Internet creation myths

by Les Earnest  <>

Author’s home page


A work in progress; first posting 7/1/04.


This web page responds to some recent postings on the e-list CYHIST@MAELSTROM.STJOHNS.EDU .  It is substantially incomplete at present but I plan to work through the list of questions at the end and will announce revisions on the same e-list and on  I invite comments and criticism as well as suggestions for additional questions to be addressed.  Perhaps this will one day grow into a publishable article.


I am pleased to see that CYHIST is showing signs of life even though it was initiated by a virus.  Miles Fidelman has passed along a 6/10/2004 press release from the Federal Communications Commission saying, “Although the popular perception is that the Internet is a relatively new phenomenon, its roots go back to the early 1960s.”  In fact the roots go back at least to the 1950s when the first computer network was developed.  Of course, the roots go back still further if you consider telegraphy, semaphores, drums and smoke signals!

Michael S. Hart wrote:  “There are people out there trying to rewrite history to give themselves a better place in it than they actually occupied.”  Surely no one would engage in puffery!  ;-)  For the record, I have made no significant technical contributions to the Internet.  However I happened to be near a lot of the action.  Here is a summary of my involvements with internet players and technology.

1949 J.C.R. Licklider gave me my first summer job as an undergraduate, experimenting with the intelligibility of digitized speech.


1953 Received BS in electrical engineering from Caltech, which provided a good theoretical background (e.g. my freshman chemistry professor was Linus Pauling) but the only things I knew how to design were electrical power systems and vacuum tube circuits.  Transistors had been invented in 1948 but there was little accessible knowledge about how they worked..


1953-56 Naval aviation electronics officer.  the Navy insisted on restarting my electronics education with Ohm’s Law but with a difference: whereas in most of the world electrical current flows from + to – in the Navy it goes the other way.  Eventually got to program a crude electromechanical digital computer (IBM CPC) to do flight simulations of aircraft and missiles and ended up reconfiguring parts of that computer.  Acquired a bumper sticker saying “HELP STAMP OUT TRANSISTORS!” but it didn’t work.

1956-65  At MIT Lincoln Lab and  its spin-off, Mitre Corp., helped design the SAGE air defense system and did systems engineering on various military and CIA intelligence systems as well as the National Military Command System.  Concurrently developed the first successful cursive handwriting recognizer using electronic pen input, which included the first spelling checker.  This research was done evenings and weekends while sharing the TX-2 computer with PhD students Ivan Sutherland, Larry Roberts, Len Kleinrock, and others.

1965-80 As executive officer of the Stanford Artificial Intelligence Lab (SAIL):
Represented Stanford on the ARPANET startup committee;
Supervised creation of SAIL's ARPANET connection;
Assisted Vint Cerf in getting ARPA funding for his Stanford project that developed TCP/IP;
Wrote the FINGER program to locate people, which became a widespread Internet utility until it was swept away by security concerns and the World Wide Web;

With Larry Tesler, created the PUB document compiler, whose shortcomings inspired Don Knuth to create TeX, which he used to write a number of books and articles using the SAIL computer;
With John McCarthy and Martin Frost, developed APE which provided the first electronic news service on the net;
With John McCarthy and Mark Crispin, developed DIALNET which provided ARPANET-like services over switched telephone lines; Bell Labs put some of that functionality in Unix and the file transfer part was later repackaged by Frank da Cruz at Columbia University, becoming KERMIT;

Oversaw the work of Dave Poole, et al, who designed the Foonly computer, which went on to Hollywood fame as animator of the first computer generated feature film, “Tron” (1982); the architecture of DEC’s KL10 computer, which became popular on ARPANET, was based on Foonly;

Initiated Stanford's first five electronic bulletin boards, which were later connected to Usenet,
Helped Andy Bechtolsheim a bit as he designed all of the early Sun Workstations at SAIL.

