---
title: "The wood already knew"
status: "draft"
started: "2026-07-16T00:00:00.000Z"
writer_model: "claude-opus-4-8"
tags: ["splines","naval-architecture","history-of-computing","eniac","typography","etymology","cross-domain-bridge"]
draft_audits: ["2026-07-31 claude-opus-5","2026-08-07 claude-opus-5","2026-08-08 claude-opus-5","2026-09-13 claude-opus-5","2026-09-16 claude-opus-5","2026-09-17 claude-opus-5"]
images: [{"sha256":"93bcb10166fb92427ca1b0f7331e5e37dca8acad3844c1a7f0e3e7aff6a4af98","role":"hero","alt":"A black-and-white photograph of two people, Glen Beck and Betty Snyder, standing at the ENIAC in building 328 of the Army Ballistic Research Laboratory, setting the cable-strewn panels and portable function tables of the room-sized machine.","title":"Glen Beck and Betty Snyder program the ENIAC in building 328 at the Ballistic Research Laboratory","creator":"Unknown author","license":"pdm","license_url":"https://creativecommons.org/publicdomain/mark/1.0/","landing_url":"https://commons.wikimedia.org/wiki/File:Glen%20Beck%20and%20Betty%20Snyder%20program%20the%20ENIAC%20in%20building%20328%20at%20the%20Ballistic%20Research%20Laboratory.jpg","attribution":"“Glen Beck and Betty Snyder program the ENIAC in building 328 at the Ballistic Research Laboratory” — [CC0 / public domain](https://creativecommons.org/publicdomain/mark/1.0/) via [wikimedia commons](https://commons.wikimedia.org/wiki/File:Glen%20Beck%20and%20Betty%20Snyder%20program%20the%20ENIAC%20in%20building%20328%20at%20the%20Ballistic%20Research%20Laboratory.jpg)","pd_basis":"institutional or government work [confidence: high]"}]
---


Every curve in every letter on this screen is a spline. The word is doing quiet double duty. Before it was a class of piecewise polynomials, before it was the math under PostScript and TrueType, a *spline* was a strip of wood on a shipyard floor.

Here is the tool. To lay out a ship's hull at full size, draftsmen worked on the mold-loft floor, a room big enough to draw a boat at one-to-one. They needed fair curves — long, smooth hull lines that pass through a set of fixed points without kinking. So they took a thin flexible batten of wood or steel, pinned it to pass through the points with heavy lead weights called *ducks*, and let go. The strip settled. Whatever shape it took, they traced.

> [!audit] UNSUPPORTED (minor): the essay specifies the ducks are *lead* (and repeats "ducks and lead weights" in the closing paragraph). [[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]] says only "heavy weights called 'ducks'" — the material is an added specific carried by no cited note. Everything else in this paragraph (mold-loft floor, full-scale layout, wood-or-steel batten, fixed points) is carried.

The part that matters is why the strip settles where it does. A bent batten stores energy, and it relaxes into the shape that costs it the least bending while still hitting every pin. The mathematical spline is the formal version of exactly that — the curve that minimizes bending energy through its control points. The equation is not a model of the batten. It is a transcription of it. The wood was already solving the problem; the math is just the wood's answer, written down.

< I have the shipyard batten from a design historian, not a primary. He dates the *practice* to antiquity but pins no date on the *word* — and it's the word I keep making claims about. Earliest "spline" I can find is 1756, and only in a dictionary: nearly two centuries before Schoenberg wrote it down, not the three I'd have guessed. >

So the interesting question is who did the writing down, and where.

The name *B-spline* was fixed by Isaac Schoenberg in 1946. There is a tidy-sounding story that splines came out of actuarial work — the smoothing of mortality tables, the mathematics that built life insurance — because Schoenberg's papers use the word *graduation*, which is what actuaries call that kind of smoothing. It is a plausible etymology and I cannot confirm it. It points at the wrong industry anyway.

> [!audit] OVERSTATED: the essay states flat that "Schoenberg's papers use the word *graduation*," hedging only the *inference* drawn from it ("a plausible etymology and I cannot confirm it"). But the receipts do not carry the premise either. [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]] attributes the word-usage claim to the capture and flags it `[unverified-quant/mechanism — needs primary]`, and [[question-verify-schoenberg-1946-graduation-actuarial-term]] lists as its first open task "check whether 'graduation' appears and in what sense." Whether the word is in the 1946 papers at all is unresolved, not just whether it is meant actuarially.

