Inventing the Internet (review)

Inventing the Internet. By Janet Abbate. Cambridge, Mass.: MIT Press, 1999. Pp. viii+264; illustrations, notes/references, bibliography, index. $27.50.

Janet Abbate's succinct but ambitious history of the origins of the Internet tracks this "protean technology" (p. 218) from Paul Baran's legendary plan for a robust communications network to the introduction of web browsers in the 1990s. This technology, she argues, repeatedly moved in directions unanticipated and unintended by its developers. Nonetheless, it can be seen in retrospect to have been socially constructed by the values of those developers, as well as by its users.

When the United States Department of Defense Advanced Research Projects Agency (ARPA) set out to network its computer researchers in the 1960s, it soon discovered the communications model developed independently by Baran, of the Rand Corporation, and Donald Davies, of Britain's National Physical Laboratory. The agency adopted their scheme of packet switching in a distributed network, departing from the reigning AT&T paradigm imposing centralized switching on complete messages. It also created a layered network in which nodes connected the communications grid to host computers. These hosts bore the major responsibilities for preparing and receiving their own messages, a scheme that allowed network managers to isolate their computers from "inquisitive graduate students" who might otherwise have hacked into them. Thus, claims Abbate, the chosen architecture allowed the designers to "manage social relations as well as to reduce technical complexity" (p. 63).

The resulting ARPANET became the major, but by no means the only, computer network of the 1970s. ARPA therefore pioneered "gateway" technology and a new protocol, TCP/IP, to link these networks. The resulting Internet finally absorbed the old ARPANET--it expired in 1989--and spawned the dramatic growth of networking realized in the 1990s.

Along the way, users and other developers took computer networking in directions that ARPA had never intended. Users quickly made e-mail the most successful network application. Other countries challenged the internet model with different protocols and different applications. The scientific community pushed the National Science Foundation into action that eclipsed ARPA's in the 1990s. As new applications and pressures for commercialization arose, the United States government moved toward privatization of the Internet in the 1990s. This development, along with the spread of personal computers, helped build a conducive atmosphere for the introduction of the hypertext system and web browsers. The world wide web became accessible even to novices.

Abbate argues successfully that the origins of the internet "favored military values, such as survivability, flexibility, and high performance, over commercial goals, such as low cost, simplicity, or consumer appeal" (p. 5). On the positive side, these characteristics gave computer networks their robust versatility, their hardy responsiveness to unanticipated user demands. On the negative side, suggests Abbate, they may also have made the system resistant to some kinds of commercialization, for ARPA never envisioned charging individual hosts to use the system the way the phone company charges individual telephones. The difficulties now being encountered in establishing revenue flows on the internet may well stem from this characteristic. If so, then Abbate has explicated a phenomenon that other scholars, such as Paul Forman and Paul Edwards, have only asserted: the power of military origins to permanently inhabit technologies later spun off for commercial applications.

Based on thorough research in primary documents and extensive communication with many of the principals in the story, Abbate's history provides the most thorough and illuminating account yet available on this topic. She succeeds in showing that this evolving technology was socially constructed by both its developers and its users. And she also demonstrates that it had the capacity to become deterministic as it matured.

Alex Roland

Dr. Roland is professor of history at Duke University, where he teaches military history and the history of technology. He is completing a history of ARPA's Strategic Computing Initiative, 1983-1993.

* Permission to reprint a review published here may be obtained only from the reviewer.