# AP30 — the OneDrive docx against the published paper

**2026-09-07 · a finding, not a ruling.** Written by the site desk at G's instruction after the republish of 7 September 2026.

## What this is

`420code_final/Standalone_Artists_Proofs/AP30_The_Resistance_FINAL.docx` (73 KB, last modified 26 May 2026) is **not** the text published at the420code.org. It differs from it in **17 substantive paragraphs**. The most visible: it carries *“zero fitted parameters”* where the published paper says *“zero free parameters”*.

Neither file is edited by this note. The published text is the paper of record — now `AP30_The_Resistance_FINAL_v1_0_dated_notes_2026-09-07.docx`, which is that same published body plus five dated notes. The OneDrive file is marked superseded.

**The open question for G, and the only one:** were any of these edits *intended*? If a sentence below is a correction G meant to make, it does not go into the body — the body is locked. It becomes a dated note beside it, as the notes of 6 and 7 September did. If none was intended, the OneDrive file is simply an earlier or divergent draft and the matter closes here.

## Method

Paragraph text from the docx; text from the PDF published to 6 September 2026 (sha256 `acb4621c1d3e6258de70897980427fc7f367dde3f57f38b1a1f35f926e3239ce`, preserved beside this note). Comparison is on letters and digits only, lower-cased — immune to line breaks, hyphenation, ligatures and page furniture. A paragraph counts as divergent only if its normalised text appears nowhere in the published text. Each is paired here with its closest published counterpart by sequence similarity; a low similarity score means the paragraph has no real counterpart and is an addition or a wholesale rewrite.

## What the 17 amount to

Every one of them runs in the same direction: **the OneDrive file is the earlier draft and the published paper is the edited one.** Not a single divergence is text that exists only in OneDrive and was lost from the published paper. They sort into four kinds:

| kind | count | example |
|---|---|---|
| **House style — "corpus" retired** | 8 | OneDrive *corpus* -> published *body of work*, *the 420 Code* |
| **"fitted" -> "free" parameters** | 3 | the phrase the whole claim rests on |
| **Registry citation de-versioned** | 5 | OneDrive *per MKSR v5.8* -> published *the Master Kill Switch Registry* |
| **Symbol and gloss corrections** | 2 | *h-bar* -> the correct glyph; and the axiom names spelled out (below) |

The one difference with real content is divergence 6, and it too is an improvement made *for* publication. OneDrive reads *"C (propagation), S (exchange), B (break)"*; the published paper reads *"C (Constraint — locality, read as propagation), S (Symmetry, read as exchange), B (Break)"*, naming the axioms rather than only their roles.

**A caution on reading the word-level diffs below.** Several show a `delete` of a trailing clause. Four of those were checked against the published text directly and the clause **is present** — the pairing window simply ended before it. The apparent deletions in divergences 2, 4, 5 and 7 are artifacts of the pairing, not losses. Divergence 6 is the only genuine content difference in the file.

**What this means for the decision.** On this evidence none of the OneDrive text is a correction G made after publication that went missing — the edits all went the other way, draft into published. If that reading is right, nothing here needs to become a dated note and the OneDrive file is simply the pre-editorial draft, correctly superseded. The judgement is G’s; the evidence is below in full so it can be made on the text rather than on this summary.

## The 17 divergences

### 1. similarity 95%

**OneDrive:**

> AP30 stands on the constants sector of Notebook IV: the leakage constant (AP06), the residual that reads every constant as a face of one break (AP24), and the channel count that derived gravity (AP28). It is the first mass ratio the corpus derives from {S, B, R, C}. If the method holds, the rest of the mass spectrum is in reach. If it fails, the kill switches say where.

