Brian MacMillan

This essay argues that the 1992 Rio Summit on Climate Change (the “Earth Summit”) was one of the catalysts for the development and success of the World Wide Web. I know because I watched it happen! The paper was presented at the “Environmental Story Telling” Conference, MAGNET, NYU-POLY, April 2015. Note that I am in the process of web-publishing this presentation. The content is pretty much done, but I need to take a morning off to add the (extensive) footnotes. A pdf version is available upon request.

Start

Non-Military Sources for the World Wide Web: The Association for Progressive Communications and the Expansion of the Internet in Brazil and Southern Africa, 1990-1996

The explosion of Internet use during the 1990s was an extremely complicated process, marked by the intermixing of political, economic and social agendas that were, often, only related to the Internet itself in an ancillary fashion. Unfortunately, the bulk of academic research on Internet history only lightly touches upon these complexities and instead focuses on simple, popular narratives, which though broadly speaking are accurate, are incomplete because they are facile and in most cases tell a story that is technologically deterministic.

The academic literature about the history of the Internet falls into several categories. The dominant narrative analyzes the development of the Internet using the tools of institutional historiography, focusing specifically on the alliance between military contractors, the American university system and the United States Department of Defense. Other dominant narratives analyze the history of the Internet from the perspective of enabling technologies, such as TCP/IP, Unix, packet switching and Ethernet; enabling applications, like email; and the technologists that developed them. Though all of these narratives have slightly different explanations for the development of the Internet until 1990, that vary more in terms of emphasis than content, all of them explain the rapid expansion of the Internet during the 1990s in terms of the spread of the World Wide Web. The rapid and extensive adoption of the World Wide Web during the 1990s is generally attributed to the development of NCSA Mosaic, and later Netscape, by Marc Andreessen, and Netscape’s subsequent adoption as the primary tool for Internet access.

Janet Abbate’s article “Government, Business, and the Making of the Internet” is among the best of the institutional histories of the early Internet. Her article clearly lays out the relationship between the department of defense, military contractors such as Lockheed Martin and the university system. A very similar, though more focused, story is told by Arthur Nortberg in his history Information Processing for the Pentagon, as well as in online histories, such as the Internet Society’s “Brief History of the Internet”.

The institutional histories of the Internet are closely paralleled by the technological histories. Among the most cited of these histories, are Netizens, by Michael and Ronda Hauben, Casting the Net, by Peter Salus, and Where Wizards Stay Up Late, by Katie Hafner and Matthew Lyon. These academic works are supplemented by hundreds of articles and web pages which lionize the creative genius and financial successes of lone entrepreneurs. Where Wizard’s Stay Up Late is probably the most representative of the genre, and exhibits all of its strengths and weaknesses. One of the myths of the Internet has always been that it has been a product of brilliant, typically maverick, individuals. The myth is compelling not only because there are so many examples in the history of the Internet of remarkable men, including Steven Jobs, Eric Brin and Tim Berners-Lee, who at least to an extent confirm the theory, but also because the theory dovetails so completely with a dominant myth of American society about the loner who goes outside “the system” in order to improve it. From this perspective, Bill Gates, the Harvard dropout who transformed the world, becoming fabulously wealthy as he did so, is the iconic example. Though Hafner’s story of internet wizards makes for a compelling read, it is somewhat aggravating to a historian because the social, economic, cultural and economic transformations that the Internet ‘wizards’ have initiated are presented as a given and not explained.

In both the scholarly literature and in the trade press, the history of the Internet is presented in a remarkably consistent, and largely non-contentious manner. Virtually all commentators agree that the development of ARPANET – the communications backbone of the university-military research efforts – in 1969 was a critical stage in the development of the infrastructure of the internet and that the development of certain key technologies by brilliant inventors such as Kernigan and Ritchie (Unix), Paul Baran (packet switching) and Tim Berners-Lee (HTML) was a necessary condition for the creation of the Internet as we know it. Even those commentators, such as Ian Peter, who challenge this narrative, do so primarily to emphasize that European universities and researchers also contributed technologically to the development of critical internet technologies, and were also part of the military-university network that evolved into the internet. The problem with these histories occurs when they attempt to explain the transformation of the Internet from a purely military-university research infrastructure, to the dominant global electronic network during the 1990s.