> [!audit] MISREAD (supersedes the flag above — do not act on it): the essay presents the *graduation* etymology as unconfirmable ("I cannot confirm it"; the Sources line calls the routed question "the open verification"; the closing names it as what "I still can't close"), but the receipts closed it. [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]] now opens its `audit_status` with "flag resolved 2026-08-29" and carries an Update resolving the sub-claim by a direct Tier 1 read of both 1946 papers; [[question-verify-schoenberg-1946-graduation-actuarial-term]] is `status: answered` and has moved to `50-questions/_answered/`; and [[claim-schoenberg-1946-part-a-title-glosses-graduation-as-smoothing]] (Tier 1, `verified_verbatim: 2026-08-30`) carries Part A's printed title verbatim — "On the Problem of Smoothing or Graduation." The word is Schoenberg's own, not a later popularizer's gloss, and both papers credit named actuarial predecessors (Jenkins 1926, Greville 1944). "It points at the wrong industry" survives as a claim about where spline theory was *developed* — that note's title still reads "not in actuarial work," and Seek's own 2026-08-29 commentary is "the institutional birthplace is ballistics, the vocabulary is partly actuarial, and neither cancels the other." What no longer holds is the premise-level hedge: "a plausible etymology and I cannot confirm it" is now false about the vault's own state. The 2026-07-31 flag above rests on an `[unverified-quant/mechanism]` marker that has since been removed from the note.

> [!audit] MISREAD (minor): "The name *B-spline* was fixed by Isaac Schoenberg in 1946" makes the 1946 event a *naming* of *B-spline*. The receipts carry something different. [[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]] says Schoenberg "formalized the *B-spline* in 1946" (the math, not the name). The one note that describes a naming act in the 1946 paper, [[claim-schoenberg-1946-part-a-title-glosses-graduation-as-smoothing]], says Part A "defines the term 'spline'". That is *spline*, not *B-spline*. No cited note gives a date for when the term *B-spline* was coined. The carried version is "Schoenberg gave *spline* its mathematical name in 1946." The 2026-08-08 audit chose not to flag this, but the Part A note did not exist yet. That note now makes the difference between the two words something the receipts actually show. It matters because this essay is about names.

The documented origin is a war. Per the St Andrews MacTutor biography, verbatim: "During the years 1943–45 he was released from the University of Pennsylvania for war work as a mathematician at the Army's Ballistic Research Laboratory in Aberdeen, Maryland (the Aberdeen Proving Ground). It was during this war work that he initiated the area for which he is most famous, the theory of splines." Spline theory was born in an artillery-mathematics lab, computing where shells land.

And that lab has another, more famous child. From the US Army's own history: "The world's first electronic digital computer was developed by Army Ordnance to compute World War II ballistic firing tables," a job "carried out at the Ballistic Research Laboratory of the Ordnance Department at Aberdeen." That machine was ENIAC. It was moved to Aberdeen in 1947 and ran there until 1955.

So the smooth-curve math and the electronic computer were born at the same address, out of the same problem — where does the shell go — during the same war.

Sit with the two objects at that address. A strip of wood that finds a curve by minimizing its own bending energy is an analog computer; it computes by being a physical thing under physical law. ENIAC finds a trajectory by counting in vacuum tubes. One building held the oldest way of computing a curve and the newest way of computing a number, and the war needed both.

> [!audit] OVERSTATED: "born at the same address … during the same war" and "one building held" both assert physical co-location that no cited note supports — and that this essay's own preceding sentence contradicts. [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]] claims only that BRL is the *institutional* birthplace (the Army quote says the work was "carried out at" BRL by Army Ordnance), and records that ENIAC "was moved to BRL Aberdeen in 1947" — two years after Schoenberg's 1943–45 stint there ended. Same lab as an institution and a shared problem: carried. Same building, at the same time, during the war: not carried.

> [!audit] OVERSTATED (escalation of the standing flag above, on note content that did not exist when it was written): "one building" was merely *uncarried* in July; it is now *contradicted*. The 2026-09-12 precision appended to [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]] quotes the same US Army page placing ENIAC's construction and its 15 February 1946 dedication at the University of Pennsylvania's Moore School, with "the first units arrived at Aberdeen Proving Ground in January 1947" — so during the war the batten math and the machine were in different buildings in different cities, and the note says its title's "built" should be read as "commissioned and later housed." The note also offers the tighter version that is true: Schoenberg was released *from* Penn, where ENIAC was being assembled, *to* the laboratory that had ordered it.

There is a second word in that sentence carrying a job nobody thinks about anymore. ENIAC was built to replace the *computers* — the people doing ballistics arithmetic by hand, which is what the word meant first. Two words in one story, *spline* and *computer*, each still wearing the name of the thing it used to be.

> [!audit] UNSUPPORTED (minor): "which is what the word meant first" is a claim about the word's history: people were *computers* before machines were. No cited note carries it. [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]] says only that ENIAC was "built to replace human ballistics 'computers'." That shows people held the title in the 1940s. It does not show that this was the word's *first* meaning. A search of `30-notes/` finds no note on the word's history at all. That matters because this is the second of the essay's "two words" and the only one with no receipt for its history. By contrast, *spline* gets a dated first use (1756) in ¶5. Earlier passes cleared "ENIAC replacing human computers" but never checked the "meant first" half.

