Jean Bartik Learned ENIAC Without a Programming Manual. Then She Helped Redesign It.

When Jean Jennings Bartik arrived in Philadelphia in 1945, “computer” was still a job title. She had just graduated with a mathematics degree and joined the U.S. Army’s Ballistics Research Laboratory, where teams of mostly women calculated artillery firing tables by hand with mechanical desk calculators. Then she was selected for a stranger assignment: program a new electronic machine called ENIAC.

There was no programming language—and barely a map

ENIAC did not resemble a modern computer with code stored in memory. To make it solve a new problem, programmers had to understand the machine’s units, design the flow of the computation, set thousands of switches, and connect panels with cables. The six women selected for the job—Jean Jennings Bartik, Kathleen McNulty, Betty Snyder Holberton, Marlyn Wescoff Meltzer, Frances Bilas Spence, and Ruth Lichterman Teitelbaum—were given diagrams and access to the hardware, but no established programming profession to copy.

Computer History Museum records describe Bartik and the team as teaching themselves how ENIAC worked and becoming its first programmers. Their work was not clerical operation. They had to translate differential equations for ballistic trajectories into a sequence that ENIAC’s accumulators, function tables, and control circuits could execute. Debugging could mean tracing a wrong numerical result back through a physical configuration of cables and switches.

The public demonstration hid the hardest part

When ENIAC was unveiled in 1946, the machine became the star. It was enormous, filled with thousands of vacuum tubes, and could calculate far faster than the electromechanical systems that came before it. The women who programmed the demonstration were barely acknowledged publicly. The ENIAC Programmers Project later documented how the six programmers had prepared the ballistics program that helped demonstrate the machine’s speed, yet were not introduced to the press as the people who made it run.

That invisibility shaped the way early computing history was told for decades. Hardware designers were easy to photograph and name. Programming itself was new, poorly understood, and often treated as secondary work. In hindsight, the opposite is obvious: a general-purpose machine is only useful when someone can turn a problem into a reliable sequence of operations.

Bartik then helped change what “programming” meant

ENIAC’s original plugboard programming method was flexible but slow to reconfigure. In 1947, Bartik joined the effort to convert ENIAC into a stored-program machine. Instead of rewiring large portions of the computer for every task, instructions could be represented in a more compact form and read from the machine’s function tables. The change did not make ENIAC identical to later stored-program computers, but it significantly reduced setup time and moved programming closer to the model that would dominate computing.

Bartik later worked on BINAC and UNIVAC I, carrying experience from one of the earliest electronic computers into the emerging commercial computing industry. Her career therefore spans two foundational transitions: from human calculation to electronic computation, and from configuring hardware directly to expressing instructions in stored form.

The deeper lesson is about abstraction

Modern developers rarely think about voltage levels, patch cables, or which accumulator physically holds a number. That distance from the hardware is possible because generations of programmers created better abstractions. Bartik’s work sits near the beginning of that story. She had to understand almost everything the machine was doing because the layer between mathematical intent and hardware execution was still thin.

It is tempting to describe early programmers as people who merely “operated” giant machines. Bartik’s career shows why that description is misleading. She and her colleagues invented procedures, decomposed problems, debugged systems, and helped reshape the architecture so that future programming could become less physical and more expressive. The tools changed dramatically; the intellectual work is recognizable to any software engineer today.

Explore

Stay connected