Janet Abbate provides one of the more scholarly summaries of the transformation,

“[in the early 1990s there] were still many obstacles to finding information on the internet, however. There was no way to link information found in different documents, and the various protocols that had evolved for exchanging information were not compatible; no one program could handle formats as diverse as ftp, mail, gopher, and WAIS. … All of these issues were addressed by a net Internet application that became known as the World Wide Web. The Web would fundamentally change the Internet, not by expanding its infrastructure or underlying protocols, but by providing an application that would lure millions of new users. The Web also changed people’s perception of the Internet: Instead of being seen as a research tool or even a conduit for messages between people, the network took on new roles as an entertainment medium, a shop window, and a vehicle for presenting one’s person to the world.”

 

David Hudson, in Rewired provides a more breathless, and typical, description of the transformation initiated by the World Wide Web:

“Marc Andreessen’s realization of Mosaic, based on the work of Berners-Lee and the hypertext theorists before him, is generally recognized as the beginning of the web as it is now known. Mosaic, the first web browser to win over the Net masses, was released in 1993 and made freely accessible to the public. The adjective phenomenal, so often overused in this industry, is genuinely applicable to the…’explosion’ in the growth of the web after Mosaic appeared on the scene. Starting with next to nothing, the rates of the web growth (quoted in the press) hovering around tens of thousands of percent over ridiculously short periods of time were no real surprise.

It is not particularly surprising that the Netscape-as-engine-of-Internet growth story should have become the dominant narrative in histories of the Internet during the 1990s. The technologist-entrepreneur story, as noted above, is an important computer-era myth. This bias is enhanced by the nature of Internet development itself. As late as 1995, 40% of all Internet traffic, according to Pacific Telephone, occurred in California ; and to this day the majority of trade journals and many of the authors of Internet histories are California-based and are active as Silicon Valley consultants. The proponents of the conventional narrative are merely describing themselves and their experiences.

Though there is a kernel of truth to the World Wide Web thesis, particularly as an explanation for Internet growth in the later part of the decade, it is a quite inadequate thesis for explaining the growth of web use from 313,000 to 93,047,785 hosts between October 1990 and July 2000 , and more importantly, it does not explain the expansion of the Internet globally: throughout this period Internet access in Latin America, Africa and Asia was predominantly, and in most developing exclusively, text-based. Even in developed countries, the penetration of World Wide Web technologies was far less pervasive and thorough than the literature suggests simply because the bandwidth of v.32bis modems – the dominant means of internet access from personal computers in this period – was only 14,400 baud, making the display of graphical web pages quite inconvenient. It is not until the appearance of the faster v.90 modem later in the decade, and the concurrent rapid expansion of corporate networks, that Internet users in developed countries began to take advantage of the multi-media capabilities of the World Wide Web. Even then, adoption of World Wide Web technologies was slow: research that suggests that as recently as 2002, up to 84% of Internet users connected to the Internet exclusively for email, a technology that even today only incidentally embodies World Wide Web technologies.

These problems in the historical literature about the expansion of the Internet out of California and academia, have been noted and partially addressed. In the social sciences there is a considerable body of literature, which addresses the mechanics of Internet growth. This literature has applied models from information theory related to the diffusion of innovation, and the importance of contagion to technology adoption. Amitava Dutt and Rahul Roy, for example, in their work “The Mechanics of Internet Growth” attempt to quantify correlations between infrastructure, education, income, government policy, and the rate of internet growth. Such work does much to broaden our understanding of Internet growth, but like all efforts to construct predictive models, results in deterministic conclusions that gloss over the role of human agency in technological diffusion.