**Published:**

> AP30 stands on the constants sector of Notebook IV: the leakage constant (AP06), the residual that reads every constant as a face of one break (AP24), and the channel count that derived gravity (AP28). It is the first mass ratio the body of work derives from {S, B, R, C}. If the method holds, the rest of the mass spectrum is in reach. If it fails, the kill switches say where. the420code.org

**Word-level:**

- `replace` — OneDrive: 'corpus' · published: 'body of work'
- `insert` — OneDrive: '' · published: 'the420code.org'

### 2. similarity 94%

**OneDrive:**

> AP30 derives the proton-to-electron mass ratio, m_p/m_e, from the four axioms {S, B, R, C} and one measured input, α ≈ 1/137. It reframes mass as geometric resistance, decomposes the integer 1836 into three independent static layers, adds the dynamic corrections the break’s self-maintenance forces, and lands within five parts per billion of the measured value at order α, with zero fitted parameters. The O(α²) term reduces the residual to about 0.008 parts per billion; the coefficient at that order (16) is carried in the verified formula but its structural decomposition from {21, 3, 4} is not yet established and is owed under KS-30.3, so the parameter-free claim is stated firmly at order α.

**Published:**

> AP30 derives the proton-to-electron mass ratio, m_p/m_e, from the four axioms {S, B, R, C} and one measured input, α ≈ 1/137. It reframes mass as geometric resistance, decomposes the integer 1836 into three independent static layers, adds the dynamic corrections the break’s self-maintenance forces, and lands within five parts per billion of the measured value at order α, with zero free parameters. The O(α²) term reduces the residual to about 0.008 parts per

**Word-level:**

- `replace` — OneDrive: 'fitted' · published: 'free'
- `delete` — OneDrive: 'billion; the coefficient at that order (16) is carried in the verified formula but its structural decomposition from {21, 3, 4} is not yet established and is owed under KS-30.3, so the parameter-free ' · published: ''

### 3. similarity 96%

**OneDrive:**

> §1 Definitions: framing and definition. §2 Lemma 1: structural argument (additivity from axiom independence; KS-30.1 live). §3 Lemma 2: derivation (the integer count 1764 + 63 + 9 = 1836). §4 Lemma 3: derivation (the dynamic maintenance term). §5 Lemma 4: derivation (the isotropic-leakage correction). §6 Proposition: derivation (the full ratio and its precision). §7 The Structural Residual: synthesis — the residual read as ε itself. §8 Connections: corpus integration.

**Published:**

> §1 Definitions: framing and definition. §2 Lemma 1: structural argument (additivity from axiom independence; KS-30.1 live). §3 Lemma 2: derivation (the integer count 1764 + 63 + 9 = 1836). §4 Lemma 3: derivation (the dynamic maintenance term). §5 Lemma 4: derivation (the isotropic-leakage correction). §6 Proposition: derivation (the full ratio and its precision). §7 The Structural Residual: synthesis — the residual read as ε itself. §8 Connections: the 420 Code

**Word-level:**

- `replace` — OneDrive: 'corpus integration.' · published: 'the 420 Code'

### 4. similarity 93%

**OneDrive:**

> Seven kill switches engage on this paper, all live. KS-30.1 (additivity formalisation), KS-30.2 (leakage isotropy), KS-30.3 (higher-order terms), and KS-30.4 (alternative formula) test the proton-mass derivation directly. KS-NPP.1, KS-NPP.2, and KS-NPP.3 test the neutron–proton mass difference (decomposition, sign, lattice-QCD consistency). KS-30.4 and the three KS-NPP switches were registered against AP30 through the Ø Predictions integration; their operational tests are carried in Ø Predictions (Part I and Part III) per MKSR v5.8.

