The World War II years were a crucial period in the history of computing, when powerful gargantuan computers began to appear. Just before the outbreak of the war, in 1938, German engineer Konrad Zuse (1910–1995) constructed his Z1, the world’s first programmable binary computer, in his parents’ living room. The following year, American physicist John Atanasoff (1903–1995) and his assistant, electrical engineer Clifford Berry (1918–1963), built a more elaborate binary machine that they named the Atanasoff Berry Computer (ABC). It was a great advance—1000 times more accurate than Bush’s Differential Analyzer. These were the first machines that used electrical switches to store numbers: when a switch was “off”, it stored the number zero; flipped over to its other, “on”, position, it stored the number one. Hundreds or thousands of switches could thus store a great many binary digits (although binary is much less efficient in this respect than decimal, since it takes up to eight binary digits to store a three-digit decimal number). These machines were digital computers: unlike analog machines, which stored numbers using the positions of wheels and rods, they stored numbers as digits.
The first large-scale digital computer of this kind appeared in 1944 at Harvard University, built by mathematician Howard Aiken (1900–1973). Sponsored by IBM, it was variously known as the Harvard Mark I or the IBM Automatic Sequence Controlled Calculator (ASCC). A giant of a machine, stretching 15m (50ft) in length, it was like a huge mechanical calculator built into a wall. It must have sounded impressive, because it stored and processed numbers using “clickety-clack” electromagnetic relays (electrically operated magnets that automatically switched lines in telephone exchanges)—no fewer than 3304 of them. Impressive they may have been, but relays suffered from several problems: they were large (that’s why the Harvard Mark I had to be so big); they needed quite hefty pulses of power to make them switch; and they were slow (it took time for a relay to flip from “off” to “on” or from 0 to 1).
Photo: An analog computer being used in military research in 1949.
Most of the machines developed around this time were intended for military purposes. Like Babbage’s never-built mechanical engines, they were designed to calculate artillery firing tables and chew through the other complex chores that were then the lot of military mathematicians. During World War II, the military co-opted thousands of the best scientific minds: recognizing that science would win the war, Vannevar Bush’s Office of Scientific Research and Development employed 10,000 scientists from the United States alone. Things were very different in Germany. When Konrad Zuse offered to build his Z2 computer to help the army, they couldn’t see the need—and turned him down.
On the Allied side, great minds began to make great breakthroughs. In 1943, a team of mathematicians based at Bletchley Park near London, England (including Alan Turing) built a computer called Colossus to help them crack secret German codes. Colossus was the first fully electronic computer. Instead of relays, it used a better form of switch known as a vacuum tube (also known, especially in Britain, as a valve). The vacuum tube, each one about as big as a person’s thumb and glowing red hot like a tiny electric light bulb, had been invented in 1906 by Lee de Forest (1873–1961), who named it the Audion. This breakthrough earned de Forest his nickname as “the father of radio” because their first major use was in radio receivers, where they amplified weak incoming signals so people could hear them more clearly. In computers such as the ABC and Colossus, vacuum tubes found an alternative use as faster and more compact switches.
Just like the codes it was trying to crack, Colossus was top-secret and its existence wasn’t confirmed until after the war ended. As far as most people were concerned, vacuum tubes were pioneered by a more visible computer that appeared in 1946: the Electronic Numerical Integrator And Calculator (ENIAC). The ENIAC’s inventors, two scientists from the University of Pennsylvania, John Mauchly (1907–1980) and J. Presper Eckert (1919–1995), were originally inspired by Bush’s Differential Analyzer; years later Eckert recalled that ENIAC was the “descendant of Dr Bush’s machine.” But the machine they constructed was far more ambitious. It contained nearly 18,000 vacuum tubes (nine times more than Colossus), was around 24 m (80 ft) long, and weighed almost 30 tons. ENIAC is generally recognized as the world’s first fully electronic, general-purpose, digital computer. Colossus might have qualified for this title too, but it was designed purely for one job (code-breaking); since it couldn’t store a program, it couldn’t easily be reprogrammed to do other things.
ENIAC was just the beginning. Its two inventors formed the Eckert Mauchly Computer Corporation in the late 1940s. Working with a brilliant Hungarian mathematician, John von Neumann (1903–1957), who was based at Princeton University, they then designed a better machine called EDVAC (Electronic Discrete Variable Automatic Computer). In a key piece of work, von Neumann helped to define how the machine stored and processed its programs, laying the foundations for how all modern computers operate.  After EDVAC, Eckert and Mauchly developed UNIVAC 1 (UNIVersal Automatic Computer) in 1951. They were helped in this task by a young, largely unknown American mathematician and Naval reserve named Grace Murray Hopper (1906–1992), who had originally been employed by Howard Aiken on the Harvard Mark I. Like Herman Hollerith’s tabulator over 50 years before, UNIVAC 1 was used for processing data from the US census. It was then manufactured for other users—and became the world’s first large-scale commercial computer.
Machines like Colossus, the ENIAC, and the Harvard Mark I compete for significance and recognition in the minds of computer historians. Which one was truly the first great modern computer? All of them and none: these—and several other important machines—evolved our idea of the modern electronic computer during the key period between the late 1930s and the early 1950s. Among those other machines were pioneering computers put together by English academics, notably the Manchester/Ferranti Mark I, built at Manchester University by Frederic Williams (1911–1977) and Thomas Kilburn (1921–2001), and the EDSAC (Electronic Delay Storage Automatic Calculator), built by Maurice Wilkes (1913–2010) at Cambridge University.