Pattern and Follower: From Nuremberg to today

Pattern and Follower: From Nuremberg to CNC

This is the last post in a series that has traced a single idea across four centuries. The idea is simple enough to state in one sentence: a shaped pattern controls the movement of a cutter, and the cutter transfers the pattern’s form to the workpiece.

That’s it. That’s the entire operating principle of every ripple molding machine ever built — from Johann Schwanhardt’s first device in Rothenburg ob der Tauber around 1600 to the CNC machines cutting organic waveforms in workshops today. The materials have changed. The power source has changed. The precision has changed by orders of magnitude. But the logic has not changed at all.

Jonathan Thornton, in his 2002 paper on waveform moldings, offered a thesis that deserves to be quoted in its most compressed form: what happened to devices that created waveform moldings can be thought of as a capsule history of woodworking. As machines have grown in complexity, the necessary skills of the operator have declined.

This post is about that capsule history — and about where the intelligence goes when it leaves the craftsman’s hands.

The Lineage

The pattern-and-follower concept did not begin with woodworking. Thornton traces it back to around 1480, to screw-cutting lathes where a carved helical guide controlled the cutting of threads. A follower riding along the helical template transferred the spiral pattern to the workpiece with each pass of the cutting tool.

From screw-cutting, the principle migrated in several directions simultaneously:

Ornamental turning. Jacques Besson, Da Vinci’s successor as engineer to the French Court, designed complex ornamental lathes using pattern-and-follower systems in 1579. The rose engine — a lathe with an offset cam that produces complex geometric patterns — uses the same principle. Ornamental turning became a gentleman’s hobby in the 18th and 19th centuries, producing objects of breathtaking geometric complexity from a fundamentally simple mechanical idea.

Gun rifling. The cutting of spiral grooves inside gun barrels, practiced since the late 15th century, used a spiral guide rod to control the rotation of the cutting head as it was drawn through the barrel. This is pattern-and-follower in its most literal form: a physical template dictating the path of a cutter.

Waveform molding. Schwanhardt’s leap — applying the rifling principle to flat wooden surfaces — opened the door to an entire family of machines. The Kaseman wriggling plane (c. 1630) used guide strips to force a plane into lateral oscillation. Moxon’s Waving Engine (1678) used interchangeable template rods to produce both lateral and vertical waveforms. Félibien’s, Diderot’s, and Roubo’s factory machines (1676–1775) mechanized the feed and pressure while retaining the template-and-follower core.

The Jacquard loom (1804). Joseph Marie Jacquard’s programmable loom used punched cards to control the raising and lowering of individual warp threads, creating complex woven patterns automatically. This is pattern-and-follower at a higher level of abstraction: the pattern is encoded not as a physical shape but as a sequence of holes in a card, and the follower is a mechanical reader that translates those holes into physical action. The punched card doesn’t look like the finished fabric — it’s a coded representation. But the logic is identical: a stored pattern controls the movement of a tool (in this case, the heddle that lifts each warp thread).

Babbage and early computing. Charles Babbage explicitly cited the Jacquard loom as inspiration for his Analytical Engine (conceived in the 1830s). Ada Lovelace, writing about the Engine in 1843, drew the analogy directly: the Analytical Engine weaves algebraic patterns just as the Jacquard loom weaves flowers and leaves. The concept of a stored program — instructions that control the behavior of a machine — is the pattern-and-follower principle taken to its logical extreme, where the “pattern” is pure information and the “follower” is a general-purpose computing mechanism.

CNC machining. Numerical control, developed in the late 1940s and 1950s at MIT, brought the story full circle back to the workshop. A CNC machine uses stored digital instructions (G-code) to control the movement of a cutting tool relative to a workpiece. The G-code is the pattern. The servo motors and controller are the follower. The relationship is mechanically identical to Schwanhardt’s template rod controlling the movement of a scraper blade — except that the pattern exists as data rather than as a carved physical object, and the precision of the follower is measured in ten-thousandths of an inch rather than whatever a craftsman’s hand could achieve.

Where the Intelligence Lives

This lineage — from carved helical guides to punched cards to digital code — traces a migration of intelligence from one medium to another. And with each migration, something is gained and something is lost.

In Schwanhardt’s hands (c. 1600): The intelligence lived entirely in the craftsman. He carved the template, ground the cutter, built the device, operated it, and judged the results. Every decision — wavelength, amplitude, depth, profile, feed rate, material selection — resided in one person’s accumulated knowledge and physical skill. The template was an extension of his judgment, carved by his hands to embody his aesthetic decisions. If the template wore out, he carved a new one, adjusting the pattern based on his evolving understanding of what worked.

In Moxon’s template rods (1678): The design intelligence was partially externalized — encoded in the physical shape of the template — but the execution intelligence still lived in the operator’s hands. The template said what wave to cut. The operator’s skill determined how well it was cut. A bad operator using a good template produced bad moldings. Moxon knew this, which is why he acknowledged that the craft “cannot be taught by words.”

In Diderot’s factory machines (c. 1760): The execution intelligence was transferred to the machine — springs, screws, gears, and crank mechanisms that controlled what the operator’s hands had previously controlled. The design intelligence still lived in the template, which still had to be carved by a skilled hand. But the gap between a skilled operator and an unskilled one narrowed dramatically. The machine compensated for the operator’s limitations.

