How to Create the Jacquard Mechanism for a Circular Loom?

 

Jacquard looms have long fascinated me. Their cobweb-like form resembles a living organism. The machine is incredibly intricate, with so many moving parts that it's almost incomprehensible how one could create it — let alone maintain it. Yet, believe it or not, its complexity is based on a remarkably simple concept.

the 'box' is the metal part on top

Let's begin with the fact that Mr. Joseph Jacquard, a French weaver, didn’t invent the Jacquard loom. [watch this video before moving on] In 1804, he introduced a major addition to an already existing principle — an extrapolation of the draw loom, a concept developed by the Chinese long before the Common Era. The challenge was to automate and control individual or sets of heddles independently and enabling complex and intricate patterns to be woven faster and more efficiently.

The Jacquard loom is said to have influenced the development of binary thinking. Its use of punch cards to control shedding later inspired early computer programming. Each card carried a set of instructions, and these cards were linked in a chain, forming a train of commands that was read by a mechanical device — let's call it "the box." 

Most punchers —
 were women.
Each card had to
be manually perforated

The box and the chain have direct parallels with an interpreter reading a program's instructions. Think of the time spent writing words line by line, saying them one after another. That’s where the chain comes from. The code becomes a textile. Textile becomes text (as in the woven prayer book, Livre de Prières tissé d'après les enluminures des manuscrits du XIVe au XVI siecle. ). Book becomes a program. Text (from "texere," meaning to weave, braid, or construct) becomes a set of instructions — and so on.

[From The Common Thread: "Some might say that history starts with writing, but in textile research, history starts with clothing, and before history was nakedness."]

 

The box introduces a limitation that persists even now — the sequential bottleneck. Most programming languages still operate on machines rooted in this same mechanism. Efforts like multithreading, concurrency, and async-await attempt to mitigate this flaw, but they still rely on underlying systems designed to process sequences. Even advanced approaches must account for certain system constraints — often described by Amdahl's Law:

"Amdahl's Law states that the maximum speedup an application can achieve is limited by the portions that must execute sequentially."

Hardware imposes fundamental limitations that can’t be fully bypassed. All the abstractions in programming languages and frameworks are just attempts to manage these constraints more efficiently. It’s like building bridges over a river – the river is still there, but we try to make it less of an obstacle.

Another major technical issue — the memory bottleneck it’s again a reminiscence of the loom. This bottleneck is a significant challenge, possibly explaining why demanding tasks like artificial intelligence or gaming require so much computational power. Yet, emerging LLMs (large language models) and generative algorithms offer hope(by helping in development) for integrated memory and computational design, such as Processing-in-Memory (PIM) technologies:

Breaking the von Neumann bottleneck: architecture-level processing-in-memory technology

The Memory Bottleneck: New AI Chip Designs That Overcome von Neumann’s Limits

 Line by line. It's hard to imagine something fundamentally different. Perhaps if the Jacquard machine had been designed for creating baskets or 3D structures from the start — structures that reject sequential thinking — we might have something more powerful.

 One might argue that simplicity leads to complexity and refinement over time — simple rules building it up, with mechanisms correcting errors along the way. And repeat! - but don’t even try to DRY it...

What if we mimicked the way bacteria — a seemingly simple organism — operates, where everything happens simultaneously: reading, translating, and replicating DNA sequences with various enzymes?

Factors Behind Junk DNA in Bacteria

The dark side of the genome - does it matter?

Genes are not "code" or "instructions", and creationists oversimplify biology by claiming that they are.

 

Tadek Beutlich: On and Off the Loom,
installation view,
Ditchling Museum of Art and Craft,
 18 January – 22 June 2025. Photo: Phoebe Wingrove.

What would happen if we adopted principles from beehives or ant colonies — systems in which individual units collaborate to create complex outputs?

Here is a  different approach different to ‘sequential’ in the field of textiles:

Mr. Toyoda (founder of Toyota) designed a circular loom, where curved shuttle (imagine multiple instead of one) — move in circular motion to create cloth resembling a big sleeve.

[VIDEO of the working loom]

 

What if the Jacquard mechanism were designed for a machine like this? What would replace punch cards? Could we eliminate the sequential square-based logic altogether?

The circle defies the square. At the very least, it marks a major step toward moving beyond the limitations of sequential thinking. The movement is no longer back and forth, side to side — sets of dots after another set — but something more spiral-like, fluid, and constant.

Perhaps the real question is whether we need the division between the interpreting body and the mechanism performing it. How do we blur that line? How do we create a system where interpretation and execution are one and the same — where maybe a motion itself is the key? The swing, the breath (the amplitude) — these rhythmic movements might hold the answer.

How do we create the Jacquard mechanism for a circular loom?  I don't know, but "we cannot solve our problems with the same thinking we used when we created them".