Here at PennApps, on the University of Pennsylvania campus, there is a rich history of computer innovation. Many undergraduate computer science majors have heard of ENIAC in passing, but what exactly was it? Was it the “first” computer? Who built it — and why? What is its legacy?
There are a lot of myths and urban legends surrounding these questions. Below is an abbreviated version of the story. More important, this version is true! Wikipedia and other sites often just repeat the same old mythology.
ENIAC, built from 1943-1945, was the Electronic Numerical Integrator and Computer. It was a general-purpose computer, so it could (hypothetically) solve any problem. It filled a whole classroom, was programmed with plug boards, and used 18,000 vacuum tubes for its digital zeros through nines — because parts of the system were decimal instead of binary.
ENIAC was not the “first” computer, and you shouldn’t let anyone tell you otherwise, not even your CS/EE professors. “First” is a very dangerous word about scholars of computer history, because it is subjective.
Charles Babbage designed (but never quite finished) a general-purpose mechanical computer in the 1830s. It didn’t work, but researchers in the 2000s determined it could have, if only Babbage had better precision in his machine tools. Most efforts during the next century focused on mechanical calculating, not computing. The difference is vital: a calculator has no decision-making (if-then) capability. There were several efforts to make electromechanical, general-purpose computers in the 1930s and early 1940s, featuring gears, relays, and solenoids. Examples are the IBM-designed Harvard Mark-1 and German engineer Konrad Zuse’s machines. The world also saw many general-purpose analog computers during this time.
British scientists made a computer named Colossus during the second World War. This was notable as the earliest known fully electronic computer, but it was a special-purpose machine, meaning that it only did one thing — compare patterns to break Axis codes.
You may also hear about the work of an Iowa professor named John Atanasoff. His team designed a special-purpose calculator. They never finished it, and that was that. They became famous 30 years later when Honeywell sued Remington Rand. Remington at that point owned and tried to enforce the (by then) largely obsolete ENIAC patents, which Honeywell and others wanted to invalidate. Honeywell’s attorneys found that Mauchly during his own research period had visited Atanasoff and stated his intention to build a similar design. This was a smoking gun, the judge said, in his decision that Honeywell won the case.
The decision was a landmark for the computer industry, because it allowed for much more competition. But it is almost universally accepted by modern historians that the judge got it wrong on several major points: Atanasoff’s design was a calculator, not a computer; it was not fully general-purpose; it was not fully electronic; Atanasoff’s architectural designs weren’t unique; and although its chief designed John Mauchly saw it and was polite to Atanasoff, the actual ENIAC design was nothing like that machine. Ironically, Atanasoff never gave his machine a name — it’s popularly known today as the “Atanasoff-Berry Computer”, but that name came from Honeywell lawyers. Unfortunately, a team of ENIAC engineers upset for not getting enough credit of their own, and a widely discredited author who wrote a book called “The Man Who Invented the Computer”, all give credence to the judge’s decision.
ENIAC’s place in history, solidified by serious historians despite its legal and popular discrediting by those with an agenda, became even stronger in 2012-2013 due to a happy discovery of archival material. British historians previously believed that U.K. computers ran in stored-program configurations before ENIAC was converted to that setup later in its lifespan. However, original documents proved that ENIAC has that record, too, if only by a few months.
So there you have it: ENIAC was a general-purpose, fully-electronic, stored-program computer — quite apparently the “first” such system. The best part: it was built here at Penn Engineering.
ENIAC’s lead designers were the physics professor Mauchly and engineering graduate student Pres Eckert, both working at Penn’s Moore School of Electrical Engineering. The programmers — “computors” — were the university’s smartest female math majors. The machine was originally called Project PX and its eventual destination was the Army’s Aberdeen (Maryland) Proving Ground. Common understanding states that ENIAC was designed to compute trajectories for firing tables during the war effort. However, that’s not true! Mauchly and Eckert cited the ballistics application in their funding proposal to the U.S. Army, in hopes that bullets-and-brass types would understand that — but they had a wide variety of applications in mind, such as cryptology and meteorology, to name a few. Historians many decades later pointed to the funding proposal and misread it as the machine’s primary function. (Mauchly in the 1970s told a Smithsonian historian that he was “impelled” to build a general-purpose computer starting around 1941, when he was contracted to do math for parabolic radar antennas at the Army’s Evans Signal Laboratory, located in Wall, New Jersey, which is now a science museum.)
Years later, an urban legend stated that the lights in Philadelphia dimmed whenever the machine was powered on. Eckert unequivocally said that wasn’t true. ENIAC’s real-world weakest link was how long it took to program — a single application could take days or even weeks to wire up before the computer was changed to its stored-program design. However, when actually running the code, ENIAC was thousands of times faster than any of its electromechanical peers.
Parts of ENIAC are on public display at the Moore School, which you should visit if you’re in town for PennApps.
ENIAC’s legacy lives on the Philadelphia area. After the war, its designers left Penn and formed the Eckert-Mauchly Computer Company. Their main product was the series of UNIVAC mainframes. Remington Rand acquired the company, and also acquired Minnesota’s Engineering Research Associates, combining both firms into Remington’s new UNIVAC division. Remington then merged with Sperry and became known as Sperry Rand. Finally, Sperry Rand merged with Burroughs, taking the new name Unisys — which is still based in the Philadelphia area.