Somewhere in your house right now, there is a device built according to a plan written in 1945. It might be the phone in your pocket. It might be the laptop on your desk. It might be the thermostat on your wall or the chip inside your car. Every one of these machines follows an architectural blueprint laid out in a single document: the First Draft of a Report on the EDVAC, written by John von Neumann and distributed on June 30, 1945.
The report was 101 pages long. It was typed, not published. It was stamped “Confidential” and circulated among a small group of engineers and mathematicians at the University of Pennsylvania’s Moore School of Electrical Engineering. It was never intended for wide distribution. And yet it became the most influential unpublished document in the history of computing.
This is the story of how it came to be, what it said, and why it made some very talented people extremely angry.
The Problem With ENIAC
To understand the EDVAC report, you first need to understand what came before it. In 1943, the U.S. Army was fighting a war, and that war required an enormous number of mathematical calculations. Ballistic firing tables – the charts that told artillerymen how to aim their guns – had to be computed by hand, and the process was painfully slow. Each table required weeks of work by teams of human “computers,” mostly women with mathematics degrees.
J. Presper Eckert and John Mauchly, two engineers at the Moore School, proposed building an electronic machine that could do these calculations automatically. The result was ENIAC – the Electronic Numerical Integrator and Computer. When it was completed in late 1945, it was enormous: 30 tons, 18,000 vacuum tubes, and a power consumption that allegedly dimmed the lights in an entire section of Philadelphia.
ENIAC worked. It was astonishingly fast for its time. But it had a fundamental problem: programming it was a nightmare.
Rewiring the Machine
To set up a new calculation on ENIAC, operators had to physically rearrange cables and flip switches. The machine’s “program” was its wiring. Changing from one problem to another could take days of manual labor. This was not a minor inconvenience. It meant that ENIAC was essentially a different machine each time it was reconfigured.
Everyone involved knew this was unsustainable. Even before ENIAC was finished, the team began thinking about its successor: a machine called EDVAC (Electronic Discrete Variable Automatic Computer). And this is where John von Neumann entered the picture.
Von Neumann Arrives
By 1944, von Neumann was already one of the most famous mathematicians alive. Born in Budapest in 1903, he had made fundamental contributions to quantum mechanics, game theory, and mathematical logic before he turned forty. He worked at the Institute for Advanced Study in Princeton, alongside Einstein and Godel. He was a consultant to the Manhattan Project. He was, by every account, one of the fastest thinkers anyone had ever encountered.
Von Neumann learned about ENIAC almost by accident. In the summer of 1944, he ran into Herman Goldstine – a mathematician and Army liaison for the ENIAC project – at a train station in Aberdeen, Maryland. Goldstine recognized von Neumann and mentioned the computing project. Von Neumann was immediately fascinated. He began visiting the Moore School regularly, talking to Eckert, Mauchly, Goldstine, and the rest of the team about what the next machine should look like.
The conversations were intense. The engineers had been wrestling with the stored-program concept for months. The core idea was radical but logical: instead of wiring a program into the hardware, why not store the program in memory, alongside the data? The machine could then read its instructions from memory, just as it read numbers. Changing the program would mean changing the contents of memory, not rewiring the machine.
The Report Takes Shape
Von Neumann, with his extraordinary ability to absorb and synthesize ideas, began writing up these discussions. The result was the First Draft of a Report on the EDVAC. It described a computer architecture with five key components:
- A central arithmetic unit (what we now call the ALU) to perform calculations
- A central control unit to coordinate operations and interpret instructions
- A memory unit to store both data and programs
- An input mechanism to feed data and programs into the machine
- An output mechanism to deliver results
The critical innovation was the stored-program concept. Programs and data would share the same memory space. Instructions would be encoded as numbers, just like the data they operated on. This meant a program could modify itself – a powerful and slightly unsettling capability that remains fundamental to computing today.
Von Neumann also drew an analogy between the computer and the human nervous system, using terminology borrowed from neuroscience. He called the logical components “neurons” and described their behavior in terms that would be familiar to anyone studying the brain. This was not mere metaphor. Von Neumann was genuinely interested in the parallels between electronic computation and biological information processing – an interest that would consume his later career.
