I just read John von Neumann’s First Draft of a Report on the EDVAC, a tremendously influential 1946 document about computer architecture. A paper copy is available on Amazon, and the identical document as a PDF here.

Written after his experience with ENIAC (where he, famously, configured the plugboards to implement stored instructions in memory), the document describes the architecture of a (then) next-generation machine. Supposedly this early draft (the only draft he ever produced) was circulated widely and led to many implementations of similar machines in the late 1940s.

It’s a fascinating historical document. One thing that jumped out at me (no pun intended) is that nowhere in the document does he mention the idea of conditional branches – without those it’s extremely difficult or impossible to make the machine Turing-complete.

Also, he seems to have conceived the machine purely as a programmable calculator. The concept of what we’d call “data processing” is completely absent.

Von Neumann is considered one of the smartest people to have ever lived – Hans Bethe said “I have sometimes wondered whether a brain like von Neumann’s does not indicate a species superior to that of man”, and Edward Teller (a very competitive guy) admitted that he “never could keep up with John von Neumann.”

Teller also said “von Neumann would carry on a conversation with my 3-year-old son, and the two of them would talk as equals, and I sometimes wondered if he used the same principle when he talked to the rest of us.”

I think this just goes to show how difficult it is to predict even the future of a narrow technology.