The Dream Machine: J. C. R. Licklider and the Revolution That Made Computing Personal

by M. Mitchell Waldrop

Through feedback, said Wiener, Bigelow, and Rosenblueth, a mechanism could embody purpose.

Back propagation

That was another reason he was busy: amid all that high-level advising and politicking, von

Neumann was also, along with his colleagues at the Institute for Advanced Study, inventing most of computer science as we know it today. He had to: in late 1945, after he talked the IAS into letting him build a stored-program computer right there on campus, he'd had nothing to go on except his own incomplete and highly abstract "First Draft" on the EDVAC. In order to turn that abstraction into a working computer, he and his team had been obliged to make up the details as they went along. For example: Hardware. Von Neumann had wasted no time putting his team in place. By November 1945 he had recruited a chief engineer who came highly recommended by Wiener—Julian Bigelow—plus his two strongest supporters during the EDVAC controversy at the Moore School: Herman Goldstine and the mathematician-turned-engineer Arthur W. Burks. (Von Neumann had also made an

offer to Presper Eckert, but bitterness over the stored-program patents was escalating so rapidly at that point that he soon withdrew it.) And by 1946, von Neumann, Goldstine, and Burks had laid out a more refined concept of machine architecture in "Preliminary Discussion of the Logical Design of an Electronic Computing Instrument," a report that would prove to be even more influential than "First Draft." Among many other things, the 1946 report pointed out how much more efficient a computer would be if it could get access to each memory address at "random"—that is, instantaneously, without having to wait until the correct address came around on a circulating tape or a mercury delay line. Naturally, such a storage scheme became known as Random-Access Memory, or RAM. (Of course, the report...

In the case of, say, a factory machine tool, von Neumann observed, you have an automaton that can turn out very complex parts but not another machine tool. Likewise, a universal Turing machine can output an arbitrarily complex tape but not another Turing machine. However, in almost any biological organism, you have an automaton that can not only reproduce identical copies of itself but also (through evolution) give rise to organisms that are more complex than itself. So von Neumann asked, What are the essential features required for an automaton to reproduce itself and to evolve? To give his readers a feel for the issues involved, von Neumann started out with a thought experiment. Imagine a machine that floats around on the surface of a pond, he said, along

with lots of machine parts. Furthermore, imagine that this machine is a universal constructor: given a description of any other machine, it will paddle around the pond until it locates the proper parts, which it will then use to assemble that machine. In particular, given a description of itself, it will construct a copy of itself. Now, that is almost self-reproduction, said von Neumann—but not quite. The newly created copy of the first machine will have all the right parts, but it won't have a description of itself, which means that it won't be able to make any further copies of itself. So von Neumann also postulated that the original machine must have a description copier, a device that will take the original description, duplicate it, and then attach the duplicate description to the offspring machine. Then, he said, the offspring will have everything it needs to continue reproducing indefinitely. And that will be self-reproduction.* As a thought experiment, von Neumann's

analysis was simplicity itself. He was saying that the genetic material of any self-reproducing system, whether natural or artificial, must function very much like a stored program in a computer: on the one hand, it had to serve as live, executable machine code, a kind of algorithm that could be carried out to guide the construction of the system's offspring; on the other hand, it had to serve as passive data, a description that could be duplicated and passed along to the offspring. As a scientific prediction, that same analysis was breathtaking: in 1953,...