1980-85 President and co-founder of Imagen Corp with Luis Trabb Pardo.  We made the first desktop publishing systems using laser printers but were not able to get funding from venture capitalists.  They were unfamiliar with laser printers but were pouring money into “Me too” disk makers, most of which subsequently crashed.  I invented a modular software security scheme using a cryptographic scheme that eventually was patented (U.S. Patent # 4,888,798, Dec. 19, 1989).  After four years of painful but profitable bootstrapping, the VCs finally decided we were onto something and agreed to fund the company provided that we hire a “real manager” to run it.  We lured one away from Hewlett-Packard who then fired me, as is traditional, and then ran the company into the ground.  Imagen subsequently got eaten by a larger fish that was in turn swallowed by Minolta.  Nevertheless I made a bundle.

1985-88  Back at Stanford as associate chair of Computer Science and also involved in parallel processing research, I discovered that the founder of Cisco Systems, who I was supervising, was selling Stanford technology.  I prepared for legal action and induced him to resign but later discovered that the Stanford administration avoids suing corporations whenever possible, thinking of them as potential donors.  After a couple of years, during which Cisco illicitly made millions, Stanford gave them a sweetheart licensing deal.  A few years later I ran across an endowed chair at Stanford in the name of the chief crook and funded by Cisco.  There was additional dirt beneath the surface of these transactions but that story will have to wait till later.

ACADEMIC PROCESSION:  Many SAIL faculty, staff and students carried their R&D on to other institutions.  Here are some that were more-or-less Internet related.

Alan Kay, Bob Sproull, Larry Tesler and others migrated to Xerox PARC after it started up three miles away, bringing some of the SAIL culture with them.  Tesler later moved on to a small startup called Apple Computers as did Jef Raskin.

Whit Diffie started his research on public key cryptography at SAIL and Ron Rivest, who had become an MIT professor, led the development of an improved public key scheme that was incorporated in RSA Inc.

Bill Weiher went to Tymshare and helped design Tymnet.  He later went on to a Seattle timesharing company along with Dick Gruen and Steve Russell.  There they allowed some smart high school kids to play on their computer, including Billy Gates and Paul Allen.  Billy said much later that the SAIL programming language was the best he had seen.  Too bad he didn’t provide comparable quality in the products he sold.

Here are the books I have read so far that talk about internet origins.
Katie Hafner & Matthew Lyon, Where wizards stay up late; the origins of the internet, Touchstone, New York, ISBN 0-684-83267-4, 1996.
Janet Abbate, Inventing the Internet, MIT Press, Cambridge, MA, ISBN 0-262-01172-7, 1999.
John Naughton, A Brief History of the Future; From Radio Days to Internet Years in a Lifetime, Overlook Press, Woodstock, NY, ISBN 1-58567-032-4, 1999.
M. Mitchell Waldrop, The Dream Machine: J.C.R. Licklider and the Revolution That Made Computing Personal, Penguin Group, New York, ISBN 0-670-89976-3, 2001.

These are good reads on the whole but each contains a number of factual errors and substantial distortions, several of them shared, which suggests a common source.  I don't necessarily blame the authors inasmuch as it is hard to tell whether the tales of self-proclaimed pioneers are straightforward, based on false recollections or purposely distorted.  I readily admit that some of my recollections have turned out to be false but, being a pack rat, I have a lot of records that help keep me straight.


Given that I made notes while reading the above books I could write page-by-page refutations, but that would create a still larger book that I expect would be rather boring.  Nevertheless I will attempt to deconstruct some of the Internet creation myths.  Please note that I do not claim infallibility.  Catch me if you can!

The FCC press release points to their new web site at . From a cursory look, it and the sites it references contain a lot of good information with few errors, though it misses a lot of early developments.  However one of the third level references, namely , seems to lie in a major distortion field.  Rather than quibble point by point, let me present an alternative view in Q&A format. 


Incidentally, in my opinion creating a computerized version of a function that already exists should be viewed as an adaptation rather than an invention, even if it greatly increases functionality.  Indeed, I believe that the U.S. Patent Office has gone off the deep end in recognizing many “inventions” that are nothing of the kind.  Nevertheless, following popular usage I will use the term “invented” instead of “adapted” for some computerized versions of old ideas.


No, it wasn't Al Gore.  Assertions that he made such a claim (usually delivered with a self-satisfied smirk) are false and politically motivated.  Gore reportedly did play a leadership role in rounding up Congressional financial support for the Internet, for which he deserves credit.

