ENIAC

John Presper Eckert Jr. (1919–1995) and John W. Mauchly (1907–1980) designed and built ENIAC at the Moore School of Electrical Engineering, University of Pennsylvania. Eckert, an electronics prodigy, solved the engineering challenges of scaling vacuum tubes to reliable computation; Mauchly, a physicist, conceived the machine's architecture. Their team of roughly 50 engineers and technicians—including Betty Jennings, Betty Snyder, Marlyn Meltzer, Fran Bilas, Ruth Lichterman, and Kathleen Antonelli, the six women programmers who debugged and operated the machine—made ENIAC functional. The U.S. Army Ordnance Department funded the $486,804 project (roughly $8 million in 2024 dollars) to accelerate the computation of firing tables for artillery.

ENIAC replaced mechanical relays and electromechanical switches with vacuum tubes (thermionic valves), each acting as a fast electronic switch. The machine used the decimal (base-10) system rather than binary, encoding numbers as patterns of tube states. Two parallel computing units—the Accumulator and Function Table—processed data in parallel where possible. Eckert designed a redundancy protocol: each logical operation was performed twice and compared; if results differed, the machine halted, alerting operators to a tube failure. This reliability strategy was essential because vacuum tubes failed frequently (roughly one every 15 minutes on average, though the machine could run for hours). The machine's architecture used 'pulse trains'—sequences of electrical pulses—to represent numbers and control operations. A master clock oscillator at 100 kHz synchronized all operations. Cooling was critical: the 18,500 tubes generated enormous heat, requiring a dedicated air-conditioning system and forced-air circulation through the cabinet.

ENIAC emerged from the U.S. Army's urgent need to compute artillery firing tables during World War II. Traditional mechanical calculators and human 'computers' (mathematicians performing calculations by hand) could not keep pace with the volume of ballistic data required by new weapons. In 1943, the Army contracted the Moore School to build an electronic calculator. Eckert and Mauchly, drawing on prior work in radar electronics and the theoretical foundations of Boolean logic, proposed an all-electronic machine using vacuum tubes. Construction began in 1944; the machine was completed and tested in late 1945. On February 14, 1946, ENIAC was formally dedicated at the Moore School. It was the first general-purpose electronic digital computer—a watershed moment in the history of computation. Unlike earlier electromechanical machines (such as the IBM Automatic Sequence Controlled Calculator, or 'Harvard Mark I'), ENIAC could be reprogrammed by rewiring and resetting switches, making it adaptable to different problems. Its speed—performing thousands of arithmetic operations per second—was revolutionary. However, ENIAC's reliability remained a challenge: the machine required constant maintenance, and downtime for tube replacement was frequent. Despite these limitations, ENIAC demonstrated the feasibility of electronic computation and inspired a generation of engineers and mathematicians to pursue digital computing. The machine operated until 1955, solving problems in ballistics, weather prediction, atomic physics, and pure mathematics.

The U.S. Army Ordnance Department needed to compute firing tables for artillery with unprecedented speed and accuracy. During World War II, the Army had developed new weapons—guns with ranges and trajectories that required complex ballistic calculations. Each new gun required a complete set of firing tables (printed booklets showing the correct elevation and windage for different ranges and weather conditions). These tables were computed by teams of human 'computers' (mostly women mathematicians) using mechanical desk calculators; the process was slow and error-prone. A single firing table could take weeks to compute. The Army recognized that mechanical computation was reaching its limits. Radar technology, developed during the war, had demonstrated the power of vacuum-tube electronics for rapid signal processing. Eckert and Mauchly proposed applying similar electronics to computation itself. The Army funded ENIAC as a wartime research project, hoping to accelerate ballistic calculations. Although the war ended before ENIAC was completed, the machine's potential was immediately recognized by the scientific community. It became a tool for research in nuclear physics, weather prediction, and pure mathematics—problems that required vast computational resources.

ENIAC was operated by a small team of engineers and programmers at the Moore School. A typical day began with the morning 'warm-up': technicians powered on the machine and allowed the vacuum tubes to stabilize for several hours before computation began. The six primary programmers—Betty Jennings, Betty Snyder, Marlyn Meltzer, Fran Bilas, Ruth Lichterman, and Kathleen Antonelli—prepared a new problem by hand-wiring the machine's interconnections according to a detailed schematic. This rewiring process could take days for a complex problem. Once configured, the machine was set running; operators monitored the control console, watching for lights indicating errors or tube failures. If a tube failed, the machine halted, and technicians replaced the faulty tube (a process that took 15–30 minutes). The programmers debugged the machine by observing the pattern of lights on the control panel and tracing the flow of pulses through the circuit. Results were output on punch cards or paper tape, which were then analyzed by hand. A single problem run might take hours or days, depending on complexity. The machine was not idle; there was always a queue of problems waiting to be solved. At day's end, technicians performed routine maintenance—checking connections, replacing worn tubes, and cleaning dust from the cooling system.