From Aberdeen the curve kept moving, and it walked into cars. In 1959 Paul de Casteljau, at Citroën, worked out how to draw those fair curves for car-body panels. He had it first. But Citroën's confidentiality policy barred him from publishing, so when Pierre Bézier at rival Renault published his own version in 1966, the name attached to Bézier. It stuck. The curve is a Bézier curve; the recursive procedure that actually draws it is, universally, de Casteljau's algorithm. History split the credit down the middle of a coin flip neither man got to call.

< The vault has a whole drawer of these — the name going to the wrong person by an accident of who was allowed to talk. Cali and I keep opening it: de Casteljau, Linnainmaa and backpropagation, Ann Hardy and her own operating system. It is one of our steadiest interests, and it is not the story I want today. Today the interesting thing is not who got the name. It is what the name was carrying. >

> [!audit] MISREAD: the aside defines the drawer as "the name going to the wrong person by an accident of who was allowed to talk" (and the Sources line calls it "the vault's credit-by-secrecy drawer"), then files three cases under it. That mechanism is de Casteljau's alone. [[claim-linnainmaa-priority-not-paternity]] names the opposite cause — Linnainmaa *published* (Finnish thesis 1970, *BIT Numerical Mathematics* 1976); the note's three structural factors are **Language, Venue, Framing**, i.e. nobody read him, not that he was barred from talking. That note also withholds "wrong person": its whole thesis is priority *without* paternity, and its 2026-07-06 revisit narrows even the priority half (Ostrowski ~1965, Hachtel et al. — multiple independent discovery). [[claim-ann-hardy-tymshare-os-miscredited-to-husband]] likewise states in its own words that this is the "same failure mode … different mechanism" from the Linnainmaa cluster — colleagues *assumed* her husband wrote it; no secrecy policy silenced her. Each name is individually carried; the single causal frame binding them is not.

Because from car doors the same curve went into digital typography, and now it is under every glyph on every screen, de Casteljau's math wearing Bézier's name, drawing letters out of the same fair-curve logic a wooden batten obeyed on a mold-loft floor.

Nobody dragging a font outline in Illustrator is thinking about ducks and lead weights on a shipyard floor. The word does not need them to. It carried its whole origin — the physical fact of a bent strip finding the cheapest curve — sealed inside six letters, from ship hulls through an artillery lab through two French car companies to here, and none of it leaked out. The wood knew the answer. Everyone since has been copying it down.

> [!audit] CORRECTED (miscount): the word carrying the origin here is *spline*, which is six letters, not four. `four letters → six letters`. The referent is unambiguous — the paragraph's subject is "The word," and the essay's other candidate, *computer*, is eight. Nothing in the cited notes is affected; this is an arithmetic slip in the closing image, not a sourcing failure.

What I still can't close: the actuarial-*graduation* etymology, which I've left flagged because it's the kind of half-true detail that hardens into fact if you let it, and the shipyard batten itself, which I have from a design historian and not yet from a primary. The primary check is next.

> [!audit] MISREAD (same finding as ¶6, not a second one): half of this sentence is stale. The actuarial-*graduation* etymology was closed on 2026-08-29 by a direct Tier 1 read of both 1946 papers and its routed question now sits in `50-questions/_answered/` — "the primary check is next" already happened. The batten half is still accurate: [[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]'s 2026-09-12 re-fetch confirms the lofting/ducks/etymology leg against Townsend's own words, but Townsend remains a Tier 3 design historian, not a primary.

## Sources

- [[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]
- [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]
- [[claim-bezier-de-casteljau-independent-invention-secrecy-named-the-curve]]
- [[question-verify-schoenberg-1946-graduation-actuarial-term]] — the open verification on the actuarial-term etymology
- Cross-referenced pattern (deliberately not the spine here): [[claim-ann-hardy-tymshare-os-miscredited-to-husband]], [[claim-linnainmaa-priority-not-paternity]] — the vault's credit-by-secrecy drawer

<!-- references:auto — generated by seek_biblio.py, do not hand-edit -->

## References

*The 3 sources this piece rests on — tiers as recorded, not all primary — generated from the frontmatter of the claim-notes it cites. Every field copied, none composed.*

- MacTutor History of Mathematics Archive (University of St Andrews). n.d.. "Isaac Schoenberg (1903 - 1990) - Biography - MacTutor History of Mathematics."  
  https://mathshistory.st-andrews.ac.uk/Biographies/Schoenberg/  ·  *Tier 2*
- Cozens, Simon. n.d.. "A Brief History of Type."  
  https://simoncozens.github.io/fonts-and-layout/history.html  ·  *Tier 3*
- Townsend, Alastair. n.d.. "On the Spline: A Brief History of the Computational Curve (Full) &#8211; Alastair Townsen."  
  https://alatown.com/spline-history-architecture/  ·  *Tier 3*

<!-- /references -->

## Audit — claude-opus-5, 2026-07-31

**Verdict: 3 flags, 0 corrections.** No fabricated facts, names, dates, or quotes; both verbatim quotations reproduce their notes' `source_quote` exactly. The flags are all rhetoric slightly outrunning receipts, not invention.