The most important question left unanswered by the literature is how and why exactly did various sectors of society adopt the Internet? On one level the question is quite prosaic, for the answer lies in millions of decisions by individuals to buy personal computers, struggle with modems and pursue their idiosyncratic interests through user groups, bulletin boards and social networks. On another level, it is fundamental, touching at the very heart of the question of how the internal structure of our society and economy has changed over the past generation as a result of the widespread use of Internet technologies; and how it is that global society has truly begun to emerge.

At a broad level of analysis there are four areas that had the biggest impact on the growth of the Internet during the period under discussion. The decision by the largest purveyors of email services, American Online, CompuServe and Delphi, to transfer their user base to the Internet was responsible for a considerable portion of the growth in Internet use during 1992-94. Not only did this transfer bring a large number of people into the Internet fold, the decision was followed by an explosive growth in the users of these services in the years immediately following the transfer. A second significant factor in the spread of the Internet was the decision on the part of the owners and operators of purpose built networks to adopt Internet technologies, either partially or wholly, during this period. Among the largest networks to do so were BITNET in the area of academic communication, LEXIS/NEXIS for legal research, ERIC for educational research, EDGAR for securities filings, and the networks of the Association for Progressive Communications (APC) in the fields of social and environmental activism. Two other important developments begin in this time but are less significant until the late 1990s: the commercialization of the Internet; and the spread of Internet use into civil society as a result of the diffusion of personal computers.

Of the agents cited above, the experience of the Association for Progressive Communications provides many significant insights into the question of how it was that the use of the Internet spread from a small number of research institutions clustered in California and in the east coast of the United States in the 1980s to become a global economic, political and social force one generation later.

The Association for Progressive Communications was founded in 1990 as a merger of the networks of GreenNet (UK), IBASE (Brazil), Nicarao (Nicaragua), NordNet (Sweden), Pegasus (Australia), Web Networks (Canada) and the Institute for Global Communications (USA). The Institute for Global Communications was the result of an earlier merger of EcoNet and PeaceNet. The IGC itself was an initiative by the Tides Foundation to promote a progressive agenda within America and internationally. By 1992, the APC connected over 11,000 Non-Government Organizations (or Community Service Organizations, to use a more recent term for these groups) and individuals in over 90 countries through ten major hubs.

At the time of its creation, the APC had one of the few global electronic communications networks, encompassing not only thousands of organizations from the countries of its founding members, but also, through GreenNet’s Fido network, providing electronic email and bulletin board services to Community Service Organizations (CSOs) throughout the world. Over the course of the next four years APC, through the independent activities of its member organizations, and centrally through the Institute for Global Communications, was responsible for the expansion of Internet services throughout Latin America, Eastern Europe, Asia and Africa. In most developing countries, the APC provided among the very first non-governmental Internet gateways, and in some, like Brazil, it was the very first public Internet service provider (ISP). An examination of APC’s role as an ISP in Brazil and southern Africa between 1990 and 1996 provides many clues as to the role of human agency in the spread of the Internet, and does much to fill in gaps in the story about the global expansion of the Internet.

In some ways, Internet usage spread in Brazil along very similar lines to developed countries such as the United Kingdom and Canada. In the 1980s the backbone of the Brazilian Internet was the RNP network, which represented the communications infrastructure of the Brazilian University research system. The maintenance and development of this infrastructure was promoted by the National Research Council (CNPq), an agency of the Ministry of Science and Technology. Unlike in the developed world, however, the military importance of the Internet, though present, was decidedly secondary. Adoption of the Internet by civil society also followed a dramatically different path than in the developed world.

