shares no code
with PR5551's two moment engines or its funding peer.

## Not covered

The uniform compiler for all recursion levels, owned erasure, precision and tape costs and analytic recovery are
inherited premises. The outer conjunction and certified κ constant are PR5551's own verifier checks and are not
repeated here.
69 changes: 69 additions & 0 deletions research/p104-joint-contact-native-caller-audit/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,69 @@
# Independent native, caller and funding audit of PR5551

**Result (scoped audit, no new κ).** Every native map of the PR5551 coupled-contact word passes a dense
all-column check with absolute phase: 2,940 distinct MOVE maps, 40 distinct helper endpoint pairs and
208 intrinsic main adapters, 3,188 operators in total (3,187 distinct, since one helper pair is also a MOVE map), each compared on all 512 columns. Earlier dense
audits checked only the maps new to each word (three for PR5545 in PR5548, four for PR5551 in its own
native peer). The remaining maps were an inherited premise. On the same word, Chris R's PR5550 caller and
target-binding checker passes, an own exact replay reproduces the six ADD contact edits and the paid delta,
and an own decimal interval engine reproduces the author's raw crossing exactly.

Base: canonical main `5e0052ca0aa3036a3afa8b0438927377c8d97d53`, which already contains the PR5551 package
`research/p104-joint-contact` (author thu-nmrc). All inputs are pinned by Git blob and SHA-256 in
[SOURCES.json](SOURCES.json). This follows the PR5551 intake note: review the P104 producer/caller and the
coupled-contact delta together, recheck the native paid-map inventory, the compiler/consumer obligations,
all productive targets and complete funding, and use a serial bounded replay.

## What is checked

| Layer | Check | Result |
|---|---|---|
| 1. Own replay (no source imports) | Both words replayed from scratch: strict span nesting of every MOVE, common chart of every ADD/COPY, rational rows of all 743 endpoints, owned lifetimes | PASS on PR5551 and PR5545 |
| | Delta PR5545 → PR5551 | Scalar program identical up to frame labels; 6 ADD + 15 SCALE chart edits; local paid delta 2H1 − 3H2 + H3 − H4 + H5; all public endpoint spans and the helper profile {4:51, 5:36, 7:24, 8:122, 9:294} unchanged |
| 2. PR5550 caller checker | Byte-pinned checker, four input lines changed in a temporary copy (as in PR5567); positive control reproduces the complete PR5550 output on PR5545, SHA-256 `b51a7d19…4009` | PASS on PR5551: 69,912 paid / 12,814 productive, 288 exact signed-Pauli and phase-anchor checks, 9,486 fresh helper checks, ordinary passes 127,292,680 ≤ 184,647,552; all ten PR5550 controls reject; a changed invoice is rejected |
| 3. Dense native inventory | PR5548's credited exact Gaussian-dyadic interpreter compares the lowered circuit with the literal frame circuit on all 512 columns: exact equality, norm and support size 2^rank | PASS on all 3,188 operators (`--full`); ranks match the paid transition ranks; a second, separately written driver gave the same 3,187 digests |
| 4. Own funding engine | 80-digit decimal ln/exp with 10⁻⁷⁰ relative padding, bisection on a 10⁻¹⁴ grid | Raw crossing (152076558357/5·10¹³, 60830623343/2·10¹³) = (.00304153116714, .00304153116715), equal to the author interval; .00304117, .00304153 and .0030415305 positive, .00304154 negative |

Native fault controls on the new map 615 → 207: a global phase fault with an unchanged signed tableau and a
flipped child sign are both rejected by the dense comparison.

## Reproduce

Python 3.12, standard library only, from the repository root:

```sh
python3.12 -B research/p104-joint-contact-native-caller-audit/verify.py
```

The default run is serial and takes about five minutes. It checks layers 1, 2 and 4 completely and a fixed
deterministic subset of layer 3: all four new maps, every helper endpoint pair, all 208 adapters and every
32nd MOVE map. Each subset matrix digest must equal the digest in [FULL-CHECK.json](FULL-CHECK.json). The
complete inventory runs with

```sh
python3.12 -B research/p104-joint-contact-native-caller-audit/verify.py --full --jobs 4
```

in about eleven minutes on four cores; it recomputes all 3,188 digests and compares them with FULL-CHECK.json.
The marker is `PASS_P104_JOINT_CONTACT_NATIVE_CALLER_AUDIT`. [PROOF.md](PROOF.md) explains each layer.