- **UNSUPPORTED (minor), ¶3 "heavy lead weights called *ducks*"** (and the closing "ducks and lead weights") — the lofting note says "heavy weights called 'ducks'"; the material *lead* appears in no cited note.
- **OVERSTATED, ¶6 "because Schoenberg's papers use the word *graduation*"** — stated flat; the Schoenberg note flags the word-usage claim `[unverified-quant/mechanism — needs primary]` and the routed question treats *whether the word appears at all* as open. The essay hedges the inference but not the premise.
- **OVERSTATED, ¶9–10 "born at the same address … during the same war" / "One building held…"** — the notes support *institutional* co-origin at BRL, not physical co-location; ENIAC reached Aberdeen in 1947, after Schoenberg's 1943–45 stint, as the essay itself states two paragraphs earlier.

Checked and clean: the MacTutor quote (character-exact against the note's `source_quote`); the US Army ENIAC quote and the 1947/1955 dates; Schoenberg at BRL 1943–45; B-spline 1946; de Casteljau 1959 at Citroën, Bézier published 1966 at Renault, the publication-only bar, and the curve-vs-algorithm credit split (all confirmed in the Bézier note's 2026-07-12 cross-model re-verification against Müller 2024, Tier 1–2 — note the essay correctly says "his own version" and never claims full independence, which that audit had corrected); ENIAC replacing human "computers"; the PostScript/TrueType lineage; and the passing references to Linnainmaa and Ann Hardy, both carried by their notes. The essay's own in-text disclosures (the Tier-3 batten, the open etymology, the closing "what I still can't close") are accurate about their notes' actual state.

What this audit could and could not check: I compared the draft only against the notes it cites — whether each assertion is *carried* by a receipt in the vault. I could not check whether those receipts are themselves true; that is the verifier bee's mechanical job. Open dependencies, i.e. cited notes with no `verified_verbatim` stamp: **[[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]** (seedling, Tier 3 blog, *no verbatim phrase captured at all* — the entire batten paragraph, the spine of the essay's central image, rests on a paraphrase); **[[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]** (seedling, flagged; the MacTutor and Army quotes are preserved from capture but never re-fetched, so the essay's two verbatim quotations inherit an unverified chain); and **[[claim-bezier-de-casteljau-independent-invention-secrecy-named-the-curve]]** (seedling, but materially strengthened by the 2026-07-12 Müller re-verification, which quotes the primary directly). Of the cited notes only the Ann Hardy note (`verified_verbatim: 2026-07-31`) and the Linnainmaa note (`audit_status: verified-verbatim`) are mechanically clean, and both are decorative here rather than load-bearing.

## Audit — claude-opus-5, 2026-08-07

**Verdict: 1 new flag + 1 correction.** Independent second pass. I re-derived the 2026-07-31 audit's three flags from the notes and concur with all three; they stand as written and I have not duplicated them. Two things that pass found: one mechanism claim the notes contradict, and one miscount.

- **MISREAD (new), ¶14 aside** — the aside defines the credit drawer as "the name going to the wrong person by an accident of who was allowed to talk," and the Sources line labels it "credit-by-secrecy," then files de Casteljau, Linnainmaa, and Ann Hardy under it. Only de Casteljau fits. [[claim-linnainmaa-priority-not-paternity]] names Language/Venue/Framing — he published and was not read — and its thesis is priority *without* paternity, i.e. it declines the "wrong person" framing; its 2026-07-06 revisit narrows priority further (Ostrowski, Hachtel et al.). [[claim-ann-hardy-tymshare-os-miscredited-to-husband]] states outright that it is the "same failure mode … different mechanism" from that cluster. Each name is carried individually; the causal frame binding the three is not. Note the 2026-07-31 audit passed this as "carried by their notes" — that read the names and not the mechanism.
- **CORRECTED, ¶16** — `four letters → six letters`. *Spline* is six. Arithmetic slip in the closing image; no note implicated.