**Published:**

> Seven kill switches engage on this paper, all live. KS-30.1 (additivity formalisation), KS-30.2 (leakage isotropy), KS-30.3 (higher-order terms), and KS-30.4 (alternative formula) test the proton-mass derivation directly. KS- NPP.1, KS-NPP.2, and KS-NPP.3 test the neutron–proton mass difference (decomposition, sign, lattice-QCD consistency). KS-30.4 and the three KS-NPP switches were registered against AP30 through the Ø Predictions integration;

**Word-level:**

- `replace` — OneDrive: 'KS-NPP.1,' · published: 'KS- NPP.1,'
- `delete` — OneDrive: 'their operational tests are carried in Ø Predictions (Part I and Part III) per MKSR v5.8.' · published: ''

### 5. similarity 90%

**OneDrive:**

> AP24 (The Residual), locked in Notebook IV, treats the proton mass as one of six measured faces of ε; AP30 derives that face from {S, B, R, C}, retiring its empirical status. The relationship is framework-and-value — AP30 inherits AP24’s six-face reading and supplies the value of one face — not a dependency cycle. AP30 opens the matter sector that AP22 (The Ledger) and AP26 (The Surplus) complete. With AP30 in place, the corpus expresses all fundamental scales — c, ħ, and G (AP03, AP28) and now m_p/m_e — in terms of the axioms and the single measured input.

**Published:**

> AP24 (The Residual), locked in Notebook IV, treats the proton mass as one of six measured faces of ε; AP30 derives that face from {S, B, R, C}, retiring its empirical status. The relationship is framework-and-value — AP30 inherits AP24’s six-face reading and supplies the value of one face — not a dependency cycle. AP30 opens the matter sector that AP22 (The Ledger) and AP26 (The Surplus) complete. With AP30 in place, the body of work expresses all

**Word-level:**

- `replace` — OneDrive: 'corpus' · published: 'body of work'
- `delete` — OneDrive: 'fundamental scales — c, ħ, and G (AP03, AP28) and now m_p/m_e — in terms of the axioms and the single measured input.' · published: ''

### 6. similarity 90%

**OneDrive:**

> 3: the number of faces of the manifold (AP10). Identified with the three colours of SU(3) (AP19). Each face corresponds to one of three structural roles: C (propagation), S (exchange), B (break).

**Published:**

> 3: the number of faces of the manifold (AP10). Identified with the three colours of SU(3) (AP19). Each face corresponds to one of three structural roles: C

**Word-level:**

- `delete` — OneDrive: '(propagation), S (exchange), B (break).' · published: ''

### 7. similarity 95%

**OneDrive:**

> Among them, exactly one is hierarchically nested — each successive term using fewer corpus numbers and corresponding to a distinct geometric feature (manifold capacity, face projection, exchange matrix). This hierarchical decomposition is the one used here; whether it is the unique admissible decomposition under a precise enumeration constraint is exactly what KS-30.4 holds open, and the constraint set and the count of admissible decompositions under it are not exhibited in this paper.

**Published:**

> Among them, exactly one is hierarchically nested — each successive term using fewer body of work numbers and corresponding to a distinct geometric feature (manifold capacity, face projection, exchange matrix). This hierarchical decomposition is the one used here; whether it is the unique admissible decomposition under a precise enumeration constraint is exactly what KS-30.4 holds open, and the constraint set and the count of admissible decompositions

**Word-level:**

- `replace` — OneDrive: 'corpus' · published: 'body of work'
- `delete` — OneDrive: 'under it are not exhibited in this paper.' · published: ''

### 8. similarity 93%

**OneDrive:**

> Kill switch numbers are globally unique across the corpus. Seven switches engage on this paper; all are live.

**Published:**

> Kill switch numbers are globally unique across the body of work. Seven switches engage on this paper; all are live.

**Word-level:**

- `replace` — OneDrive: 'corpus.' · published: 'body of work.'

### 9. similarity 95%

**OneDrive:**

> Test. Exhibit a different formula in corpus numbers that matches m_p/m_e to equal precision with a different structural interpretation.

**Published:**

> Test. Exhibit a different formula in the 420 Code numbers that matches m_p/m_e to equal precision with a different structural interpretation.

**Word-level:**

- `replace` — OneDrive: 'corpus' · published: 'the 420 Code'

### 10. similarity 88%

**OneDrive:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part I); operational test carried there per MKSR v5.8.

**Published:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part I); operational test carried there per the Master Kill Switch Registry.

