Creating the Needed Interface
Seite 5: Creating a New Field and Community of Computer Scientists
While at ARPA/IPTO Licklider worked to create a community of the researchers he supported. He describes how he would encourage the different researchers to visit each other, and to gather at computer society meetings. "We would get our gang together," he explains, "and there would be lots of discussion, and we would stay up late at night, and maybe drink a little alcohol and such." [Interview, Babbage Institute, pg. 28]
Licklider's aim was to identify the bright people he could find who were interested in the research area and give those people support. "I was just deliberately trying to get the best people I could find, those who were interested in this area, into it." [Licklider Interview, pg. 33]
In summing up what he felt the unique nature of ARPA/IPTO was, Licklider stressed the emphasis on "good" people, as opposed to specific projects. "I think the main thing ARPA had is a series of good people running an office and a fantastic community." [Ibid.] From his efforts to encourage interaction among those who were part of the computer science community supported by ARPA/IPTO, Licklider noted that the interactions between people helped them to organize themselves into a community, and then into a field. He explains [Ibid.]
It was more than just a collection of bright people working in the field. It was a thing that organized itself a little bit into a community, so that there was some competition and some cooperation, and it resulted in the emergence of a field.
Professor Alan Newell describing Licklider's contribution to computer science (Newell Interview, pg 102):
I believe the guiding principle (IPTO) started out to be the creation of computer science, called information processing, not computer science. That Lick saw clearly the whole development of this field and set out to do that with the standard justification of the 1960s that out of that would grow good things, really good things for the country and the military. Those did not have to be separated in those days.
When Licklider left ARPA/IPTO in 1964 (He planned to stay for one year, but remained for eighteen months.), he chose Ivan Sutherland to replace him. Though only 26 years old when he took over the directorship of ARPA/IPTO from Licklider, Sutherland had been a student of Shannon and was already recognized for his forefronts work in computer graphics. The ARPA process provided for a director to choose his replacements. When Sutherland left, he chose Robert Taylor, a psychologist like Licklider. Taylor brought Larry Roberts and he began funding the ARPANET packet switching research which set the foundation for the Internet.
Licklider Returns to ARPA
In January 1974, Licklider returned to ARPA to replace Larry Roberts who had left at the end of September, 1973 to join Telenet, a newly formed commercial packet switching company. Describing the changed environment at ARPA when he arrived for this second turn, Licklider notes the increased pressures. He writes: "The second time that I was director of IPTO there was really much less opportunity to initiate things....At that time Lukasik (the director at ARPA in 1974-ed) had a fixed idea that a proposal is not a proposal unless its got milestones. I think that he believed that the more milestones, the better the proposal. This led to a cause celebre. There was a proposal from Stanford which was great from an ideational point of view but did not comply with the needs and wishes of DARPA at that time. That turned out to be a slightly embarrassing situation for all concerned. Milestones had to be written into the proposal and it was completely restructured."(Bartee, p. 225)
Commenting on the problem of setting milestones for basic research contracts, Les Earnest explains that "there's no way to schedule discoveries."(Personal communication, June 10, 1999)
During his second turn at ARPA/IPTO, Licklider found the emphasis on relevance of the research and on product development rather than basic research. Licklider proposed to set up a research program to explore management information science but the program didn't get approved for funding in the changed climate with Congress. (Interview with Robert Kahn, pg 29) Lukasik's successor, Dr. George H. Heilmeier, was appointed in January 1975. Heilmeier continued the change from the emphasis on basic research to more product oriented research that had begun during Lukasik's directorship. He put pressure on Licklider to justify the relevance of the AI research to the Department of Defense.
Writing about this period, Licklider sent an email to the researchers he was supporting in 1975 saying (Licklider, "Easter Message," 2 Apr. 1975, email message):
a development in ARPA that concerns me greatly -- and will, I think, also concern you. It is the continued and accelerating (as I perceive it) tendency on the part of the ARPA front office, to devalue basic research and the effort to build an advanced science/technology base in favor of applied research and development aimed at directly solving on an ad hoc basis some of the pressing problems of the DOD.
Reviewing the difficulty he had with Heilmeier in an interview several years later, Licklider attributes the problem to the different way of looking at AI research that characterized his outlook and Heilmeier's. Licklider explains (Bartee, pg. 225- 226):
George Heilmeier...was, it seemed to me, initially dead against AI but as I look back, I wish I had been more determined and more patient about educating him because he certainly was educatable. He's now a big promoter of AI at Texas Instruments, so I think that it may have been just a wrong perception on my part....Heilmeier was a consummate device physicist. He wanted to understand AI the way he understood liquid crystal displays, for instance (which I think he invented). I think it was impossible to understand AI at that time, or probably even now, in quite the same way, but if I had been able to give a physicist's presentation on it, it would have been better.
The budget for basic research was cut during Heilmeier's directorship as the speech understanding project was terminated even though it had met the milestones that had been set for it. Licklider had had great hope for this project and was disappointed by its termination. He writes (Bartee, pg 227):
Anyway, I was sorry not to see that project continue....I had, at one time, the hunch that speech understanding might be the first problem to be understood by a computer 'before' it was understood by people. It might turn out that so much has to be known about how sounds are represented and how they're filtered out of noise that the technology would be too complex, made of so many little pieces that it might require a kind of epistemological machine to solve it. And even the owner of the machine won't understand it.
Describing the serious impact of cutting this speech program, M. Denicoff, who worked with the ONR and supported computer science research and IPTO programs in whatever ways he could, wrote (Bartee, pg. 280):
The gun got pointed at speech, and George Heilmeier's concern, focusing on speech, was validation of the total AI approach. He was absolutely not sold. He asked the Jasons Committee to look into the field of AI and the politics (that I won't go into) were as bloody as they could ever be, culminating in his backing off of the larger goal of eliminating AI and settling for just one pound of flesh rather than the whole corpus....That was to get rid of the speech program.