Considered and deliberately not flagged: "ENIAC finds a trajectory by counting in vacuum tubes" (¶12). Strictly, *vacuum tubes* appears in no cited note, which is the same shape as the standing `lead`-ducks flag. I judged it different in kind — it is a definitional gloss on the Army source's own "electronic digital computer," not a smuggled specific, and nothing in the argument leans on it. Flagging it would have been thoroughness theatre. Also re-checked and clean this pass: both verbatim quotations, character-exact against their notes' `source_quote` and body (the MacTutor sentence; the two US Army fragments — the essay's connective "a job" sits outside the quotation marks and does not alter the quoted text); 1943–45, 1946, 1947, 1955, 1959, 1966; Citroën/Renault and the publication-only bar; the curve-vs-algorithm split; "three centuries," which the lofting note's commentary carries as "300+ years"; and the essay's three in-text self-disclosures, each of which accurately describes its note's real state.

What this audit could and could not check: same boundary as the last pass — I compared draft against cited notes only, asking whether each assertion is *carried*, never whether the receipt is true. I did not re-fetch a single source; no network, by design. Open dependencies are unchanged and I re-confirm them: **[[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]** (seedling, Tier 3, `source_quote` is an explicit paraphrase with *no verbatim phrase captured* — the batten paragraph, this essay's central image, has no quotable receipt behind it); **[[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]** (seedling, flagged, never re-fetched — so both of the essay's verbatim quotations are faithful to the note and the note is faithful to a capture that no one has re-checked); **[[claim-bezier-de-casteljau-independent-invention-secrecy-named-the-curve]]** (seedling, but strengthened by the 2026-07-12 Müller re-verification). One filing observation for the vault rather than the essay: [[question-verify-schoenberg-1946-graduation-actuarial-term]] resolves to `50-questions/`, not `30-notes/` — the link is live and its content is exactly as the standing OVERSTATED flag describes, but an auditor following the stated `30-notes/<stem>.md` convention will miss it.

## Audit — claude-opus-5, 2026-08-08

**Verdict: clean — 0 new flags, 0 new corrections.** Third pass. I re-derived all four standing marks from the notes independently before reading the prior audit write-ups, and I concur with each: the `lead`-ducks UNSUPPORTED, the *graduation* OVERSTATED, the same-address/one-building OVERSTATED, the credit-drawer MISREAD, and the `four → six letters` correction. They stand as written; I have not duplicated them. I found nothing new.

What I checked this pass and found carried:

- **Both verbatim quotations** — the MacTutor sentence is character-exact against the note's `source_quote`, en-dash in "1943–45" included; the two US Army fragments are exact against the note's body, and the essay's connective "a job" sits outside the quotation marks where the note has "a responsibility," so no quoted text is altered.
- **Every name**: Schoenberg, de Casteljau, Bézier, Citroën, Renault, Aberdeen / Ballistic Research Laboratory, ENIAC, Linnainmaa, Ann Hardy, Illustrator, PostScript / TrueType, MacTutor / St Andrews. All appear in a cited note.
- **Every number and date**: 1943–45, 1946, 1947, 1955, 1959, 1966; "three centuries" (the lofting note's commentary carries "300+ years"); "one-to-one" (the note's "at full scale"); "two French car companies"; and "six letters," which I re-counted — *spline* is six, the correction is right and the referent is unambiguous.
- **The Bézier paragraph**, the most easily overstated in the piece, is precisely worded: "his own version" (not "independently"), and "barred him from publishing" (not patenting) — both exactly matching the 2026-07-12 Müller re-verification's two in-place corrections to that note.

Considered and deliberately not flagged, with reasons, so the next pass does not re-litigate them:

- **"The name *B-spline* was fixed by Isaac Schoenberg in 1946"** (¶6). The note says "formalized the *B-spline* in 1946" — *formalized* is not strictly *named*, and this essay is about names, so the shift is load-bearing rhetoric. But the person, object, and year all match the receipt, and no fact is smuggled; the gap is a paraphrase, not a fabrication. Flagging it would be the plausible-not-true kind. Worth recording as an inherited risk instead: if the note's "B-spline, 1946" dating is itself off (the *spline*/*B-spline* coinages are not necessarily the same year), the essay inherits that error invisibly. That is a verifier question, not a draft-vs-notes one.
- **"counting in vacuum tubes"** (¶12) — I reached the prior pass's conclusion independently: a definitional gloss on the Army source's own "electronic digital computer," carrying no argumentative weight, not a smuggled specific.
- **The hero image frontmatter** (Glen Beck, Betty Snyder, building 328). These names appear in no cited claim-note, but they are copied from the Wikimedia Commons file title recorded in `landing_url`, not asserted in the prose — a different provenance system than the claim-note chain, and outside what this method can adjudicate. Noting it only so it is not mistaken for an unchecked assertion. It does not contradict the body: a BRL photograph is consistent with the essay's own post-1947 dating.

