Tx-0
TX-0 (the Transistorized Experimental computer zero) was a computer built at MIT's Lincoln Laboratory and running by 1956, often described as the first large, fully transistorized general-purpose computer. It used 18-bit words and a magnetic core memory, and its logic was built from roughly 3,600 Philco surface-barrier transistors, a figure usually quoted in general terms. It was designed as a test vehicle: if transistor circuits and core memory could be made reliable at this scale, the same technology could be used to build a much larger machine, and that follow-on, the TX-2, was the real goal. In the end the little prototype became more famous than the machine it was meant to test, because of what people did with it after it left Lincoln Laboratory.
The design was led by Wesley (Wes) Clark, with Ken Olsen responsible for much of the transistor circuit engineering and for the memory test computer that preceded it. Clark's idea was to prove a new way of building computers rather than a new architecture: standard circuit modules, built around the then-new surface-barrier transistor, that could be stamped out repeatedly and replaced when they failed. The surface-barrier device, developed at Philco, was fast for its day and made it possible to operate a logic cycle in a few microseconds, which a germanium alloy junction transistor of the period could not easily match. The machine itself was spare. Its instruction set was tiny, with only a handful of basic operations, and it had a cathode-ray tube display, a light pen and a paper-tape reader, which together gave it unusually direct human interaction for a 1950s computer.
The TX-0 acquired its second life in 1958, when it went on long-term loan to the MIT Research Laboratory of Electronics, and it was there that it became an object of enthusiasm rather than a research instrument. Because it was small and not heavily scheduled, students and young researchers could sign up for blocks of time and use it directly, rather than handing a card deck to an operator and waiting for printouts, as was normal on the big mainframes of the day. Members of MIT's Tech Model Railroad Club were among those who gravitated to it, and the vocabulary and habits that later became known as hacker culture, including the delight in clever programs and the idea that the machine was something to be explored, are often traced to this setting. Programs written for it included early games and music playback, and its display made small graphical experiments possible. Spacewar, written in the early 1960s, ran on its commercial descendant, the PDP-1, but the people and the programming habits behind it grew up around the TX-0.
Its link to the computer industry runs through Ken Olsen, who left Lincoln Laboratory in 1957 to found Digital Equipment Corporation (DEC) with Harlan Anderson. DEC began by selling small plug-in transistor logic modules, and its first computer, the PDP-1, appeared at the end of the decade and was strongly influenced by Clark's and Olsen's TX-0 work, being a small 18-bit machine intended for direct use by engineers and scientists. That reframed what a computer could be: not an expensive centrally run mainframe but an interactive tool for a person at a console, which is the line of thinking that leads through the minicomputer to the personal computer. The TX-2, the larger sibling, extended the same ideas with a bigger memory and was used for early interactive graphics work.
A common misconception is that the TX-0 was the first transistor computer; it was not. Experimental transistor computers had already appeared, including a prototype at the University of Manchester in 1953 and the Bell Labs TRADIC machine built for the US Air Force, which ran in 1954. What makes the TX-0 stand out is its scale and its engineering for dependable operation using large numbers of transistors and core memory, not an absolute first. The word fully also deserves care: it describes the logic and memory circuitry, and the machine still relied on conventional peripheral equipment such as paper tape and a keyboard printer.
The technology it tested is no longer in use: surface-barrier transistors were eclipsed by diffused silicon devices and then by integrated circuits, and the TX-0's few thousand transistors would now fit in a tiny fraction of one microprocessor. What survives is the pattern. A research prototype demonstrates that a component technology can be trusted in volume, a spin-off company turns the idea into a product, and a community of users discovers what the machine is good for. Many of the habits of modern computing, such as direct interaction and an immediate response to the programmer, are visible in miniature in this first generation of transistor machines.
Read TX-0 as the proof that a computer made of thousands of transistors could be dependable and could be put in the hands of the people who wanted to use it, which set the template for the minicomputer and the culture of interactive programming.