Perhaps the most important influence on the expansion of the Internet in Brazil was the exile of Carlos Afonso – the future director of IBASE – to Canada following the military coupe in Chile, where he then lived, in 1973. Afonso, while in Canada, established contacts with various activists, including Herbert de Souza (Betinho), with whom he founded the Brazilian Institute for Social and Economic Analysis (IBASE) shortly before returning to Brazil in 1980. In an interview published by the International Development Research Corporation – a Canadian Crown Corporation that plays a significant role in the story of APC’s Internet work, Carlos described his return to Brazil as follows,

I came home, late in 1980, with an Apple II computer in my luggage. It had 64 Ks of RAM, which was considered huge at the time… I had, in addition, two floppy disks and a printer. That is all we had in the way of equipment, to begin with. Things were made all the more difficult by the extremely severe regulations restricting equipment imports in those days.

That Carlos should mention the importance of the Apple II computer is not incidental to this story: most of the activist work that Carlos did for the next 20 years both facilitated and benefited from technology transfer and diffusion.

In July 1989 IBASE, with assistance from the Institute for Global Communications, one of the founding members of the Association for Progressive Communications, established Alternex, the first non-governmental Internet Service Provider in Brazil. For its first years as an Internet service provider IBASE/Alternex was a tiny operation providing Internet email connectivity to a handful of Community Service Organizations. Its technological capabilities were expanded dramatically as a result of the United Nations Conference on Environment and Development, held in Rio de Janeiro in June 1992, a story that underlines the critical importance that human networks had on the expansion of the Internet in Brazil.

The United Nations had, since the Stockholm Conference on the Human Environment in 1972, actively invited Community Service Organizations as participants, and was indirectly responsible for the creation of a number of them, for example the international branches of the Sierra Club and Friends of the Earth in the United States and Pollution Probe in Canada. This practice of inviting CSOs to United Nations Conferences had over the following two decades following the Stockholm Conference increased until, at the time of the Rio summit, CSO participation in the UNCED NGO forum was far greater than the participation of UN bureaucrats and diplomats in the Conference itself.

The extensive interest of CSOs in attending the Rio Summit posed serious logistical problems for the UNCED Secretariat, in terms of disseminating information about and soliciting input for UNCED’s various working groups and committees. Shortly following the first of the three preconference meetings of UNCED in Nairobi in 1990 Maurice Strong, the head of the Secretariat and his Senior Advisor Peter Thatcher, began discussions with Bill Leland of EcoNet and Bob Loeb and Galen Dixon of the Telecommunications Cooperative Network. The American environmentalist Robert Pollard acted as an intermediary and was the person responsible for drafting the proposal that was ultimately presented to Strong by the APC for providing computer services to UNCED. The introduction to this proposal is quoted below,

ACTION ARENAS FOR THE 1992 UNITED NATIONS CONFERENCE

The proceedings leading into UNCED have a number of distinct constituencies, action arenas and agendas that the design of the UNCED electronic conferencing needs to address. These include the United Nations Conference on Environment and Development itself and the workings of the Preparatory Committee, in which delegates from national governments will participate, along with non-government organizations (NGOs) in consultative status to the UN; there is also a need for structuring access to the formal UN proceedings and deliberations by other NGOs, by private voluntary organizations (PVOs) and by individuals.

In each country, there will be a need for NGOs and others to organize to raise the issues of UNCED on their national agendas. In this context, there will be several different areas of focus; some organizations will be directing their attention towards government and the development of national legislative agendas; others will be will be developing action agendas within a variety of grassroots constituencies at a national level; others still will be developing local or regional forums in which issues of environment and development will be addressed.

Furthermore, there have been preliminary discussions about convening a “Congress of the People of the Earth” in Brazil at the time of UNCED (June 1992) as a parallel conference for NGOs and other grassroots groups with a concern for addressing issues of environment and development. Appropriate use of electronic conferencing could play a key role in preparations and planning for such a gathering.