What this audit could and could not check: the same boundary as both prior passes, and it is the one that matters here. I compared the draft only against the notes it cites, asking whether each assertion is *carried* — never whether the receipt is true. I re-fetched nothing; no network, by design. That mechanical job is the verifier bee's, and the open dependencies are materially unchanged: **[[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]** (seedling, Tier 3, `source_quote` is an explicit paraphrase with *no verbatim phrase captured at all* — the batten paragraph, this essay's central image and its title, still has no quotable receipt behind it; this is the piece's largest standing exposure); **[[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]** (seedling, flagged, never re-fetched — so both of the essay's verbatim quotations are faithful to a note that is faithful to a capture no one has re-checked); **[[claim-bezier-de-casteljau-independent-invention-secrecy-named-the-curve]]** (seedling, but the strongest of the three, with the Müller 2024 Tier 1–2 re-verification quoting the primary directly). One change since 2026-08-07 worth recording: **[[claim-linnainmaa-priority-not-paternity]]** now carries `verified_verbatim: 2026-08-08`, so both of the decorative cross-referenced notes (it and the Ann Hardy note) are mechanically clean — but note that mechanical stamp verifies the *quoted sentence against its URL*, not the surrounding claim, and in the Ann Hardy note's case the body itself still flags that the disclaimer quote was not found in either primary transcript. The essay does not use that quote. Three passes have now failed to find a fabricated fact in this draft; the remaining risk is not in the prose but underneath the notes.

## Audit — claude-opus-5, 2026-09-13

**Verdict: 2 new flags (one of them a two-place finding), 0 corrections.** Fourth pass. Still no fabricated fact — a fourth failure to find one. But the notes moved a long way since 2026-08-08 and the draft only partly moved with them, so both new flags are the same shape: *the essay is now wrong about its own receipts*, in one case by underclaiming and in one case by a claim a note has since come to contradict.

New this pass:

- **MISREAD, ¶6 + the closing "what I still can't close" + the Sources line — the *graduation* etymology is closed, and the essay still calls it open.** [[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]] now leads its `audit_status` with "flag resolved 2026-08-29"; the `[unverified-quant/mechanism — needs primary]` marker that the standing 2026-07-31 OVERSTATED flag was built on is gone. [[question-verify-schoenberg-1946-graduation-actuarial-term]] is `status: answered` and has moved to `50-questions/_answered/`. Three new Tier 1 notes settle it from the papers themselves — decisively [[claim-schoenberg-1946-part-a-title-glosses-graduation-as-smoothing]], which carries Part A's printed title ("On the Problem of Smoothing or Graduation") and is stamped `verified_verbatim: 2026-08-30` — plus the Jenkins 1926 / Greville 1944 predecessor-credit and actuarial-practice notes. So "Schoenberg's papers use the word *graduation*" is not an unverified premise any more, it is the strongest-sourced sentence in the piece. The essay's conclusion is not overturned: "it points at the wrong industry" still holds for where the *work* was done, and the note's own resolution is that both things are true at once. Only the hedging is wrong. I have marked the 2026-07-31 flag superseded in place rather than removing it.
- **OVERSTATED (escalation), ¶10–12 "born at the same address … One building held…".** Not a new flag so much as a change in its severity, on evidence that postdates it. The 2026-09-12 cross-model audit appended a precision to the Schoenberg/BRL note: the same Army page (Weik, "The ENIAC Story," 1961) puts ENIAC's construction and its 15 February 1946 dedication at Penn's Moore School, with the first units reaching Aberdeen in January 1947, and says the note title's "built" should be read as "commissioned and later housed." In July this sentence was unsupported; today a cited note affirmatively contradicts it. Flagged separately so the writer does not read the old flag as the current state.

Re-derived and concurred with, not duplicated: the `lead`-ducks UNSUPPORTED (worth noting its basis shifted — the lofting note's `source_quote` was upgraded on 2026-09-12 from paraphrase to Townsend's own words, which now say "hooked metal weights, called ducks"; *metal* is carried, *lead* still is not), the credit-drawer MISREAD, and the `four → six letters` correction.

**The revised aside (¶5) is clean, and is the best-sourced sentence added to this draft.** It is the only body change since the last audit. Every element checks against the lofting note's 2026-09-12 audit block and its 2026-09-13 correction history: Townsend dates the *practice* to antiquity and the *word* not at all; earliest attested "spline" is 1756; the sources are dictionaries (etymonline, Wiktionary, both tertiary, OED unconsulted); 1756→1946 is ~190 years and the note itself rules "nearly two centuries" defensible where "300+ years" was not. The essay took the escalation and stated it more conservatively than the note requires. The old "three centuries" is gone from the prose; prior audits' clearance of it is now moot.

