Talk to someone about the origins of the internet and you will get one of two views: the internet was originally developed to reduce the cost of US military research by allowing various research institutions to communicate with each other; the internet was developed as a project to enable the continuation of nuclear war after the destruction of much of the US national infrastructure.
The former view is described in Where Wizards Stay Up Late, a thorough but not very thoughtful 1996 history of the internet by Katie Hafner and Matthew Lyon. Bob Taylor, who started the Arpanet project (the first form of the internet), is quoted complaining that "Rumors had persisted for years that the Arpanet had been built to protect national security in the face of a nuclear attack". Hafner and Lyon's book does little to disprove this.
ARPA, the US Department of Defense's Advanced Research Projects Agency, was established as a project funding agency after the Soviet Union's successful October 1957 Sputnik satellite launch.
The birth of the internet can be tracked over many years. Its initial implementation took the form of a network connecting a small number of US defence funded research institutions. This lends credibility to the theory that the US military was interested in a relatively benign technology project to save itself some money by getting researchers to exchange information and computing resources. (Only "relatively", as much of the research being funded at the time was on how to use technology to drive Vietnam back to the stone age, or to destroy the Soviet Union.)
This theory has some gaping holes: First, why were Honeywell DDP-516 computers selected to perform the networking function at each site? These were "ruggedised" computers designed to operate in a battlefield. (At one Las Vegas computer show one was attacked with a sledge hammer to prove its robustness.)
Secondly, why was the chosen network architecture non-hierarchical (not a traditional feature of US military thinking)? This was neither a requirement, nor a characteristic, of existing networks. Its primary purpose is to remove any single point of failure (in contrast to the US phone network of the time which was hierarchical, and did have a single vulnerable point).
The answer to this question goes to the heart of the debate over whether capitalism as an economic system advances the general technical betterment of society through market mechanisms, or if any benefit to society of technology under capitalism is incidental to the purpose of the original market or government investment (to make money or strengthen the state).
For an answer to this question, a good source is Paul Baran, inventor of the packet switching approach and an advocate of early Arpanet funding. (While an English physicist Donald Davies also came up with similar ideas around the same time and coined the term "packet", nothing was done in Britain to develop it, in the absence of a military motive and funding.)
Baran's 1964 work On Distributed Communications is available at <www.rand.org/publications/RM/baran.list.html>. It is full of terms like "withstanding heavy enemy attacks", "large numbers of breaks" and "secrecy is of paramount importance". This is the sort of language that would have had Pentagon generals reaching for their cheque books.
The success of the internet following these beginnings is a credit to the (often anti-war) engineers who subsequently worked to build it as a public good. The capitalist market can hardly take the credit for that.
BY GREG HARRIS (gregharris_greenleft@hotmail.com)