The Computational Burden Behind Firing Tables
During the Second World War, the U.S. Army needed extensive firing tables describing projectile flight under different conditions. Human computers used desk calculators and differential analyzers to solve equations, but repeated operations created backlogs. In 1943, the Army supported the University of Pennsylvania's Moore School of Electrical Engineering in Philadelphia in developing the Electronic Numerical Integrator and Computer, ENIAC, to replace many slow electromechanical operations with fast electronic circuits.
John Mauchly proposed the electronic-computing approach, Presper Eckert led key engineering design, and Herman Goldstine coordinated Army requirements. Engineers, technicians and human computers at the school jointly advanced construction. Major assembly was completed and computations began in 1945; public demonstrations and the unveiling in February 1946 brought the project into public view.
References: [3]
Numbers Moving Through Pulse Circuits
ENIAC used decimal representation. Twenty accumulators could hold signed ten-digit numbers and perform addition and subtraction; multiplication, division and square-root units handled more complex operations. Function-table switches stored prearranged values, while punched cards handled some input and output. Common timing signals coordinated the units, with numbers transmitted between components as electrical pulses.
The 1946 technical report listed an addition time of 200 microseconds, equivalent to a rate of 5,000 additions per second. This was internal arithmetic speed, not total efficiency from preparing a problem to obtaining an answer. Card reading, punching, circuit setup, fault checks and assessment of trustworthy results required additional time and labor.
References: [1]
A Program First Became Diagrams and Wiring
The original ENIAC did not read a complete instruction sequence from a unified memory as later computers did. Staff first divided an algorithm into steps executable by the available components, then set switches and connected cables. Data lines sent numbers to the correct locations, while pulses on control lines determined which units started when. The master programmer coordinated repetition and transitions.
Adding a number in one accumulator to another, for example, required both a data path and correctly timed action by sender and receiver. Repeatedly accumulating a group of numbers also required input, counting, repetition and stopping conditions. Wiring was not mechanical plugging according to a formula, but organization of an autonomous computational process within limited components and timing constraints.
How Six Women Made the Plan Run
The first six women programmers were Kathleen McNulty, Frances Bilas, Betty Jean Jennings, Betty Snyder, Ruth Lichterman and Marlyn Wescoff. Previously involved in manual numerical computation, they learned to map problems onto ENIAC using logical diagrams, circuit information and discussions with engineers.
Participant recollections collected by W. Barkley Fritz describe pairs working on different tasks. Some integrated the ballistics program, while others separately calculated test trajectories in the steps the machine would use to check intermediate results. They had to distinguish mathematical-planning errors, switch or wiring mistakes and hardware faults. Their knowledge therefore combined equations, timing, machine capabilities and verification.
References: [2]
The Public Saw an Organized Computation
Ballistic-trajectory computation was a central part of the February 1946 demonstration, with Jennings and Snyder helping complete the demonstration program. Audiences saw rapid results, but behind them lay prior design, setup, testing and debugging. A successful run demonstrated that coordinated electronic components could execute complex calculations continuously, not that the machine independently understood any mathematical problem presented to it.
Adele Goldstine's technical description became formal documentation as part of the ENIAC report that June, alongside operating and maintenance manuals. These materials made arithmetic units, program control and testing methods teachable knowledge. Beyond the giant machine and a few designers featured in publicity, programming, documentation and routine maintenance were equally decisive in making the system usable.
From Unveiling to Sustained Use and Modification
In 1947, ENIAC moved to Aberdeen Proving Ground in Maryland. It continued performing different numerical tasks, and its setup methods were repeatedly revised. A 1948 programming conversion used function-table switches to store coded instructions, bringing execution closer to the emerging sequential-instruction model. This switch storage, however, was not modern memory freely writable by a program.
That change came after the public unveiling. Distinguishing the 1943 start, completion and initial computation in 1945, the 1946 unveiling and later modifications reveals a system continually remade by its users. Its significance lay not only in speed but in showing the division of work required for general-purpose electronic computing: engineers built reliable components, programmers arranged steps, and operators and maintenance staff sustained computation.