The Controversy
Here is where the story gets uncomfortable. When Goldstine distributed the report, it carried only one name on the cover: John von Neumann. There was no mention of Eckert, Mauchly, or anyone else on the EDVAC team.
Eckert and Mauchly were furious, and it is hard to blame them. They had been developing stored-program ideas before von Neumann ever set foot in the Moore School. Eckert, in particular, had been thinking about mercury delay line memory – the technology that would make the stored-program concept physically possible – since at least early 1944. Mauchly had been dreaming of general-purpose computing machines since the early 1940s.
The question of who actually invented the stored-program computer has been debated by historians for decades, and the answer is genuinely complicated:
- Eckert and Mauchly contributed critical engineering ideas, particularly around memory technology
- Von Neumann contributed the logical framework and the mathematical rigor that made the architecture precise
- Other team members, including Arthur Burks, Goldstine, and several others, contributed ideas during group discussions
- Alan Turing’s 1936 paper on computable numbers had already described a theoretical universal machine – a concept von Neumann knew well (more on this in our post on Turing’s work)
The fairest reading is that the stored-program architecture emerged from a collaborative process, and von Neumann’s report was a brilliant synthesis rather than a solo invention. But because his name was on the document, history attached his name to the architecture. We call it the “von Neumann architecture” to this day.
The Patent Disaster
The attribution question had real consequences beyond bruised egos. Eckert and Mauchly wanted to patent the EDVAC design. But because von Neumann’s report had been distributed without restrictions, it functioned as a prior publication, making the ideas unpatentable. The stored-program computer entered the public domain before anyone could claim ownership of it.
This infuriated Eckert and Mauchly, who left the Moore School and eventually founded their own company. They built UNIVAC, the first commercially produced computer in the United States, but they spent years in patent disputes that drained their resources and their spirit. Mauchly later said that the distribution of von Neumann’s report was “the most damaging thing” that ever happened to them professionally.
Whether this outcome was accidental or deliberate remains unclear. Some historians believe Goldstine and von Neumann wanted the ideas to be freely available. Others think it was simply careless.
Why It Still Matters
Whatever the politics behind it, the architecture described in the EDVAC report proved extraordinarily durable. Pick up any computer science textbook today and you will find the same basic diagram: processor, memory, input, output, connected by buses that shuttle data and instructions back and forth.
Your smartphone follows this architecture. Your laptop follows it. The servers that run the internet follow it. Even modern processors with multiple cores, deep pipelines, and layers of cache memory are essentially elaborations on the blueprint von Neumann wrote in 1945. The fundamental idea – that a machine can store its own instructions and treat them as data – remains the bedrock of general-purpose computing.
There have been challenges. The “von Neumann bottleneck,” the limitation imposed by a single channel between processor and memory, has been a persistent headache for hardware designers. Inventors and engineers have devised increasingly clever workarounds: caches, branch prediction, out-of-order execution. But the basic model endures.
Newer paradigms are emerging. Quantum computing operates on entirely different principles. Neuromorphic chips try to mimic the brain more literally than von Neumann’s loose analogy ever intended. But for the vast majority of computation happening on Earth right now, the EDVAC report’s architecture is still the one doing the work.
The Document as Artifact
There is something remarkable about the fact that a typed, 101-page draft, never formally published, never peer-reviewed, never even finished (it was, after all, a “first draft”), became the founding document of the computer age. It was not a polished treatise. It was a working document, full of technical detail and neurological analogies, written by a mathematician trying to make sense of a new kind of machine.
That is precisely what makes it fascinating. The EDVAC report captures a moment when the future was being figured out in real time, in conversations between engineers and mathematicians who did not yet have the vocabulary for what they were building. Von Neumann gave it a vocabulary. Whether he also gave it proper credit is another question entirely.
If you are drawn to the moments where ideas first crystallize into the technologies we take for granted, you may want to explore Portraying Science, which examines how breakthroughs have been visualized and communicated across centuries. The EDVAC report is a reminder that the most consequential documents are not always the most famous – and that the line between collaboration and credit is never as clean as we would like it to be.
No products in the cart.