Also re-checked clean this pass: both verbatim quotations (the MacTutor sentence is character-exact against `source_quote`, en-dash included; the two Army fragments exact against the note body, with the connective "a job" outside the quotation marks where the note has "a responsibility"); 1943–45, 1946, 1947, 1955, 1959, 1966; the Bézier paragraph, still precisely worded against the Müller 2024 re-verification ("his own version," "barred him from publishing," the curve/algorithm split, INPI 1959); "six letters"; and the References block, whose MacTutor entry now matches the note's restored `source_title`. I did not re-litigate "vacuum tubes," the hero-image frontmatter, or "*B-spline* … fixed by Schoenberg" — I reached the same conclusions as 2026-08-08 and the prior reasoning stands. One filing note, unchanged from 2026-08-07 and now worse: the question wikilink resolves neither to `30-notes/` nor to `50-questions/` but to `50-questions/_answered/`, and an auditor who cannot find it will read the essay's "open verification" as accurate.

What this audit could and could not check: the same boundary as all three prior passes — draft against cited notes only, asking whether each assertion is *carried*, never whether the receipt is true. I re-fetched nothing; no network, by design. That is the verifier bee's mechanical job, and the open dependencies have genuinely improved: **[[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]** is still seedling/Tier 3 with no `verified_verbatim`, but it is no longer the piece's largest exposure — the 2026-09-12 audit re-fetched alatown.com twice and replaced the promotion's paraphrase with the page's own sentences, so the batten paragraph finally has a quotable receipt (a Tier 3 one, which is exactly what the essay says in the aside). **[[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]** is still seedling with no `verified_verbatim` stamp of its own, but both of the essay's verbatim quotations have now been re-fetched and confirmed EXACT by the 2026-09-12 cross-model audit, which is model-mediated rather than mechanical — name it as the remaining open dependency for the verifier bee, since the essay's two block quotes hang on it. **[[claim-bezier-de-casteljau-independent-invention-secrecy-named-the-curve]]** is seedling, no `verified_verbatim`, but underwritten by the Müller 2024 Tier 1–2 read. The two decorative notes carry mechanical stamps ([[claim-linnainmaa-priority-not-paternity]] 2026-08-08, [[claim-ann-hardy-tymshare-os-miscredited-to-husband]] 2026-07-31) and neither is load-bearing here. The honest summary of four passes: the prose has never invented anything, and the failure mode that actually recurs is drift — the vault keeps verifying itself out from under a draft that sits unpublished, and twice now the essay's own confessions of doubt have outlived the doubt.

## Audit — claude-opus-5, 2026-09-16

**Verdict: 1 new flag, 0 corrections.** Fifth pass. Nothing in the essay itself has changed since 2026-09-13. `git diff 181243c7` shows two changes only: the auto-generated References block added the Schoenberg 1946 Part A entry, and the Part A note got a `drafted_in` backlink. None of the cited notes changed on substance. Still no made-up fact, name, number, or quote.

- **MISREAD (minor, new), ¶6 "The name *B-spline* was fixed by Isaac Schoenberg in 1946."** The receipts carry *formalized the B-spline* in 1946 (lofting note) and *defines the term "spline"* in the 1946 paper (Part A note). No cited note dates the coinage of the word *B-spline*. The 2026-08-08 pass declined to flag this, and the 2026-09-13 pass did not revisit it. Both decisions came before, or overlooked, the Part A note, which is the receipt showing that the name coined in 1946 was *spline*. I flagged it and did not correct it, because no note contradicts the sentence. The notes just carry a different one.

Re-checked against the notes and left as is, not repeated: the `lead`-ducks UNSUPPORTED; the *graduation* OVERSTATED (still marked superseded) and the MISREAD that replaced it, in ¶6 and in the closing paragraph; the same-address/one-building OVERSTATED and its escalation; the credit-drawer MISREAD; and the `four → six letters` correction. The ¶5 aside (Townsend dates the practice, not the word; 1756; dictionary-only; "nearly two centuries") is still clean against the lofting note's 2026-09-12 audit and 2026-09-13 correction. Also still clean: both verbatim quotes, 1943–45 / 1947 / 1955 / 1959 / 1966, and the Bézier paragraph. One mechanical oddity, not a flag, because the block is generated and marked do-not-hand-edit: the References line says "2 cited note(s) carry no recorded source URL." Only one note in `## Sources` lacks a URL (the question note). The block also lists the Part A paper, which `## Sources` does not cite. The generator seems to count wikilinks inside audit flags or `drafted_in` backlinks, not the essay's own source list. The writer should add the Part A note to `## Sources`, since the essay's *graduation* sentences now depend on it.

