| | {{infobox |
| | | title = Jacquard loom |
| | | image = jacquard-loom.jpg |
| | | caption = A Jacquard loom — the chained card deck rides above the weaving head |
| | | invented = 1801 (demonstrated), patented 1804 |
| | | inventor = [[Joseph-Marie Jacquard]] |
| | | built_on = Bouchon, Falcon, and Vaucanson mechanisms |
| | | program_medium = Chained [[Punched card|punched cards]] |
| | | famous_output = Jacquard's woven portrait — ~24,000 cards |
| | | significance = First machine run by a stored, swappable program |
| | | legacy = Direct ancestor of the program deck and the computer |
| | }} |
| | |
| | The **Jacquard loom** is a weaving machine invented by **[[Joseph-Marie Jacquard]]** of Lyon, France, first demonstrated in **1801** and patented in **1804**. It wove complex figured patterns automatically by reading instructions from chains of **[[Punched card|punched cards]]** — making it the first practical machine controlled by a stored pattern of instructions, and the ancestor of every idea of a "program" that followed. |
| | |
| | ## Before Jacquard: the long road |
| | |
| | Jacquard's breakthrough was the last step of an eighty-year effort to mechanize pattern weaving: |
| | |
| | - **1725 — Basile Bouchon** replaced the draw boy's cord-pulling with a perforated paper roll that pressed against needles — the first automatic selection, but the paper tore too easily for real use. |
| | - **1728 — Jean-Baptiste Falcon**, Bouchon's assistant, swapped the roll for individual **punched cards** chained in a loop — the right idea, still too delicate. |
| | - **1745–1750 — Jacques de Vaucanson**, the famous automaton-maker, tried a punched metal cylinder — precise, expensive, and never adopted by the trade. |
| | |
| | Each inventor moved the idea one notch toward practicality. Jacquard — a working weaver, not a gentleman-scientist — combined Falcon's cards with a rugged needle-and-hook mechanism that any workshop could build, maintain, and afford. That is the version that spread. |
| | |
| | ## The problem it solved |
| | |
| | An ordinary loom interlaces threads by lifting some warp threads and not others on each pass of the shuttle — the *pattern* is which threads go up. Plain cloth needs simple repeating lifts, but figured fabric — brocade, damask, anything with an image or intricate design — needs a *different* selection on every pass. |
| | |
| | Before Jacquard, that selection was a person. A **draw boy** sat above or beside the loom and pulled cords to raise the correct threads, pass after pass — thousands of selections a day, each made by hand. One mistake spoiled yards of silk; one skilled pattern required weeks of labor for feet of cloth. The pattern existed only in the weaver's head and the draw boy's hands — it could not be stored, copied, or moved. |
| | |
| | ## How it worked |
| | |
| | [[File:jacquard-loom-1891.jpg|right|300px|A carpet loom with Jacquard attachment — 1891 engraving. The card chain and reading head sit above the warp]] |
| | |
| | Jacquard's attachment — mounted above an ordinary loom — replaced the draw boy with punched cards: |
| | |
| | - Each **card** carried one row of the pattern, encoded as a grid of holes. |
| | - A bank of spring-loaded **needles** pressed against the card each pass. Where a hole was punched, the needle passed through; where there was no hole, the needle pushed its **hook** out of engagement — and that warp thread stayed down (or rose, depending on the loom's rigging). |
| | - Cards were **chained together** into a continuous loop, advanced one card per shuttle pass. |
| | |
| | So: hole = this thread lifts; no hole = it doesn't. One binary decision per thread per row, thousands of decisions per card, chained into sequences long enough to weave pictures. |
| | |
| | [[File:jacquard-cards.jpg|left|260px|Chained Jacquard cards — each card is one pass of the shuttle; the loop is the program]] |
| | |
| | The famous woven portrait of Jacquard — produced by the Lyon firm Didier, Petit et Cie — required about **24,000 cards**, and its fidelity still startles people who see it. |
| | |
| | ## Why it is the ancestor of programming |
| | |
| | The loom did three things no earlier machine had: |
| | |
| | - **The pattern was data.** The same loom wove any design; you changed the deck of cards, not the machine. Hardware and its "code" were finally separate things. |
| | - **Programs were portable.** A card deck was a library — one atelier's pattern could be punched, sold, shipped, and woven on any Jacquard loom. Patterns became a tradeable commodity in themselves. |
| | - **Errors were in the cards, not the machine.** Punch a hole wrong and the cloth came out wrong in exactly that spot — the world's first *bugs*, traceable to a specific instruction. |
| | |
| | Weavers even developed **card-punching as a skilled trade**: a designer's draft was transferred to squared pattern paper, then punched row by row — the direct ancestor of "coding from a spec." Lyon developed a whole cottage industry of punchers and card-lacing specialists. |
| | |
| | ## Controversy |
| | |
| | The loom was not welcomed. Lyon was the center of French silk weaving, and the draw boys and journeymen whose work the cards automated attacked the early machines — looms were smashed and Jacquard was reportedly dragged from his loom and thrown into the Rhône. The technology won anyway: by **1812** there were some **11,000 Jacquard looms** in France, and the riots became one of history's first recorded episodes of labor unrest against automation — decades before the Luddites made the pattern famous in England. |
| | |
| | The economics were unstoppable: figured silk that took weeks now took days, at a fraction of the cost, and the same design could be reproduced identically a hundred times. Lyon's output boomed; the draw-boy trade disappeared. The argument — does a machine that makes goods cheaper destroy the people who made them — is the same one industries have repeated ever since. |
| | |
| | ## Lineage to computing |
| | |
| | [[File:jacquard-woven-portrait.jpg|right|220px|The woven silk portrait of Jacquard himself, ~1839 — roughly 24,000 cards of "program"]] |
| | |
| | The loom's cards passed the idea of *instructions on paper* directly into computing: |
| | |
| | - **[[Charles Babbage]]** explicitly adopted Jacquard-style punched cards for his **[[Analytical Engine]]** (1830s) — its operation and variable cards were program and data, and [[Ada Lovelace]]'s 1843 notes assumed that model. Babbage owned a copy of the woven portrait and loved showing it to visitors as proof of what cards could encode. |
| | - **[[Herman Hollerith]]**'s 1890 census tabulator used punched cards for data processing, and the company that grew from it — **[[IBM]]** — fixed the card as an industry standard for nearly a century. |
| | - Program and data punched cards remained the normal way to *store* a program until interactive terminals arrived in the 1970s — a direct line of descent a century and a half long. |
| | |
| | When computing historians say the Jacquard loom "invented programming," they mean it separated the **machine** from the **sequence of operations** it performs — the single idea every language since is built on. |
| | |
| | ## See also |
| | |
| | - [[Joseph-Marie Jacquard]] |
| | - [[Punched card]] |
| | - [[Charles Babbage]] · [[Analytical Engine]] · [[Ada Lovelace]] |
| | - [[Herman Hollerith]] · [[IBM]] |
| | - [[History of Programming]] |
| | |
| + | ## References |
| + | |
| + | 1. Essinger, James. *Jacquard's Web: How a Hand-Loom Led to the Birth of the Information Age* (Oxford UP, 2004). |
| + | 2. Bowness, Charles. "The Jacquard Machine." *Proceedings of the Institution of Mechanical Engineers* (1880). |
| + | |
| | [[Category:History]] |
| | [[Category:Hardware]] |
| | |