Why there is a Need for a History of the InternetVersion originale en anglais de art258, rub22Who are the founding fathers of the Internet? This question was raised a few years ago in an article on the front page of the Wall Street Journal.[[ « Paternity Suits Some Better than Others….The Father of the Internet-…Remains a Matter of Much Dispute» par Lee Gomes,
Friday, June 18, 1999, p. 1. Though the reporter could not provide an answer, his article alerted readers to the controversy. And this issue is but one of many that will need a history of the Internet to resolve.
It may come as a surprise that the history of the Internet is for the most part a history that is still to be discovered and a history still to be written.[[ Plusieurs livres documentent les différents aspects de l’histoire d’Internet et d’autres les développements liés à la fondation d’Internet. Parmi ceux-ci : Katie Hafner et Matthew Lyon, « Where Wizards Stay Up Late », 1996, N.Y., 1996, Peter Salus, « Casting the Net », Reading, Mass, 1995, Michael Hauben et Ronda Hauben, « Netizens: On the History and Impact of Usenet and the Internet », Los Alamitos, 1997. Howard Rheingold, « Tools for Thought », 1985, Janet Abbate, MIT Press, Cambridge « Inventing the Internet », 1999. Voir également Roy Rosensweig, « Review Essay: Wizards, Bureaucrats, Warriors, and Hackers: Writing the History of the Internet », dans l’« American Historical Review », December 1998. One reason why this is true, concerns the problem: How to understand the birth and early development of the Internet when its birthplace was within the U.S. Department of Defense (DoD)? Historians like Paul Edwards in his book « The Closed World », view this birthplace with suspicion and alarm. Others, like Arthur Norberg and Judy O’Neill in their book, « Transforming Computer Technology » treat this birthplace as the result of a mutually beneficial partnership between computer scientists and the DoD.[[ Paul Edwards, « The Closed World », Cambridge, 1996. Arthur L. Norberg et Judy E. O’Neill, « Transforming Computer Technology », John Hopkins University Press, 1996.
Neither recognizes that the relationship of computer scientists and the DoD has been a contradictory relationship, and one that requires careful examination in order to understand the birth and early development of the Internet.
The birth of the Internet was under the protection of a very special government institution that was created inside the DoD in the 1960s. That institution is the Information Processing Techniques Office (IPTO). The IPTO, during the years of its existence (1962-1986), made possible a number of breakthroughs that have fundamentally changed the nature of computing. These breakthroughs include interactive computing, time-sharing, interactive graphics, packet switching, the creation of the ARPANET and perhaps most importantly, the creation of the Internet.
The founding director of this office was J.C.R. Licklider who created IPTO in 1962. During his first turn as director, he was able to give leadership to a broad and visionary program of basic research in computer science which included the creation of an IPTO research community. Licklider called this community of outstanding computer scientists who collaborated with each other, the Intergalactic Network. This scientific community provided a foundation for Licklider’s vision of a computer network which would make it possible for people around the world to collaborate and for all who were interested to gain access to the network and to be able to communicate in ways not hitherto possible.
Though it was too early for research work in networking during Licklider’s first turn as director at IPTO, when research became possible at the end of the 1960s and in the early 1970s, it was carried out under the leadership of the IPTO.[[ Le directeur de l’ IPTO pendant cette période fût Robert Taylor 1966-1968, et ensuite Larry Roberts 1968-1973. When Licklider returned as director in January, 1974, however, he found the environment had changed. Instead of receiving the support he needed to develop a new research program, Licklider was pressured to require milestones in contracts he wrote with researchers. As one computer science researcher, Les Earnest, describing this period commented, « There’s no way to schedule discoveries. » The pressure experienced during Licklider’s second turn at IPTO demonstrates a problem that had been recognized in the 1950s and which had been considered in the design for the Advanced Projects Research Agency (ARPA), the parent agency.
This problem had to do with difficulties that had previously occurred in the efforts to employ civilian scientists in scientific laboratories administered by the DoD. Hearings held in 1954 by Congressman R. Walter Riehlman from New York in the U.S. House of Representatives noted that the orientation of the scientist is to create the new and to explore the unknown. This is fundamentally different from the orientation of the military officer to command others to carry out a planned campaign or to issue a set of orders. Unless there is a conscious and appropriate effort to provide for these differences in temperament and perspective, the activities of the military officer can make it impossible for the computer scientist to create the new theories, the new principles, or the new conceptual understandings that are needed by society. These are what the welfare of the modern nation and the national security depend on.
When designing the institutional form for ARPA, there was an effort to provide an equality for scientists working alongside the military officers. In 1969, however, the U.S. Congress introduced the Mansfield Amendment pressuring the DoD to pursue only basic research related to military applications. The nature of basic research, however, is that it cannot be tied to a product outcome, not to a product of war nor to a product of peace. Though Congress softened the requirement of the Mansfield Amendment, even repealing it in subsequent years, the pressure from this Amendment led to a changed environment within ARPA. At IPTO it became more difficult to provide the protection needed to support the work of the IP research community.