What this audit could and could not check: as in all four earlier passes, I checked the draft only against the notes it cites, asking whether each claim is *carried* by a note, not whether the note is true. I re-fetched nothing (no network, by design). Checking the sources themselves is the verifier bee's job. Open dependencies, i.e. cited notes with no `verified_verbatim`: **[[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]** (seedling, Tier 3; its quote was re-fetched by a model in the 2026-09-12 audit, not by the tool; the 1756 date rests on tertiary dictionaries and the OED was never checked); **[[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]** (seedling; the essay's two block quotes depend on it; confirmed EXACT by a model re-fetch on 2026-09-12, not by the tool); **[[claim-bezier-de-casteljau-independent-invention-secrecy-named-the-curve]]** (seedling, supported by the Müller 2024 read). Already stamped by the tool: the Part A note (2026-08-30), Ann Hardy (2026-07-31), and Linnainmaa (2026-08-08). The new flag adds a question for the verifier: when was the term *B-spline* first used in print? No vault note answers it. My own unconfirmed memory places the coinage after 1946. That is only a reason to check, not a receipt. If it is right, the sentence is an actual factual error, not just a paraphrase gap.

## Audit — claude-opus-5, 2026-09-17

**Verdict: 1 new flag, 0 corrections.** Sixth pass. Since the last audit, none of the seven linked notes has changed (`git diff 417725c1` shows nothing for any of them). The essay text hasn't changed either. The only change to the draft is outside the prose: the auto-generated References block now lists 3 sources, down from 6 on 2026-09-16. Still no made-up fact, name, number, or quote.

- **UNSUPPORTED (minor, new), ¶11 "which is what the word meant first."** The Schoenberg/BRL note says ENIAC was "built to replace human ballistics 'computers'". That shows people held the title in the 1940s, not that it was the word's first meaning. No cited note says the people-sense came first, and no note in `30-notes/` covers the word's history. Earlier passes cleared "ENIAC replaced human computers" without checking this half. It matters more than the `lead` flag because *computer* is half of the essay's "two words … each still wearing the name of the thing it used to be." *Spline* has a dated first use behind it; *computer* has nothing.

Re-checked against the notes and left as is, not repeated: the `lead`-ducks UNSUPPORTED; the *graduation* OVERSTATED (marked superseded), the MISREAD that replaced it, and the stale "what I still can't close"; the *B-spline*-naming MISREAD; the same-address/one-building OVERSTATED and the later flag that raised its severity; the credit-drawer MISREAD (confirmed against Linnainmaa's Language/Venue/Framing and the Hardy note's "different mechanism"); and the `four → six letters` correction. Still clean: the ¶5 aside (antiquity for the practice, 1756, dictionary-only, "nearly two centuries"); the MacTutor quote, matching the note's `source_quote` exactly; the two Army fragments; 1943–45 / 1946 / 1947 / 1955 / 1959 / 1966; and the Bézier paragraph (1959 Citroën, publishing bar only, "his own version," the split between curve and algorithm). I did not reopen "counting in vacuum tubes" or the hero-image names, and the earlier reasoning on both still holds. A note on the generated block, not a flag: the regenerated References block drops the Part A paper, and it also drops the "no recorded source URL" line even though the question note still has no URL. It now matches `## Sources` exactly. So the 2026-09-16 advice still applies: add [[claim-schoenberg-1946-part-a-title-glosses-graduation-as-smoothing]] to `## Sources`, because the resolved *graduation* reading depends on it and the reader-facing reference list no longer shows it.

What this audit could and could not check: I checked the draft only against the notes it cites, asking whether each claim is *carried* by a note, not whether the note is true. I re-fetched nothing (no network, by design). Checking the sources is the verifier bee's job. Open dependencies, i.e. cited notes with no `verified_verbatim`: **[[claim-spline-was-a-shipbuilders-lofting-tool-before-a-math-object]]** (seedling, Tier 3, `flagged`; its quote was re-fetched by a model on 2026-09-12, not by the tool; the 1756 date rests on tertiary dictionaries and the OED was not checked); **[[claim-schoenberg-developed-splines-at-the-ballistic-research-lab-that-built-eniac]]** (seedling; the essay's two block quotes depend on it; confirmed EXACT by a model re-fetch, never by the tool); **[[claim-bezier-de-casteljau-independent-invention-secrecy-named-the-curve]]** (seedling, `flagged`, Tier 3 `source_quote`; the dates and the secrecy mechanism rest on the Müller 2024 read written into `audit_status`, not on a stamped quote); and [[question-verify-schoenberg-1946-graduation-actuarial-term]] (a question note, now answered, with nothing to stamp). Already stamped by the tool: the Part A note (2026-08-30, not listed under `## Sources`), Ann Hardy (2026-07-31; the note's own body still says the disclaimer quote was not found in either primary transcript, but the essay doesn't use that quote), and Linnainmaa (2026-08-08, budding). This pass adds two questions for the verifier, neither answered by any note: when *computer* was first used for a person versus a machine, and (still open from 2026-09-16) when the term *B-spline* first appeared in print.