One episode in the history of computing described in the new book Invisible in the Storm: The Role of Mathematics in Understanding Weather by Ian Roulstone and John Norbury (Princeton University Press, 2013) invoked a very personal memory.
In December 1945 ENIAC - Electronic Numerical Integrator And Computer - became operational at the University of Pennsylvania. ENIAC has been constructed under a contract to the United States Army. By the time a team of meteorologists gathered by John von Neumann at Princeton University obtained access to the computer, it was located at the Aberdeen Proving Ground, Maryland. The backbone of the project was an iterative process devised by von Neumann and Jule Charney. ENIAC was operated via punched cards. The calculations were performed by the computer but it was humans who fed the computer punched cards input. The output also consisted of punched cards, and again it were the human operators who had to collate the cards and submit the bunch to the computer for the next iteration.
Imperfect and sluggish as the whole configuration has been, the researchers may have been lucky to have a one-of-a-kind computer. Almost three decades later while working on my thesis at the Hebrew University of Jerusalem, I got a side project from the University of Ben-Gurion in Be'er Sheva to develop a simulation for arrays of solar batteries. Luckily or so it seemed to me at the time, both universities ran the same mainframe - the CDC-6400. It was a great convenience to work in Jerusalem, where I lived, and only visit Be'er Sheva for a demonstration of the finished program.
When the time came I drove to Be'er Sheva with a neatly packaged bunch of punched cards. To my horror, the demonstration started as a complete failure. The Be'er Sheva's CDC-6400 found multiple errors in the program that Jerusalem's CDC-6400 ran successfully. After a hurried investigation it came about that the card punching and reading devices at the two locations were interpreting several symbols differently. The remedy was in writing a program that took as input my original bunch of punched cards and produced another bunch suitable for the other computer as the output - a feat that I am sure von Neumann's team was more than able to perform but did not have to.
As a side recollection, I had to run the programs I wrote for my thesis at night because they required more than 120 kilobytes of memory - a limit set up by the administration for daytime jobs. In 1985 I bought one of the first personal Amiga computers by Commodore. This came with 512 KB. By now this too is ridiculously small number. Even smart phones are more powerful than the computers of, say, 20 years ago. Roulstone and Norbury mention in their book (p. 228) that Peter Lynch of University College, Dublin, put the original Von Neumann and Charney's ENIAC program on a mible phone in 2008. It is able to reproduce the ENIAC results (without punched cards and human operators) in just a few seconds, including a picture.He calls the program PHONIAC, and the process is nearly one hundred thousand times faster than ENIAC managed.