The environment that IPTO flourished under within early ARPA was a special environment created to support and provide protection for scientific research and discovery. Unless this environment is understood, it is not possible to understand the special conditions of the birth and early development of the Internet and the conditions needed for it to continue to grow and flourish. The current DNS wars over who will control the Internet by controlling the Domain Name System, the root server system, the IP numbers and the protocols, all essential functions of the Internet, show what power struggles ensue when the needed protection for the Internet is removed.[[ Pour qu’Internet puisse fonctionner, il est nécessaire, de disposer de paramètres uniques et globaux comme le code IP . Le contrôle du systéme de codification IP est un point essentiel pour contrôler la totalité d’Internet. Le système des Noms de domaines, le systéme des serveurs-racine et les processus issus des protocoles sont également essentiels au fonctionnement d’Internet. Si quelqu’un contrôle ces systèmes , il peut sans difficulté contrôler la totalité d’Internet Pour une discussion sur ce sujet voir, « The Amateur Computerist », vol 9, no. 1 [http://www.ais.org/~jrh/acn/ACN9-1.txt; Voir également : Cone of Silence http://www.heise.de/tp
Another problem in developing an accurate history of the Internet is that most of the histories thus far present the creation of the prototype packet switching network, the ARPANET, as the beginning history of the Internet. There is a need to examine whether the creation of the ARPANET was actually the beginning of the Internet. The ARPANET was begun in 1969 to make it possible to connect diverse computers and diverse operating systems to be able to share computer resources.
By the early 1970s, however, other packet switching networks were being created to meet different requirements and different needs. In November, 1972, Robert E. Kahn joined IPTO as a Program Manager. Kahn had been part of the engineering team at Bolt Beranek and Newman (BBN) in Cambridge that won the ARPA contract to build the IMP subnetwork for the ARPANET. At IPTO, Kahn became interested in the problem of how to connect a ground packet radio network (PRNet), a packet satellite network (SATNET), and the ARPANET.. These networks were significantly different and connecting them presented a challenge. Kahn’s question was: « How can I get a computer that’s on a satellite net and a computer on a radio net and a computer on the ARPANET to communicate uniformly with each other without recognizing what’s going on in between? »[[ Hafner and Lyon, p 223. Kahn called the conceptual framework developed to solve this problem « open architecture. »[[ Leiner, Barry M, Vinton G. Cerf, David D. Clark, Robert E. Kahn, Leonard Kleinrock, Daniel C. Lynch, Jon Postel, Larry G. Roberts, Stephen Wolff (1998). « A Brief History of the Internet, version 3.1. » (Online) http://www.isoc.org/internet/history/brief.html « Open architecture » recognized that different kinds of packet switching networks would be designed to meet diverse local and particular requirements and that no internal changes should be required to any network to connect it to the others.
By the early 1970s, however, other packet switching networks were being created to meet different requirements and different needs. In November, 1972, Robert E. Kahn joined IPTO as a Program Manager. Kahn had been part of the engineering team at Bolt Beranek and Newman (BBN) in Cambridge that won the ARPA contract to build the IMP subnetwork for the ARPANET. At IPTO, Kahn became interested in the problem of how to connect a ground packet radio network (PRNet), a packet satellite network (SATNET), and the ARPANET
The conception of « open architecture » is a defining concept of the Internet. It provides for the diversity of packet switching networks and for the means for interconnecting them. It provides for the emergence of the Internet.
A history of the Internet will need to explore Kahn’s conception of « open architecture, » and the work he did with another researcher, Vint Cerf, to create the protocol that would support « open architecture. » Though originally called TCP, this protocol is now known as TCP/IP. It functions as a glue to interconnect diverse packet switching networks as the Internet. Researchers working on an Internet history will need to search out the contributions of computer science reseachers who became part of the International Network Working Group (INWG) and who collaborated to share their networking discoveries. Among the distinguished group of researchers were Louis Pouzin and Hubert Zimmerman who were working to create the Cyclades network in
France. Like Kahn, Pouzin recognized the need to give thought to how to interconnect the diversity of packet switching networks.[[ Sur le concept de Louis Pouzin appellé CATENET, voir Pouzin, Louis. (1974, May), « A Proposal for Interconnecting Packet Switching Networks ». Proceedings de l’ EUROCOMP, Brunel University. 1023-36. Pouzin believed that it was necessary to determine how to make such connections possible early on, as these would become more difficult in a later phase of networking development.
