Konrad Zuse Z1
Konrad Zuse was a German civil engineer, not a mathematician or an academic, and he built the Z1 alone, working nights and weekends in his parents' living room in Berlin, with no university backing, no government funding, and no knowledge of Charles Babbage's nineteenth-century designs or Claude Shannon's 1937 thesis on relay logic. Zuse had grown frustrated with the repetitive arithmetic required by structural engineering calculations at his day job and concluded, independently of everyone else working on calculating machines at the time, that the fastest way to eliminate that drudgery was to build a general-purpose machine that could follow a program rather than be rewired for every new calculation. He started designing in 1936 and finished a working machine in 1938, financed out of his own pocket and with help from friends who assembled parts in the family apartment, a detail that is easy to read past but is genuinely remarkable: the Z1 was built from scratch, in cramped domestic quarters, by one person who had reasoned his way to binary programmability from first principles.
Where Hollerith's tabulator and the relay logic Shannon formalized used electrical circuits to represent binary states, the Z1 was purely mechanical: it had no vacuum tubes and no relays, and instead represented binary 0 and 1 as the positions of thin sliding metal plates, with pins and slots that physically blocked or permitted motion depending on the state of other plates, all of it driven by a single electric motor that supplied mechanical power rather than electrical signaling. This made the Z1 the series' first machine to implement genuine binary logic gates, AND, OR, and NOT equivalents, entirely in mechanical hardware, arriving at the same binary insight Shannon had just formalized in relay-switch terms, but through an entirely different engineering path and without either man aware of the other's work. The sliding-plate approach also meant Zuse's binary logic was at the mercy of 1930s manufacturing tolerances: metal plates that needed to slide freely and register precisely were hard to machine to the necessary accuracy with the tools and budget available to one civil engineer, and the Z1 was consequently slow and notoriously unreliable, prone to jamming and misregistration under its own mechanical imprecision.
The Z1 used binary floating-point arithmetic, representing numbers with a sign, a mantissa, and an exponent in base two, a design choice decades ahead of the fixed-point decimal arithmetic most calculating machines of the era used, built into a roughly 22-bit word length with a mechanical memory store holding 64 such words. Programs were not entered by rewiring or by hand-set switches for every step; Zuse read instructions from a punched tape, and in a detail that captures the era's wartime scarcity as well as his improvisational engineering, the tape he used was discarded 35mm movie film stock, with binary instruction holes punched directly into the film's surface in place of the sprocket images it once carried. This gave the Z1 a second, independent link back to the series' punched-tape lineage that started with the Jacquard loom and ran through Hollerith's cards: the same physical idea, holes in a strip encoding instructions, but now the holes encoded binary machine instructions for a programmable processor rather than card-sorting patterns or loom-weaving commands.
The Z1 did not survive the decade: it was destroyed in 1943 in an Allied bombing raid on Berlin, along with Zuse's later machines and most of his original design documentation, so no original hardware or drawings of the Z1 remain today. What exists instead is a faithful modern reconstruction that Zuse himself built between 1986 and 1989, working from memory and from the surviving fragments of his notes, now on permanent display at the Deutsches Technikmuseum Berlin, the closest thing left to seeing the sliding-plate mechanism actually move. Zuse went on after the war to found one of the earliest computer companies and to develop further machines, the Z2 and Z3, improving on the Z1's reliability with electromechanical relays rather than pure mechanics, but the Z1 remains the foundational claim: the first machine built anywhere to be freely programmable and to compute in binary, arrived at by one engineer working alone, independently and almost simultaneously with Shannon's theoretical formalization of binary switching logic on the other side of the Atlantic. The next step in this series moves to the United States, where between 1937 and 1942, at Iowa State University, the Atanasoff-Berry Computer took the binary idea Zuse proved mechanically and Shannon proved theoretically and implemented it for the first time in electronic vacuum-tube hardware, with memory function separated from computing function for the first time as well.