Within each of these contexts, the use of electronic conferencing can play a valuable role. While there will be a value in having separate electronic conferences within each action arena, all will need to maintain access to a core set of information on the UNCED proceedings, and there will be many areas of overlap and/or need for dialogue between participants in the different action arenas.

The intended role for the APC was to provide support services for UNCED. Nevertheless, it was openly acknowledged – as the proposal above indicates – that one of the goals behind APC’s participation was to effect legislative change in participant countries. Commenting on this, Rory O’Brien of Web Networks (the Canadian founding member of APC) says, “It must be stressed that the UN efforts were overlaid on the concurrent independent efforts of the APC to electronically network civil society both locally and globally” with the intention of promoting the APC’s global environmental and social justice agenda. Interestingly, this overt political agenda provoked very little negative commentary. When asked about this, O’Brien noted that at the time promoting international communication through the use of the Internet was viewed as a politically neutral activity so organizations and individuals with widely diverging political viewpoints were able to benignly cooperate with each other.

Though no one disputed that the ability of the APC to provide the support services required for UNCED, the technological infrastructure of Alternex, the Brazilian branch of APC, was inadequate to the task at hand. According to Maureen James of Web Networks Alternex’s technological challenges were so severe – and importing computer technology so costly and time consuming – that the first IGC staff member to fly to Brazil on the UNCED project had to smuggle a computer in his personal luggage in order to expedite the upgrade process.
Carlos Afonso, writing in 1996, summarized the tremendous impact that the Rio Conference was to have on Alternex’s infrastructure,

AlterNet provided only e-mail connectivity to the Internet. However, at the end of 1990 several environmental organizations approached IBASE to suggest the development of a major independent project of electronic communication to be made available at the United Nations Conference on Environment and Development (UNCED) … IBASE prepared a detailed Internet project for UNCED and submitted it to the UNCED Secretariat. It was accepted and included as part of the general host country agreement between the government of Brazil and the United Nations for the conference. As a consequence AlterNex underwent a major upgrade, replacing its aging 386-based UNIX machines with a network of Sun SPARC stations specially donated by Sun Microsystems.

At the same time that Alternex was being upgraded for UNCED, an extremely well connected, Canadian-based Texan, Langston James Goree VI, was contracted by the International Institute for Sustainable Development, a Canadian based environmental organization, to work with the Brazilian NGO – the Institute for Prehistory, Anthropology and Ecology (IPHAE) – both during and after the Rio Conference to promote the “redemocratization” (sic) of Brazil, and to use computer technology to connect NGO’s in Brazil’s northeastern states and the Amazon with each other, and through the Alternex hub, with the world. The goal of the project was to ensure that the international ties being forged by Alternex were cemented through the enhancement of its national network.

These efforts were very successful. In 1994, the Movement for Ethics in Politics, an IBASE initiative which extensively used Alternex for mobilization, played an instrumental role in forcing President Fernando Collor to resign because of allegations of corruption. Over the next decade IBASE’s extended these policy successes to the point that most of its agenda has been adopted and implemented by the government of Luiz Inácio Lula da Silva.

Though most discussions of Alternex focus on its political role, it is important to note that Alternex also played a significant role in the expansion of open source technologies throughout Brazil and Latin America. When Alternex began providing Internet services, installing a telephone line in Brazil could cost upwards of $3,000, a steep price by developed world standards and prohibitive within the Brazilian context. Alternex, throughout this period took a proactive view of its role in Brazil, actively working with the Brazilian University community, including RNP and Brazil’s network of computer technical colleges (CSI) to provide Internet services and technical training to the poorest components of Brazilian society. This work not only helped facilitate Internet penetration within Brazil, it also firmly established the use of open source technologies throughout Latin America, for Alternex acted not merely as a Brazilian ISP, but also (briefly) as a regional hub which provided internet gateway services to organizations in Ecuador and Uruguay.

