The curve math under every digital font traces through car-company secrecy and ship lofting to the WWII lab that built ENIAC

Claims

1. The Bézier curve that defines every PostScript/TrueType font outline was invented independently and near-simultaneously at two rival French carmakers, and corporate secrecy — not mathematical priority — decided whose name it got. Pierre Bézier developed his curve method at Renault (published 1966); Paul de Casteljau solved the same problem earlier at Citroën (1959) but the company barred him from publishing or patenting it, so credit defaulted to Bézier. The algorithm is now named for de Casteljau anyway, the curve for Bézier. source_tier: 3 (converging independent secondary/design-history accounts; historical claim, uncontested).

2. "Spline" is not originally a math word — it's a shipbuilder's tool. Before computation, a spline was a flexible wood or steel strip bent across a hull mold and pinned by weighted "ducks" to trace a fair curve by hand (naval lofting). Isaac Schoenberg's 1946 papers that formalized the B-spline used the word "graduation" for curve-smoothing — actuarial terminology for smoothing mortality tables — which is why B-splines are often (mis)attributed to insurance work. [unverified-quant/mechanism -- needs primary] on the actuarial-term claim specifically; source_tier: 3 (design-history blog, historical/craft claim).

3. Schoenberg wasn't doing actuarial work when he invented the B-spline — he was a WWII Army mathematician, at the same lab that built ENIAC. MacTutor, 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." That same lab, per the US Army's own ENIAC history: "The world's first electronic digital computer was developed by Army Ordnance to compute World War II ballistic firing tables," and "this responsibility was carried out at the Ballistic Research Laboratory of the Ordnance Department at Aberdeen." ENIAC itself was later moved to and operated at BRL, Aberdeen (1947–1955). source_tier: 2 (MacTutor academic biography; US Army archival technical-history page).

Why this was hop-worthy

Five domains that don't normally share a sentence — printing-press economics, 1960s automotive CAD, wooden shipbuilding, WWII artillery math, and the birth of general-purpose computing — turn out to be one lineage: the same curve-fitting math traveled from ship hulls to insurance-table language to bombs to car doors to font outlines. vault_novelty's closest matches for the final synthesis were scattered across completely unrelated existing clusters (blind mathematicians' private notations, an artificial-life simulator, Braille/Barbier, early computing hardware) — no single cluster owns this, which is exactly the bridge signal the protocol flags as highest-value.

Further leads

• NURBS (Boeing, 1979–81) unified Bézier curves and B-splines for aerospace manufacturing before Pixar's subdivision-surface work (Catmull, Geri's Game, 1997) replaced NURBS for organic animation — a further zoom-out not pursued this chain (max_cosine 0.612, saved).
• Alois Auer, Austrian punchcutter credited with ~110,000 forms across his career — a "person behind the thing" hook from the typefounding-economics leg, not pursued (max_cosine 0.596).
• Don Knuth's METAFONT (1977) required designers to write equations instead of drawing curves — Knuth himself called this "too much" to ask of an artist; a mechanism hook adjacent to this chain but not followed (max_cosine 0.643).

Hop chain

Hop 1: "A Brief History of Type" (Simon Cozens) — https://simoncozens.github.io/fonts-and-layout/history.html

• Hook type: Cross-domain bridge
• Hook: Bézier curves, core to PostScript/TrueType outlines, were born at Renault and Citroën as automotive CAD tools, not as a type-design invention
• Why followed: highest-priority hook type per protocol; also a strong zoom-in from the seed's general typefounding-economics framing to a specific mechanism
• Key findings: Pierre Bézier (Renault, published 1966) and Paul de Casteljau (Citroën, developed 1959) solved the same curve-fitting problem independently; Citroën's confidentiality policy suppressed de Casteljau's publication for years, so Bézier's name attached to the curve while de Casteljau's attached only to the underlying algorithm

Hop 2: "On the Spline: A Brief History of the Computational Curve" (Alastair Townsend) — https://alatown.com/spline-history-architecture/