**Word-level:**

- `replace` — OneDrive: 'MKSR v5.8.' · published: 'the Master Kill Switch Registry.'

### 11. similarity 90%

**OneDrive:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part III) per MKSR v5.8.

**Published:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part

**Word-level:**

- `delete` — OneDrive: 'III) per MKSR v5.8.' · published: ''

### 12. similarity 90%

**OneDrive:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part III) per MKSR v5.8.

**Published:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part

**Word-level:**

- `delete` — OneDrive: 'III) per MKSR v5.8.' · published: ''

### 13. similarity 90%

**OneDrive:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part III) per MKSR v5.8.

**Published:**

> Status. LIVE — STRUCTURAL. Registered against AP30 via Ø Predictions (Part

**Word-level:**

- `delete` — OneDrive: 'III) per MKSR v5.8.' · published: ''

### 14. similarity 97%

**OneDrive:**

> The proton-to-electron mass ratio is derived from {S, B, R, C} and one measured input, to within five parts per billion at order α with zero fitted parameters; the O(α²) term reduces the residual to about 0.008 parts per billion, using a coefficient whose structural decomposition is owed under KS-30.3. Mass is established as geometric resistance, and the integer 1836 is decomposed into three independent static layers, each tracing to a distinct axiom subset.

**Published:**

> The proton-to-electron mass ratio is derived from {S, B, R, C} and one measured input, to within five parts per billion at order α with zero free parameters; the O(α²) term reduces the residual to about 0.008 parts per billion, using a coefficient whose structural decomposition is owed under KS-30.3. Mass is established as geometric resistance, and the integer 1836 is decomposed into three independent static layers, each tracing to a distinct

**Word-level:**

- `replace` — OneDrive: 'fitted' · published: 'free'
- `delete` — OneDrive: 'axiom subset.' · published: ''

### 15. similarity 97%

**OneDrive:**

> With AP30 in place, the corpus expresses every fundamental scale in terms of the axioms and one measured input: c, ħ, and G through AP03 and AP28, and now m_p/m_e through the channel geometry. The constants are no longer a list of independent numbers; they are faces and weights of one break.

**Published:**

> With AP30 in place, the body of work expresses every fundamental scale in terms of the axioms and one measured input: c, ℏ, and G through AP03 and AP28, and now m_p/m_e through the channel geometry. The constants are no longer a list of independent numbers; they are faces and weights of one break.

**Word-level:**

- `replace` — OneDrive: 'corpus' · published: 'body of work'
- `replace` — OneDrive: 'ħ,' · published: 'ℏ,'

### 16. similarity 96%

**OneDrive:**

> §1 Definitions [FRAMING / DEFINITION]. Mass is geometric resistance; the electron is the unit; the corpus numbers 21, 3, 4 and α are installed.

**Published:**

> §1 Definitions [FRAMING / DEFINITION]. Mass is geometric resistance; the electron is the unit; the 420 Code numbers 21, 3, 4 and α are installed.

**Word-level:**

- `replace` — OneDrive: 'corpus' · published: '420 Code'

### 17. similarity 98%

**OneDrive:**

> §6 Proposition [DERIVATION]. m_p/m_e = 1836 + α×21×(1 − 1/(84π)) + O(α²): the O(α) term gives 5 ppb with zero fitted parameters; the O(α²) term reduces the residual to ~0.008 ppb (1836.15267344 vs measured 1836.15267343) using a coefficient whose structural derivation is owed under KS-30.3.

**Published:**

> §6 Proposition [DERIVATION]. m_p/m_e = 1836 + α×21×(1 − 1/(84π)) + O(α²): the O(α) term gives 5 ppb with zero free parameters; the O(α²) term reduces the residual to ~0.008 ppb (1836.15267344 vs measured 1836.15267343) using a coefficient whose structural derivation is owed under KS-30.3.

**Word-level:**

- `replace` — OneDrive: 'fitted' · published: 'free'

---

*Nothing here is edited into either file. Frozen bodies are never edited; corrections supersede, dated, in the open.*