Whereas the role of the APC in the Rio Summit underlines how various components of civil society, including international aid organizations, multinational corporations, environmental lobby groups, and individual activists could cooperate to initiate technological diffusion, the role of APC in extending internet service to southern Africa provides a much different account of the importance of human agency to the spread of Internet technology in the early 1990s.

The situation in southern Africa during the early 1990s was quite different than in Brazil. Whereas Brazil already had an established, albeit expensive and incomplete, telecommunications infrastructure, in much of southern Africa, particularly rural areas, there was no infrastructure whatsoever. The political situation in southern Africa was also much different. Brazil’s government was democratically elected and somewhat supportive of CSO’s, particularly those with religious affiliation. In sharp contrast, Zimbabwe and Malawi had authoritarian governments that were overtly hostile to any form of non-governmental political activity; and the Republic of South Africa was in transition to a multi-racial government (the first non-racial election was held in 1994, which the African National Congress handily won). Only Botswana and Mozambique were full democracies.

In 1992, the APC convened a Global Networking Workshop in Toronto to consider issues related to extending its services throughout the developing world. The working paper of the Africa Workshop provides valuable insights into the state of Africa’s networking infrastructure at that time.
As of February, 1992, there were eight Fido systems in Africa operating as hosts for about over 100 end users who are able to send and receive electronic mail and conferences… These systems are operated by a broad range of different organisations: quasi-state bodies like the Centre for Scientific and Industrial Research in Accra, Ghana; international non-governmental bodies such as the UN Economic Commission for Africa in Addis Ababa, Ethiopia; single large International NGOs like Environmental Liaison Centre International (ELCI) in Nairobi, Kenya; coalitions of NGOs as in MANGO in Harare, Zimbabwe; autonomous NGO service organisations such as WorkNet in Johannesburg, South Africa.

… Some systems run on 386s with 120 MB hard disks, some run on slow XTs with only 20MB hard disk. An XT is adequate to serve the needs of 15-50 users when Fido software is used to automate and compress message transfers. Each host has a Telebit T2500 dual standard modem connected to a single phone line. This modem supports high-speed links (9600 or 19200 baud) as well as the standard 1200 and 2400 baud. The standard speeds are mainly for users, while the high-speed links provide gateways to systems of the Association for Progressive Communication (APC).

Not only were Internet users in Africa using legacy technologies that were between 1 and 3 generations behind the developed world, but African telephone networks, when they existed at all, were so noisy that two way communication was impossible. The Working Group on Africa put together a list of recommendations for hardware and software that could inexpensively address these issues.

In 1994, the Canadian Government, through its foreign aid agency CIDA, contracted the Canadian branch of APC, Web Networks, to begin working on the integration of CSOs in Ecuador and Africa. CIDA money was supplemented by funds from the Netherlands Foreign Ministry and the International Development and Research Centre (IDRC). Other organizations, which provided non-monetary support included Partnership Africa Canada and OXFAM Canada.

The work that was done by IGC employees Maureen James and Rob Ellis on this project for SANGONET in the Republic of South Africa was not unlike the work done by IGC employees several years earlier for Alternex. SANGONET was an already existing computer network in a country, like Brazil, that was already partially developed. There were some differences in terms of the goals of their work. Whereas the work in Rio was strongly influenced by an environmentalist and social activist agenda, the work in South Africa and Botswana had a public health component, focusing particularly on issues related to the spread of AIDS; as well as women’s issues.

The APC efforts in Malawi and Zimbabwe were quite different, however, than in Brazil. In these countries, the IGC employee Mike Jensen, with a knapsack stuffed full of network gear, hiked around the countryside as a form of digital missionary, setting up connectivity to the APC network for those NGOs which were interested and which APC deemed had operations consistent with their mandate.