• Hook type: Cross-time-period bridge (a special case of cross-domain bridge, per protocol's extra weighting)
• Hook: "spline" pre-dates computing by centuries as a physical shipbuilding lofting tool (flexible strip + weighted "ducks")
• Why followed: zoom-out from the specific 1960s algorithm to its much older physical/craft root — direction alternation after hop 1's zoom-in
• Key findings: naval lofting → French curves (Burmester, ~1900) → Citroën/de Casteljau and Renault/Bézier (1959–66) → B-splines (Carl de Boor at GM, adapting a 1940s function) → NURBS (Boeing, 1979–81) → Pixar subdivision surfaces (1997); the article names the 1940s B-spline origin as tied to "actuarial graphs"

Hop 3: Isaac Schoenberg biography search + "B-spline" term origin — https://mathshistory.st-andrews.ac.uk/Biographies/Schoenberg/ (via search)

• Hook type: The unfamiliar name
• Hook: Isaac Schoenberg coined "B-spline" in 1946; his papers use "graduation," the actuarial word for smoothing mortality data
• Why followed: zoom-in on the specific person/term behind the B-spline, continuing to test the "actuarial origin" claim from hop 2
• Key findings: Schoenberg's two-part 1946 paper, "Contributions to the problem of approximation of equidistant data by analytic functions," addressed "smoothing or graduation" of data — but "graduation" turned out to be a red herring pointing at the wrong industry (see hop 4)

Hop 4: MacTutor biography of Isaac Schoenberg — https://mathshistory.st-andrews.ac.uk/Biographies/Schoenberg/

• Hook type: The surprising claim (a correction of the popular "actuarial" attribution)
• Hook: Schoenberg developed spline theory not in insurance/actuarial work but as a WWII Army mathematician at the Ballistic Research Laboratory, Aberdeen Proving Ground (1943–45)
• Why followed: per sources.md, a surprising/contested biographical claim escalates the sourcing floor to Tier 1–2; MacTutor cleared that bar and directly falsified the actuarial-graphs framing carried by the (Tier 3) design-history blog in hop 2
• Key findings: "It was during this war work that he initiated the area for which he is most famous, the theory of splines" (MacTutor, verbatim) — spline theory's true institutional origin is a WWII artillery-math lab, not an insurance office

Hop 5: Ballistic Research Laboratory / ENIAC history — https://ftp.arl.army.mil/~mike/comphist/eniac-story.html and https://en.wikipedia.org/wiki/Ballistic_Research_Laboratory

• Hook type: Cross-domain bridge
• Hook: the exact lab where Schoenberg worked (BRL, Aberdeen) is the same lab that commissioned ENIAC to replace human ballistics computers, and later operated ENIAC on-site
• Why followed: zoom-out from Schoenberg-the-person to the institution's other, far more famous output — closing the chain on a second cross-domain bridge (typography's curve math and the literal birthplace of general-purpose computing share an address)
• Key findings: "The world's first electronic digital computer was developed by Army Ordnance to compute World War II ballistic firing tables" (US Army archival source, verbatim); ENIAC was moved to BRL Aberdeen in 1947 and operated there until 1955

Saved hooks not followed:

• NURBS / Boeing (1979–81) and Pixar subdivision surfaces (1997) — from Hop 2 — a further zoom-in mechanism thread on curve math's later industrial history, interesting but the chain had already found its cross-domain destination
• Alois Auer, Austrian punchcutter (~110,000 forms) — from the original seed search (TypeDrawers thread) — a "person behind the thing" hook in the pre-digital typefounding-labor economics story itself, orthogonal to the curve-math thread actually followed
• Don Knuth's METAFONT (1977) and his own admission that requiring 60-parameter equations was "too much" to ask an artist — from Hop 1 — a mechanism hook (math notation vs. artistic drawing) adjacent but not pursued

post-worthy: yes — a clean five-hop chain with two genuine cross-domain bridges (automotive/typography, ballistics/computing), a falsified popular claim corrected against a Tier 2 source, and a final synthesis that bridges multiple disconnected existing vault clusters rather than sitting in just one.