The ENIAC team comprised roughly 50 people, organized into several groups. The core design team included John Presper Eckert Jr. (chief engineer, electronics), John W. Mauchly (principal investigator, architecture), and Herman Goldstine (liaison with the Army, mathematician). Supporting engineers included Arthur W. Burks (logic design), Harry Huskey (memory and control systems), and others. The six primary programmers were Betty Jennings (née Holberton), Betty Snyder (née Holberton), Marlyn Meltzer, Fran Bilas (née Bilas), Ruth Lichterman (née Teitelbaum), and Kathleen Antonelli (née McNulty). These women, most with backgrounds in mathematics or physics, developed the first programming techniques and debugged the hardware by hand. Technicians and assemblers—whose names are largely unrecorded—soldered connections, tested tubes, and maintained the machine. The Army Ordnance Department provided funding and oversight through officers such as Colonel Paul N. Gillon. The project was housed at the Moore School of Electrical Engineering, University of Pennsylvania, under the direction of Dean J. Peirce.

Construction of ENIAC began in 1944 and proceeded in phases. Eckert and Mauchly first designed the overall architecture and logic circuits, then oversaw the fabrication of individual modules—accumulators, function tables, and control circuits—each comprising dozens of vacuum tubes and supporting components. Technicians hand-soldered approximately 500,000 solder joints. The machine was assembled in a large room at the Moore School, with modules mounted on metal frames and interconnected by thousands of cables. Each cable was hand-routed and labeled. The power supply, a massive transformer and rectifier system, was built to deliver the precise voltages required by different tube types. The cooling system—a forced-air duct and air-conditioning unit—was installed to manage the heat generated by 18,500 tubes. Testing began in late 1945; the machine was brought up gradually, with individual modules tested before integration. Debugging was painstaking: engineers traced the flow of pulses through circuits using oscilloscopes and logic analyzers. The machine was declared operational in December 1945 and formally dedicated on February 14, 1946. Final construction cost was $486,804 (roughly $8 million in 2024 dollars).

ENIAC was a unique machine; no identical copies were built. However, its design influenced several successor machines. The EDVAC (Electronic Discrete Variable Automatic Computer), designed by Eckert and Mauchly and completed in 1951, incorporated lessons learned from ENIAC, using binary representation and stored-program architecture. The IAS machine (Institute for Advanced Study Electronic Computer), designed by John von Neumann and built at Princeton, also drew on ENIAC's innovations. The UNIVAC I (Universal Automatic Computer), the first commercial electronic computer, was built by Eckert and Mauchly's company (Eckert-Mauchly Computer Corporation, later acquired by Remington Rand) and delivered in 1951. These machines shared ENIAC's vacuum-tube technology but improved reliability, speed, and programmability. ENIAC itself was modified during its operational life: additional memory units were added, and the control system was upgraded to support new programming techniques. However, these modifications were incremental; the fundamental architecture remained unchanged.

ENIAC itself is the famous example—there was only one machine. However, its impact on subsequent computing is profound. The EDVAC (1951), designed by Eckert and Mauchly, incorporated ENIAC's vacuum-tube technology but added stored-program architecture (following John von Neumann's theoretical work). The IAS machine (1952), designed by von Neumann at Princeton, also drew on ENIAC's design principles. The UNIVAC I (1951), the first commercial electronic computer, was built by Eckert and Mauchly's company and became the standard for business and scientific computing in the 1950s. The IBM 701 (1952), IBM's first electronic computer, competed with UNIVAC and incorporated lessons from ENIAC. These machines, and dozens of others built in the 1950s and 1960s, all traced their lineage to ENIAC's pioneering design.

ENIAC was not buried or lost; it was dismantled in 1955 and its components were salvaged. However, portions of the machine survive in museums and archives. The Smithsonian Institution holds documentation, photographs, and some original components. The University of Pennsylvania's Moore School of Electrical Engineering maintains archival materials related to ENIAC's design and construction. The Computer History Museum in Mountain View, California, preserves ENIAC-related artifacts and documentation. No intact ENIAC machine exists; the original was too large and power-hungry to preserve as a working artifact. However, a full-scale replica of a portion of ENIAC was constructed at the University of Pennsylvania in the 1990s for educational purposes. The replica, using modern components but following ENIAC's original logic, demonstrates the machine's operation to visitors.