## Scope and limits

This audits the finite 9-qubit local word and its complete invoice. It does not materialize the uniform word,
so the uniform compiler, owned erasure, precision and tape costs and analytic recovery remain inherited premises,
as in PR5551 and PR5550. The selected record stays .003027288235; PR5551's certified endpoint is the author's
checker constant. This is an agent check, not human or proof-kernel acceptance. The checked native maps cover the
"native all-column implementation" premise only for this fixed word.

## Open next questions

- Binding the uniform-compiler contract of PR5550 to the FULL-CHECK.json digests, so the finite native check
carries to every recursion level. This is the remaining inherited native premise.
- Maintainer admission of PR5551, after which PR5553, PR5561 and PR5569 can be judged on an admitted parent.

## Credit

thu-nmrc (PR5551 construction and word). Chris R / Clear1nsight (PR5550 checker, used unchanged except four
input lines; PR5548 dense interpreter, imported unchanged). Ali Romman (PR5544/PR5545 and the P104 caller).
The P104 lineage and all retained native, Clifford, compiler and analytic contributors are credited in the
cited packages. This package was prepared for Matarener with Claude (Anthropic), via Claude Code. Apache-2.0.
99 changes: 99 additions & 0 deletions research/p104-joint-contact-native-caller-audit/RESULT.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,99 @@
{
"layer1_replay": {
"add_sites": [
[
10140,
297,
303
],
[
10141,
303,
297
],
[
12479,
152,
297
],
[
13062,
198,
319
],
[
13172,
205,
198
],
[
13173,
198,
205
]
],
"chart_edits": {
"ADD": 6,
"SCALE": 15
},
"endpoint_rows_checked": 743,
"helper_profile": {
"4": 51,
"5": 36,
"7": 24,
"8": 122,
"9": 294
},
"local_paid_delta": {
"1": 2,
"2": -3,
"3": 1,
"4": -1,
"5": 1
},
"moves": 3624,
"paid_local": {
"1": 2430,
"2": 554,
"3": 390,
"4": 111,
"5": 108,
"6": 17,
"7": 13,
"9": 1
}
},
"layer2_caller": {
"changed_invoice": "rejected",
"exact_native_checks": 288,
"fresh_helpers": 9486,
"ordinary_allowance": 184647552,
"ordinary_passes": 127292680,
"output_sha256": "b78408c1aa71698cb657d7bc6b6b31b2cc3eff29d4de751f31b96babee6027e1",
"paid_jobs": 69912,
"positive_control_pr5550_output": "b51a7d19c10faea7d569a48e9536db8ba1e7418f8dd8dfb25646cc87cc9c4009",
"productive_fields": 12814,
"status": "PASS_SCOPED_CALLER_AND_TARGET_BINDING"
},
"layer3_dense": {
"controls": {
"phase_only_fault": "rejected",
"wrong_child_sign": "rejected"
},
"full_check_sha256": "d33e47ca4cb9afb95a4f6e95a42f2173d2bb20037d0e998da3cba3d332a34efa",
"tasks_total": 3188
},
"layer4_funding": {
"engine": "decimal, 80 digits, correctly rounded ln/exp padded by 1e-70 relative; independent of both source engines and the PR5551 funding peer",
"raw_crossing": [
"152076558357/50000000000000",
"60830623343/20000000000000"
],
"signs": {
".00304117": "positive",
".00304153": "positive",
".0030415305": "positive",
".00304154": "negative"
}
}
}
102 changes: 102 additions & 0 deletions research/p104-joint-contact-native-caller-audit/SOURCES.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,102 @@
{
"schema_version": 1,
"base_commit": "5e0052ca0aa3036a3afa8b0438927377c8d97d53",
"files": [
{
"path": "certificates/selected-result.json",
"usage": "provenance",
"git_blob": "fc3d4ba105e1f097d0d7d051a308281e86ab1446",
"sha256": "66974efdced4083c00025f97d95d83909a226cdf801edb9301c195c6641f6903"
},
{
"path": "docs/research/agent-task-catalog.json",
"usage": "provenance",
"git_blob": "f126dd2e2dad2b669e75d1d0d0bd05c600725602",
