EDVAC and the Stored-Program Idea: Moving Instructions into Memory

Early electronic computers made calculation fast, but changing what they did could remain cumbersome. ENIAC’s original programming method relied heavily on switches and cable connections. The EDVAC project developed at the Moore School pursued a different architecture in which instructions could be represented within the machine’s memory system. John von Neumann’s 1945 ‘First Draft of a Report on the EDVAC’ became one of the most influential descriptions of this emerging design.^[1]

The familiar phrase ‘von Neumann architecture’ can make the story sound like a single person’s invention. The historical record is more collaborative. J. Presper Eckert, John Mauchly, Herman Goldstine, von Neumann and others participated in overlapping discussions about the limitations of ENIAC and the architecture of its successor. The report mattered enormously because it circulated the ideas, but attribution remains a subject of historical debate.^[2]

Why rewiring was a bottleneck

ENIAC could execute arithmetic electronically at remarkable speed, yet configuring a new problem required substantial physical preparation. The separation between fast execution and slow reconfiguration made programmability an architectural problem. Designers wanted a machine whose behavior could change by loading instructions rather than reconstructing a web of control connections.^[2][5]

Programming speed and execution speed are different

This distinction remains fundamental. A processor may execute billions of operations quickly, but software value depends on how efficiently humans can express, revise and test new behavior. Stored programs attacked the changeability problem by making instructions into information the machine itself could hold.

The 1945 First Draft

Von Neumann’s report described an automatic digital computing system divided into functional parts for arithmetic, control, memory, input and output. The Smithsonian Libraries preserves a public-domain scan of the document, allowing the architecture to be read as a historical artifact rather than only through later summaries.^[1]

A report became a transmission mechanism

The First Draft circulated beyond the immediate EDVAC team and influenced researchers building early computers in the United States and Britain. IEEE’s later publication notes its pervasive influence and observes that Alan Turing cited it in connection with the Pilot ACE.^[3]

Instructions in memory

The defining architectural shift was that program instructions could be encoded and stored within the machine’s memory system rather than existing only as external wiring or setup. This greatly increased flexibility: changing the computation could become a matter of changing stored information.

Programs become manipulable representations

Once instructions are represented internally, important possibilities follow. Programs can be loaded, copied and potentially altered by other processes. The stored-program idea therefore strengthened a conceptual relationship already visible in Turing’s universal machine: instructions can be treated as representations rather than being inseparable from physical machinery.

Stored-program does not automatically mean modern software

Early machines had tiny memories, idiosyncratic instruction sets and primitive input/output. Programming remained difficult. The architecture created conditions for software to become more flexible, but compilers, operating systems, libraries and interactive tools would require further decades of invention.

The credit problem

Computer History Museum historians caution against simplistic ‘first’ stories. The phrase ‘stored program’ itself was not the language used consistently by participants in the 1940s, and several related architectural ideas emerged through collaboration and parallel development.^[4] Calling the whole architecture solely von Neumann’s can obscure Eckert and Mauchly’s prior work and the institutional context of the Moore School.

From EDVAC plans to a wider movement

The Moore School lectures of 1946 helped spread new computer-design ideas to researchers who would build machines elsewhere. CHM describes the lectures as a crucial transmission point where attendees learned about ENIAC and the emerging concept of programs held in memory.^[2]

Architecture became a community project

By the late 1940s, groups at Manchester, Cambridge, Princeton and other institutions were turning variants of stored-program design into working machines. No single implementation exhausted the concept. The important historical transition was collective: programmable electronic computing moved toward memory-centered architectures.

Why EDVAC belongs in coding history

EDVAC matters to programmers because it helped move instructions inside the computational system. That change reduced the distance between program and data, made software easier to replace than hardware, and set the stage for programming to become a rapidly evolving layer of its own. The First Draft is therefore both an architecture document and a milestone in the history of code.

A better way to state the legacy

Rather than saying one person invented the modern computer, the evidence supports a more precise statement: the EDVAC project and the 1945 report helped consolidate and disseminate a stored-program architecture that became enormously influential. Its history is a case study in how collaborative engineering ideas can later acquire overly simple names.

RESEARCH / PROVENANCE

Works Cited

5 SOURCES

1. 01
   Smithsonian Libraries — John von Neumann, First Draft of a Report on the EDVAC (1945) ↗library.si.edu
2. 02
   Computer History Museum — Making the Case: Birth of the Computer ↗computerhistory.org
3. 03
   IEEE Xplore — First draft of a report on the EDVAC ↗ieeexplore.ieee.org
4. 04
   Computer History Museum — The Neverending Quest for ‘Firsts’ ↗computerhistory.org
5. 05
   University of Pennsylvania Almanac — 75th Anniversary of ENIAC ↗almanac.upenn.edu

CodeHistory is a living archive. Citations document the evidence used for this edition; later evidence may refine the account.