Electronic Computer
An electronic computer is a machine that does its computing by switching electrical signals with devices that have no moving parts — vacuum tubes at first, then transistors and integrated circuits — rather than with the gears, rods, and clattering relays of mechanical and electromechanical calculators. The distinction is about speed and physics: a relay has to physically move a metal contact, which limits it to a few operations per second per switch, while a vacuum tube or transistor changes state by steering electrons, thousands to billions of times faster. Almost everything a modern person calls a computer, from a phone to a data-center server, is an electronic computer whose basic element is a switch that happens to be built from a semiconductor.
The electronic era began in wartime, and several machines have a fair claim to being first, depending on which features are demanded. The Atanasoff-Berry computer, designed by John Atanasoff and Clifford Berry at Iowa State College between roughly 1939 and 1942, used about 300 vacuum tubes to solve systems of linear equations in binary, with a rotating capacitor drum for memory — electronic, but built for one job and not programmable. Colossus, designed by Tommy Flowers and installed at Bletchley Park in 1943–44 to attack German Lorenz teleprinter ciphers, was a large electronic machine of more than a thousand tubes (about 2,400 in the later Mark 2) that could be set up for different logical tests with switches and plugs, but it was a code-breaking engine, and its existence stayed secret for decades. ENIAC, finished in 1945 at the University of Pennsylvania, was the first large general-purpose electronic digital computer, though it was programmed by rewiring.
The next real leap was not electronics alone but electronics plus the stored program, the idea that instructions live in the same fast memory as data. The Manchester Baby, built by Frederic Williams, Tom Kilburn, and Geoff Tootill at the University of Manchester, ran its first program on 21 June 1948 from a small Williams-Kilburn cathode-ray-tube memory, and EDSAC at Cambridge entered regular service in 1949. Commercial machines followed quickly: the UNIVAC I, from the company founded by J. Presper Eckert and John Mauchly, was delivered to the US Census Bureau in 1951, and IBM's 700 series arrived in the early 1950s. Changing a program now meant loading new numbers instead of re-plugging cables, which is what turned the computer from a bespoke instrument into a general tool.
The usual way to describe the history is by generations, each defined by its switching component. First-generation machines used vacuum tubes, which were bulky, hot, power-hungry, and prone to burning out. Second-generation machines replaced tubes with transistors, which were smaller, cooler, and far longer-lived. Third-generation machines used integrated circuits, with many transistors fabricated together on one chip, and the fourth generation is marked by the microprocessor, an entire central processing unit on one chip, beginning with Intel's 4004 in 1971. The labels are a convenient shorthand rather than sharp boundaries, since tube and transistor machines overlapped for years, and the later generations differ more by scale of integration than by any single change in design.
The transistor entered computing in the mid-1950s, in a flurry of experimental machines that proved the idea before it was economical. The University of Manchester's experimental transistor computer was running in 1953, Bell Labs built TRADIC for the US Air Force around 1954, and the Atomic Energy Research Establishment at Harwell completed the CADET in 1955, a machine of a few hundred point-contact transistors that was prized for running reliably for long stretches. MIT's Lincoln Laboratory built the TX-0 in the same period, a transistorized machine that later became famous as an early interactive computer. Their payoff was reliability and size: a tube computer might need a room-sized power and cooling plant and constant tube replacement, while a transistor design cut power draw sharply and lasted far longer between faults.
A common misconception is that one machine, usually ENIAC, was the electronic computer, and that the history is a clean single-file line. In reality the credit is distributed: Atanasoff and Berry demonstrated electronic binary arithmetic, a United States court in 1973 declined to treat ENIAC's key patent as valid partly because of that prior work, the British Colossus predated ENIAC but was classified, and the stored-program idea came from the design work around the EDVAC and was first realized in Manchester. Another pitfall is equating electronic with digital; analog electronic computers existed in numbers, and electronic simply says how the switching is done, not whether the numbers are discrete.
The modern relevance is that the definition has not changed, only the scale. A current processor still computes by switching electrons through transistors, but with tens of billions of them on one chip, switching in fractions of a nanosecond, so that the same logical idea that filled ENIAC's room now fits in a fingernail-sized die inside a phone, and GPUs and AI accelerators simply do vast numbers of electronic multiply-and-add operations in parallel. The long-run story is the shrinking of the switch, which is why the history of the electronic computer and the history of the transistor are so tightly intertwined.
Read electronic computer as a statement about the switch, not the shape of the machine: the moment computation moved from moving metal to moving electrons, speed, size, and reliability could improve without the old mechanical limits, and every later generation, from tubes to transistors to microprocessors, is that same step repeated at smaller and smaller scales.