In Jacquard’s punched cards (1804): The design intelligence was abstracted from physical shape to coded representation. The punched card didn’t look like the pattern it produced. A human designer still made the aesthetic decisions, but those decisions were encoded in a medium (holes in cards) that had no visual relationship to the finished product. This was a conceptual revolution: the pattern and the thing-it-looks-like were separated for the first time.

In CNC cam profiles (present): The design intelligence lives in software — in mathematical descriptions of curves, in cam profile pairs, in G-code that describes toolpaths. The execution intelligence lives in servo motors, ball screws, and PID controllers that follow programmed paths with micrometer accuracy. The human operator’s role is reduced to setup, monitoring, and quality control. The creative decisions — the aesthetic judgments about what makes a good ripple pattern — still require a human mind. But those decisions are made once, at the design stage, and can then be reproduced indefinitely without degradation.

What’s Gained

The gains of this migration are real and substantial:

Precision. A CNC machine can hold tolerances that no human hand and no wooden template rod could approach. The waveforms are mathematically perfect, consistent from the first inch to the last, repeatable across hundreds of feet of molding.

Durability of patterns. Digital cam profiles don’t wear out. The problem that plagued historical makers — template rods degrading with use, producing increasingly shallow and inconsistent relief — is eliminated. A pattern designed today will produce identical results in a hundred years, assuming the data survives.

Design freedom. CNC control allows patterns that would be impractical or impossible to produce with physical templates. Complex mathematical curves, variable-frequency waveforms, asymmetric profiles, patterns that change gradually along their length — all are trivial to program and execute. The design space is vastly larger than what any fixed template could offer.

Reproducibility. A customer can order more molding in ten years and get an exact match. A museum can commission a reproduction frame and know it will be dimensionally identical to the original. Production runs can be interrupted and resumed without the quality variations that afflicted template-based work.

Speed and volume. CNC machines can produce molding at rates that hand-pulled or hand-cranked machines cannot match, making ripple moldings economically viable for markets that couldn’t support the labor costs of traditional production.

What’s Lost

But the losses are real too, and they matter to anyone who cares about surface quality, historical authenticity, or the particular kind of beauty that comes from the interaction between a skilled hand and a resistant material.

The hand-pulled surface. Thornton and others have documented that hand-cut ripple moldings have a different surface character than machine-cut. The subtle variations in speed and pressure — the fact that no two passes of a hand-pulled slide-board are exactly identical — produce a surface that shimmers under changing light in a way that perfectly uniform CNC-cut ripple does not. This is not mysticism. It is a measurable difference in surface geometry that produces a measurable difference in light behavior.

The apprenticeship relationship. When the intelligence lived in hands and eyes and ears, it was transmitted through the physical presence of a teacher and the extended practice of a student. That transmission created relationships — between master and apprentice, between generations of makers, between communities of practice. Digital knowledge transfer is efficient but lonely.

The constraints that drive creativity. When your design space is limited by what a physical template can produce — when the wavelength is set by the template rod you’ve carved, and changing it means carving a new one — you work within constraints that force focus. You refine within a narrow range rather than ranging freely across an infinite design space. Many makers find that constraints improve their work. The unlimited freedom of CNC can be paralyzing rather than liberating.

The contingency of material. A hand-pulled pass through ebony is a negotiation between the blade and the grain. The wood pushes back. Knots, interlocked grain, density variations — all of these require the operator to respond in real time, adjusting pressure and speed. The finished surface records this negotiation as a kind of texture that is absent from CNC work, where the machine’s power overwhelms the material’s resistance.

The Same Conversation

Here is the thing that makes this four-century story cohere, and the reason it matters to anyone making moldings today rather than just studying them:

Every person who has ever produced a ripple molding — from Schwanhardt in his Rothenburg workshop to a CNC operator in a modern production facility — has been engaged in the same fundamental act. A pattern is controlling the movement of a cutter through wood. The intelligence that governs that act has migrated from medium to medium — from the craftsman’s hands to carved template rods to crank-driven mechanisms to punched cards to digital cam profiles — but the act itself has not changed.

When you run a cam-profile-driven ripple cut on a CNC router, you are doing what Schwanhardt did. You are using a stored pattern to control a cutter’s path through wood. Your pattern is stored as data rather than as a carved wooden rod. Your follower is a servo motor rather than a peg in a track. Your depth control is a stepper motor rather than a hand-turned screw. But the conversation between pattern, cutter, and wood is the same conversation.

This continuity is not sentimental. It is structural. Understanding it — understanding that you are working within a tradition that has solved and re-solved the same fundamental problem for four hundred years — gives you access to the accumulated wisdom of that tradition. The 17th-century makers learned things about wavelength, amplitude, profile shape, and visual rhythm that remain valid regardless of how the molding is produced. Their aesthetic discoveries are not obsolete just because their machines are.

The intelligence migrates. The logic endures. And the wood, as always, has the final word.

 

Sources:

• Jonathan Thornton, “The History and Technology of Waveform Moldings: Reproducing and Using Moxon’s ‘Waving Engine,’” WAG Postprints, 2002. Free PDF at wag-aic.org.
• Joseph Moxon, Mechanick Exercises (1678–80).
• André-Jacob Roubo, L’Art du Menuisier (1769–1775).
• Denis Diderot and Jean le Rond d’Alembert, Encyclopédie (1751–1772).
• 1642 Designs, “About” page (1642designs.com/about).
• Ada Lovelace, “Notes” on the Analytical Engine, in Richard Taylor’s Scientific Memoirs Vol. 3 (1843).
• Kingswood Frames (kingswoodframes.com/ripple-and-wave).