June 30, 1945

Von Neumann architecture

Initial circulation of von Neumann's First Draft of a Report on the EDVAC

The report dated June 30 may already have been in 24 people's hands five days earlier — a 101-page working draft bearing one name, for a design developed through collaborative EDVAC discussions.

June 1945: Goldstine sends 24 copies

The document bears the date 30 June 1945, but the best-supported account says Herman Goldstine sent 24 copies of John von Neumann's typescript to people connected with the EDVAC project on 25 June, with wider circulation around and after the printed date.

The incomplete, 101-page report set out the logical organization of an electronic digital computer whose memory held both numerical data and coded operating instructions — making a change of program an operation on symbols rather than a rewiring job.

Five organs and a 32-bit word

The draft divided the proposed machine into five organs — arithmetic, central control, memory, input and output — and discussed binary, serial operation, addressable storage and a suggested word length of 32 binary digits. An external recording medium was used through the input and output organs, not counted as a sixth organ.

Calling it a 'publication' needs care: von Neumann and Goldstine regarded it as an internal working document. Its uncontrolled copying gave it publication-like influence anyway.

One name on a collaborative design

Goldstine's typescript named only von Neumann. Its circulation obscured the contributions of other EDVAC participants — J. Presper Eckert and John Mauchly among them — and became a source of conflict within the Moore School group over credit and patentability.

Historians dispute individual priority for the stored-program concept: the report synthesized collaborative EDVAC work, and should not be read as proof that von Neumann alone invented the architecture that carries his name.

Sources

Researched 23 Aug 2026 7 sources 3 audit passes Date disputed

How this was checked

Researched from the web into a fact sheet, rewritten from that sheet, then audited against it by a different model. How the pipeline works →

Date disputed. Sources disagree; the research places it at 1945-06-25.

What the sources leave uncertain

June 30 is the date printed on the report, but the best-supported account says Goldstine distributed 24 copies on June 25. Later copies circulated more widely around and after June 30, so sources often use the printed date as the event date.
Calling the draft a “publication” is potentially misleading: von Neumann and Goldstine regarded it as an internal working document, although its uncontrolled circulation gave it publication-like influence.
Historians dispute individual priority for the stored-program concept. The report synthesized collaborative EDVAC work and should not be treated as proof that von Neumann alone invented the architecture.