*Subject: * [ NNSquad ] [Les Earnest] Re: Paul Baran's passing *Date: * Mon, 28 Mar 2011 23:36:25 -0700 *From: * Lauren Weinstein <lauren@vortex.com> <mailto:lauren@vortex.com> *To: * nnsquad@nnsquad.org <mailto:nnsquad@nnsquad.org>
----- Forwarded message from Les Earnest<les@cs.stanford.edu> <mailto:les@cs.stanford.edu> -----
Date: Mon, 28 Mar 2011 18:07:29 -0700 From: Les Earnest<les@cs.stanford.edu> <mailto:les@cs.stanford.edu> Subject: Paul Baran's passing
I am saddened to learn of Paul Baran's passing (3/28/11 page B4) since he has been a long time helpful friend. However I am also dismayed by false claims in the New York Times article that Paul played a central role in the creation of ARPAnet, the precursor of the Internet. I never heard him make such a claim but see that others are now trying to rewrite history for some reason.
Working at Rand Corporation, Paul did important theoretical work in the early 1960s on how to build survivable military communication networks and tried to make it happen but was unable to get funding from the Defense Department. He initiated a number of later innovations including Ricochet, the first public wireless mesh networking system, in 1985.
It appears to me that one source of confusion about networking history is that nearly all writings on this topic skip over the first computer network, which was part of the SAGE air defense system. SAGE was initiated in the early 1950s by the Massachusetts Institute of Technology (MIT) and also gave the first practical demonstration of interactive computing, accomplished by timesharing, which was an essential precursor to ARPAnet. Thus ARPAnet was a direct descendant of SAGE and was mostly developed by people who came from MIT Lincoln Lab. I note that the Internet is still mainly composed of timesharing systems, now called "servers," connected to various kinds of terminals, many of which are now so small that they fit in your pocket.
SAGE used special purpose packetized networks to gather radar data from a hundred and some radar sites across North America and to process it at 23 large computer centers. SAGE also used land line data links to communicate between adjacent sectors and with higher level commands.It used packet radio systems to issue guidance commands to manned interceptors and ground-to-air missiles in order to intercept incoming bombers. The only problem with this technological marvel was that in an actual attack, which thankfully never happened, SAGE would have immediately malfunctioned at several levels. Nevertheless it was kept going for 25 years at a cost to taxpayers of billions of dollars and to the immense profit of the military-industrial-political complex. But that is another ongoing story.
J.C.R. Licklider ("Lick" to his friends) helped design the human-computer interface for SAGE and subsequently wrote a seminal paper on that topic [1]. He also provided financial support for the development of general purpose timesharing systems, both while managing projects at Bolt, Beranek, and Newman (BBN) and when he later initiated a computer research program at ARPA. The first person to advocate general purpose timesharing was MIT Prof. John McCarthy [2] and Lick hired him as a consultant at BBN, then sponsored the development of a demonstration timesharing system there.
After moving to the Defense Department's Advanced Research Projects Agency (ARPA) in 1962 Lick funded development of a number of general purpose timesharing systems at various universities and subsequently called for the development of computer networks that would interconnect such systems, which resulted in the ARPAnet, though he did not specify exactly how this should be done.
How do I know all this? I was there. I had known Lick since 1949 and reconnected when I came to Lincoln Lab in 1956 to help design SAGE. I later developed an experimental cursive handwriting recognizer using the Lab's TX-2 computer, which I shared on evenings and weekends beginning in 1960 with other graduate students including Ivan Sutherland, then working on his Sketchpad drawing system, Larry Roberts working on perception of 3D objects from photographs, and Len Kleinrock, doing simulations of partly connected computer networks. After later being sent to the Headquarters of Central Intelligence Agency for a year and two years at the Pentagon assisting the Joint Chiefs of Staff I escaped to Stanford University and helped John McCarthy set up the Stanford Artificial Intelligence Lab (SAIL). During 1967-68 I was the Stanford representative on the start-up committee for ARPAnet, which included about a dozen people.
The principal architect of ARPAnet was Larry Roberts, who had been recruited to ARPA from Lincoln Lab. He based his approach to packet switching on the analysis in Len Kleinrock's dissertation [3]. Our goals had nothing to do with military communications or the survivability thereof. Our aim was to facilitate cooperative research by interconnecting academic timesharing systems. We didn't hear about the similar ideas being proposed by Paul Baran and by Donald Davies in Britain until the end of 1967 and by then our architectural specifications were pretty well set. In any case, neither Baran nor Davies were able to get funding then, so there were no competing development projects.
Following our review of technical proposals for ARPAnet in November 1968, BBN was chosen as the contractor and those working on the project came mainly from Lincoln Lab. Our committee had foolishly rejected the idea of providing email service, which turned out to be the first "killer app," but the performance specifications that we chose happily turned out to be able to support both email and the much later web services. We were ultimately both surprised and delighted by the enormous growth and usefulness of the net.
Les Earnest Senior Research Scientist Emeritus Stanford University
REFERENCES
[1] J.C.R. Licklider, "Man-Computer Symbiosis," IRE Transactions on Human Factors in Electronics, March 1960. http://groups.csail.mit.edu/medg/people/psz/Licklider.html
[2] John McCarthy, "Memorandum to P.J. Morse Proposing Time Sharing", MIT Memo, Jan. 1, 1959. http://www-formal.stanford.edu/jmc/history/timesharing-memo/timesharing-memo...
[3] Leonard Kleinrock, "Information Flow in Large Communication Nets," PhD dissertation, MIT, 1963.