"sha256": "f7a48bfa2659b88378282a8eee7d9fe52cdf505594124c4d1686b58510fffd72"
},
{
"path": "research/p104-joint-contact/WORD.json.gz",
"usage": "input",
"purpose": "PR5551 coupled-contact word (canonical package on main)",
"git_blob": "c04a815c2662bcd0c48ce9c3067a0fe4a041573a",
"sha256": "48f8734ca9c5adf5d60ded3a5fefbc24894d2c358ed7cb30fcc0d71718599446"
},
{
"path": "research/p104-joint-contact/RESULT.json",
"usage": "input",
"purpose": "PR5551 complete invoice and claimed raw root interval",
"git_blob": "a83db6d025b3755176449e679e58f8a2e16f3e5c",
"sha256": "36a5680f43c7da67e570b6c919de2934c6a5ce46db002993ff162aa36972d0ba"
},
{
"path": "research/p104-full-frame-contact/WORD.json.gz",
"usage": "input",
"purpose": "PR5545 predecessor word for the delta",
"git_blob": "1ce96b0130daeea46ef53fdbd32bca2727aa6b35",
"sha256": "a1e416add9d2bb44fce923111ebc50513e5d845e56aabe214f2f0bc7d5f2b45d"
},
{
"path": "research/p104-caller-independent-audit/CAPSULE-INVENTORY.json",
"usage": "input",
"purpose": "PR5550 capsule byte inventory",
"git_blob": "07c7b818f02c2aabab4798c3e8203ffd2fe1e827",
"sha256": "c3e525f5d42ad5b274b039db42f97a2b6e372b583d382c73f59dcb3142f1960d"
},
{
"path": "research/p104-caller-independent-audit/capsule/outputs/p104-caller-route-audit-v2-20261025/check.py",
"usage": "input",
"purpose": "PR5550 caller and target-binding checker (patched in a temporary copy only)",
"git_blob": "506882c655d02cbf1ce6a61aabc8ea15982561e8",
"sha256": "0e5e0e3d56165437fc2821f41927e3f0cae838c60841a93e8b5cb46caeb7657c"
},
{
"path": "research/p104-native-channel/vendor/clifford_exact.py",
"usage": "input",
"purpose": "canonical exact Clifford compiler (lower_child)",
"git_blob": "166841a037f069249f6f7f6dedb89248812f59e2",
"sha256": "2982ead07721e83358dc67c6f49c2b4c16a236a6bebdfd72f58221be1f5d0962"
},
{
"path": "research/p104-native-channel/vendor/native_verify.py",
"usage": "input",
"purpose": "canonical literal circuit descriptions",
"git_blob": "2c209dfa9cfade3bc20ac5fc9d876b433a147f8c",
"sha256": "3049471905b47b385c88378a1a3be7a18b8858aadd035b8a6da2f4a03b470f38"
},
{
"path": "research/p104-contact-independent-audit/capsule/outputs/p104-native-contact-audit-20261025/check.py",
"usage": "input",
"purpose": "PR5548 exact dense Gaussian-dyadic interpreter",
"git_blob": "e7e663bebaf13f44c95fb8eb118b472cf377e137",
"sha256": "43fd40847a0853c0ac511d58ce77ef822a5c6dd1b25f662f01cefaa3196a2f7d"
},
{
"path": "research/pr4359-p7-padding/native_inventory.py",
"usage": "input",
"purpose": "native inventory read by the PR5550 checker",
"git_blob": "8d2d96e3998254c2333c43fe47cdfd074cbd811e",
"sha256": "65539acaf45bdfe8c5aee6830f9c9b3d1107c442a60842ef803fdd7688de8f32"
},
{
"path": "research/kappa-0023-productive-jobs/dependencies/kappa-next-regime/direct-decoder/sources/DATA-PHASES.json",
"usage": "input",
"purpose": "phase data read by the PR5550 checker",
"git_blob": "d9a4f84211182ca200c081ea061d724c8a2cfb70",
"sha256": "c5916ff312678f12aba403847643fbbc29f643acec60bdf836fa56cd3e8251fa"
},
{
"path": "research/p104-joint-contact/PROOF.md",
"usage": "provenance",
"purpose": "PR5551 claim being audited",
"git_blob": "d1c5fb2211e95c3528b0b5ab99567012e73f98ef",
"sha256": "944e11a9a739b2e8cb5516f14959675d1eea7542b6508d7cfb034447e78260ac"
},
{
"path": "research/p104-caller-independent-audit/PROOF.md",
"usage": "provenance",
"purpose": "PR5550 helper-endpoint contract",
"git_blob": "3a74b71a55753a6e7cb372d98b0074cf98dc8708",
"sha256": "d73ff9587a85db8063391919ffc4a17521ccd5759cdaf6338ba036703239d839"
}
]
}
Loading