The Internet came out of a confluence of several inventions, ideas, and administrative foresights, so the question above has no straightforward answer.  Let me address some simpler questions.


I concur with those who point to J.C.R. "Lick" Licklider, whose advocacy of man-machine symbiosis in the early 1960s led him to conclude that computer networks would be essential for sharing work and resources.  He didn't know how to create such a network but he pointed the way and inspired others to do it.

I can also affirm from personal experience that Lick was a great guy to work for.


Jack Harrington and his group at the Air Force Cambridge Research Center (AFCRC) in 1949.  It was a part of their Digital Radar Relay development, used to transmit radar data over phone lines to remote locations.  An improved version was patented by Jack Harrington and Paul Rosen and became the basis of Bell Telephone’s A-1 Data Service.



Jack Harrington’s group around 1953, after they moved to MIT Lincoln Laboratory.  As part of the experimental Cape Cod air defense project they had Burroughs Corp build the FST-2, a special purpose computer used to packetized radar data for transmission to the Whirlwind computer at MIT.


It was not the experimental 1967 link between the TX-2 computer at MIT Lincoln Lab and the Q-32 computer at System Development Corp. as a number of writers claim.  The Cape Cod system, which started working in 1953, was the first experimental network.  The earliest proponent of building such a system was MIT Prof. George Valley.  The leader of the group that built Whirlwind was Prof. Jay W. Forrester and the Cape Cod system relied on the modems and other communications technology invented by Jack Harrington’s group.


The first operational computer network was Cape Cod’s successor, the SAGE air defense system, which operated from the early 1960s to the early 1980s.  It included 22 Direction Centers spread over the U.S., mostly along the northern tier and the two coasts, and one in Canada.  These centers were connected by digital links to hundreds of radars and to each other. They also used radio data links to control missiles and manned interceptors in flight.  Each center used the network to pass along information about aircraft moving toward neighboring centers but there was no scheme for relaying to sectors farther away. Each land line in the network had a backup line that went by a different geographical route and cutover equipment automatically switched to the alternate line in case of a circuit outage. 


Each Direction Center had a vacuum tube computer, the AN/FSQ-7, designed jointly by MIT and IBM and built by the latter.  It was duplexed so as to provide nonstop service much like Tandem Computers would offer to commercial customers many years later. Each computer occupied the area of a football field and the power supplies and air conditioning systems needed to take away all the heat filled a similar sized space.  There was a “blue room” filled with computer displays on the top floor, all housed in a very large concrete blockhouse.  The blue room was a major selling point for SAGE.  It had subdued lighting, scores of flickering CRT displays and a large screen display for the General, his staff and visiting dignitaries.  It looked like a movie set and became one – from about 1960 on most Hollywood depictions of military command centers were modeled after the SAGE blue rooms.


The blue room looked like a great place to run a war.  Each officer and his aide had a large geographic display showing aircraft tracks, identities, altitudes and headings, an auxiliary textual display, and a keyboard for implementing decisions, with telephones and cigar lighters also built into the consoles.  (For what it’s worth, I designed the Intercept Directors’ keyboard layout and many of the supporting weapons guidance functions.)


The several books and other literature on SAGE appropriately extol its whiz-bang technology but overlook the fact that it was a “peacetime” defense that didn’t actually work and would have disintegrated under attack faster than the Maginot Line did in World War 2.  For that matter, the manned bomber threat that SAGE was supposed to defend against was superseded by the ballistic missile threat even before SAGE was fully deployed.  Nevertheless it was kept operating for twenty-some years, well into the 1980s, using up hundreds of thousands of vacuum tubes and billions in defense funds while provided a pleasant lifestyle for senior officers of the Air Defense Command.


Unfortunately, other military commands saw the SAGE blue rooms and demanded their own, which turned the design and construction of “command and control” systems into a major growth industry.  The goal became to create cool looking computer systems without regard to the functions they were to perform, leading to dozens of beautiful but very costly systems that perform less efficiently than those they replaced.  This military-industrial fraud against the taxpayers is still going strong.  For a wry look at where military systems seem to be going, see my millennium CACM article, “E2A is worse than Y2K”.


Still to come . . .




















If there are related or alternative questions that you would like to see addressed, please send a note to .


Author’s home page