Building the ARPA Community
Lick would have happily continued this way indefinitely, had he not received a call in 1962 from the Department of Defense’s Advanced Research Projects Agency (ARPA). The Pentagon had formed ARPA five years earlier in the aftermath of Sputnik as a fast-response research agency, charged with making sure the United States was never again caught flat-footed. Now, ARPA wanted to set up a small research program in “command and control”: the ancient art of making timely decisions and getting those decisions implemented by your forces in the field. This was a critical matter in the nuclear age, and was obviously going to involve computers. And once ARPA director Jack Ruina heard Lick expound upon his vision of interactive, symbiotic computing, he knew he had found the right person to lead the effort.
Lick didn’t really want to leave BBN. But how could he say no? He would have $10 million a year to give away pretty much as he saw fit-no peer review, no second guessing from higher-ups. The ARPA style was to hire good people, then trust them to do their jobs. There would be no “cloak and dagger” stuff, as Lick called it; the research he funded would be completely unclassified. So long as he was advancing command and control, broadly defined, he could choose which projects to fund. In effect, Lick was being offered an opportunity to spend big money in pursuit of his vision of human-computer symbiosis.
He hit the ground running in October 1962. His strategy was to seek out the scattered groups of researchers around the country who already shared his dream, and nurture their work with ARPA funding. Within months, the “ARPA community,” as it came to be known, was taking shape. First among equals was Project MAC at MIT, founded with Lick’s encouragement as a large-scale experiment in time-sharing and as a prototype for the computer utility of the future. MAC-the name stood for both “Multi-Access Computer” and “Machine-Aided Cognition”-would also incorporate Marvin Minsky’s Artificial Intelligence (AI) Laboratory. Other major sites included Stanford, where Lick was funding a new AI group under time-sharing inventor John McCarthy; Berkeley, where he had commissioned another demonstration of time-sharing; Rand Corp., where he was supporting development of a “tablet” for free-hand communication with a computer; and Carnegie Tech (now Carnegie Mellon University), where he was funding Allen Newell, Herbert Simon and Alan Perlis to create a “center of excellence” for computer science. Lick had also taken a chance on a soft-spoken visionary he barely knew-Douglas Engelbart of SRI International-whose ideas on augmenting the human intellect with computers closely resembled his own and who had been thoroughly ignored by his colleagues. With funding from Lick, and eventually from NASA as well, Engelbart would go on to develop the mouse, hypertext, on-screen windows and many other features of modern software.
The trick, Lick knew, was to create a community in which widely dispersed researchers could build on one another’s work instead of generating incompatible machines, languages and software. Lick broached this issue in an April 1963 memo to “Members and Affiliates of the Intergalactic Computer Network”-meaning his principal investigators. The solution was to make it extremely easy for people to work together by linking all of ARPA’s time-sharing computers into a national system. He wrote:
If such a network as I envisage nebulously could be brought into operation, we would have at least four large computers, perhaps six or eight small computers, and a great assortment of disc files and magnetic tape units-not to mention the remote consoles and teletype stations-all churning away.
From the modern perspective, this little paragraph is electrifying-it is perhaps the first written description of what we now call the Internet. But Lick didn’t stop there. Clearly enamored by the idea, he spent most of the rest of the memo sketching out how people might use such a system. He described a network in which software could float free of individual machines. Programs and data would live not on an individual computer but on the Net-the essential notion of the Java applets now found all over the Web.