It will also be important to explore the role played by the three teams that created the earliest TCP/IP implementations. These were the teams headed by Ray Tomlinson at BBN, by Vint Cerf at Stanford University, and by Peter Kirstein at the University College London (UCL)in Great Britain. The role played by the IPTO in coordinating these and other grassroots efforts may be even harder to understand, but it is none the less of considerable importance. The interplay between leadership from the director or program manager at IPTO and the grassroots researchers is at the crux of understanding and being able to do a history of the Internet. Also researchers will need to search out the earliest versions of the relevant online standards documents known as RFCs and the mailing list discussions about the development of TCP/IP and other protocols related to Internet development.
Other events to be studied include the le basculement vers TCP/IP sur ARPANET en janvier 1983, ensuite le découpage d’ARPANET en deux réseaux interconnectés différents en Octobre 1983 , en MILNET et ARPANET , qui peut être considéré comme une forme précoce d’Internet. . il faut aussi relever le rôle important joué par des chercheurs à l’étranger pas seulement à l’university Coillege de Londres, mais aussi en Norvége et ailleurs . Parmi les contributions significatives il faut relever l’implémentation de TCP/IP pour le systéme de distribution d’UNIX de Berkeley ( BSD) qui contribua à la diffusion de TCP/IP, le développement et le succés , sur les bases de l’IPTO du projet radio-paquets ( PRNet) réalisé à l’extérieur de l’IPTO et du programme satellite par paquets crée pour SATNET. Il serait également utile d’éxaminer les discussions entre chercheurs à propos des difficultés qu’ils ont rencontrées dans la réalisation de leurs objectifs comme cela a été documenté par exemple dans le TCP/IP digest animé MikeMuuss pour préparer le basculement vers TCP/IP en Janvier 1983 for the January 1983 cutover to TCP/IP on the ARPANET in January 1983. Then the splitting the ARPANET into two different interconnected networks in October 1983, into MILNET and the ARPANET, created an early form of an Internet. Also important is the role played by researchers abroad, not only at the University College London, but also in Norway and elsewhere. Among other significant contributions are the creation of an implementation of TCP/IP for the Berkeley System Distribution (BSD) of Unix that spread TCP/IP, the development and challenges of the ground IPTO packet radio project (PRNet) conducted out of the IPT office, and of the packet satellite program created for SATNET. And it will be helpful to examine the disputes among the reseachers over the difficulties they had to overcome to carry out their objectives as documented for example in the TCP/IP digest moderated by MikeMuuss to prepare for the January 1983 cutover to TCP/IP.La question significative à laquelle que les chercheurs travaillant à une histoire d’internet doivent se confronter est la compréhension de comment il a été possible d ‘obtenir le leadrship scientifique nécessaire au développement d’Internet. Le pouvoir se bat contre qui veit contrôler l’essentiel des fonctions d’Internet et contre les formes institutionnelles appropriées à ces fonctions montre qu’il sagit une tâche importante pour les gouvernements qui veulent assurer un développement continu d’Internet. La maniére de déterminer ce rôle et d’arriver au shéma institutionnel est une question cruciale pour le développement d’Internet. En apprenant à partir de l’expérience de L’IPTO durant son existence de 1962 à 1986, les gouvernements y compris le gouvernement de s Etats -Unis doivent être en capacité de dire comment concevoir les institutions nécessaires appropriées à leurs conditions nationales. L’expérience de l’INRIA ( Institut national de De Recherce Informatique ert Automatique) où le travail sur CYCLADES fût effectué, peut fournir une utile source comparative .Les institutions gouvernementales qui ont rendu possible une activité coopérative entre scientifiques de divers pays pour batir Internet ont besoin de modèles pour rendre possible la cooprération dans le développement continué d’internet La recherche portant sur, comment cela a été possible dans le passé , créera les formes institutionnelles appropriées aux nouvelles exigences du futur. La recherche historique sur les origines du développement d’Internet est un pré requis pour trouver les principes et les formes institutionnelles nécessaires pour orienter les futurs développements d’Internet
The significant question that researchers of the history of the Internet most need to grapple with, however, is an understanding of how it was possible to provide scientific leadership for the development of the Internet. The power struggle over who will control the essential functions of the Internet and over the institutional form that is appropriate for such functions show that there is an important task for governments in the continuing development of the Internet. How to determine this role and to provide the institutional form for it is a crucial outstanding question in the development of the Internet. Learning from the experience of the IPTO during its existence from 1962-1986, governments, including the U.S. government will be able to consider how to design the needed institutions appropriate to their national conditions. The experience in France with IRIA (Institut de Recherche d’Informatique et d’Automatique, a government sponsored research laboratory reporting to the Ministry of Industry) where work on CYCLADES was done, can provide a helpful comparative source for study. The government institutions which made it possible to support cooperative activity among scientists from diverse nations in the building of the Internet are needed models to make possible the international cooperation for continued Internet development. Research into how this has been done in the past will make it possible to create the appropriate institutional forms to provide for the new demands for the future. Research into the history of the origin and development of the Internet is a prerequisite to finding the principles and institutional forms needed to guide the development of the Internet for the future.