The extremely colorful experiences of Mike Jensen underline a number of interesting points about the manner in which the Internet initially penetrated southern Africa. The most obvious, of course, is how personal the process was. In part, this was a reflection of the lack of resources available – one dedicated and capable individual could do the work that dozens did elsewhere. Another important consideration was political. The authoritarian governments in Zimbabwe and Malawi did not appreciate the work that Jensen was doing. Because he acted as an individual he could hide his activities from the authorities far more effectively than an institution could. Jensen was a very creative and resourceful individual who actively sought out solutions to problems posed by government hostility and censorship, for example by using satellite phones to circumvent the Zimbabwe telephone system entirely. Jensen’s work points to a considerably different dynamic in the technology transfer process in southern Africa. Whereas in Brazil, Carlos Afonso of IBASE/Alternex and Wim Groeneveld of IPHAE actively solicited financing for their network projects, the African work by Web Networks and the IGC was initiated by Canadians who were – at least an to extent – proselytizers for the new technologies. The mechanics of this process is hinted at in the introduction to Arnie Mikelson’s 1992 report “Networking in Africa”,
On first consideration, computer communications in Africa would seem an inappropriate form of development when basics such as food, sanitation, clean water and health care are in short supply. In fact, computer networking is the ideal tool for enabling and improving communications in these areas because it makes such efficient use of scarce resources.

There is no mention in Web Network’s documentation about any interest on the part of Africans to connect to the APC, and there was only a cursory discussion about whether CIDA’s money would have been better spent on food aid or schoolbooks. Though there undoubtedly was African interest in connecting to APC, for example through the Partnership Africa Council, it is important to note that the Web Networks African project was a classic example of tied aid, where a donor country dictates how it’s funding is to be spent. In this instance it is possible to argue that there was a push factor in the diffusion of Internet technology to southern Africa, unlike the case of Brazil, where Carlos Afonso very clearly was encouraging Internet diffusion to promote his own political and social agenda.

The social and political mechanisms by which the Internet spread to Latin America and southern Africa point to the critical importance that human agency played in the process of technological diffusion and the relative unimportance of technology itself in this process. Had the NSF decided not to commercialize the Internet as a result of military considerations, it is certain that the same actors would have promoted a similar agenda using a slightly different communications infrastructure. Most of the players viewed the Internet as merely one tool among many that could be used to further particular goals, like mobilizing global public opinion in order to increase AIDS awareness or reduce economic inequality.

These insights about the role of human agency in technology transfer apply equally to the process of technology diffusion in developed countries. It is very useful to consider exactly what the distinct motivations were of the millions of people who adopted the Internet in developed countries during the 1990s. Many were pushed into the Internet through their affiliation with purpose built networks that chose to connect for particular institutional reasons. Many more chose to connect for a host of idiosyncratic reasons that they considered sufficient to justify the considerable effort that Internet adoption entailed.

When the study of Internet adoption focuses on human agency it is possible to consider the transformations that Internet use inducing in our society, and the mechanisms of these transformations. Clay Shirkey, in his work Here Comes Everybody has written about the way in which the Internet has altered the way people organize; David Weinberger, in his work Small Pieces Loosely Joined has considered the manner in which the Internet has altered our conceptions of space, time and society; Peter Morville in Ambient Findability and Information Architecture for the World Wide Web has discussed a number of ways in which the Internet is altering who we are and how our world is (and should be) designed. In addition, a host of thinkers have considered how the theory of emergence – “the arising of novel and coherent structures, patterns and properties during the process of self-organization in complex systems” – relates to daily activities on the Internet.

One of the surprising gaps in Internet historiography is that none of these approaches, which are explicitly concerned with the structure, development and functioning of the Internet, has been applied to the history of the Internet itself. This study suggests that applying some of the methodologies employed by philosophers, designers and social scientists, could be extremely useful to explaining, in a nuanced fashion, the development and spread of the Internet. The explosion of communications networks during the 1980s, for example, and the final merger of these networks into the Internet, follows many of the patterns found in emergent systems, not just in terms of which networks triumphed and how, but also in terms of which networks, like BITNET, ultimately failed. The APC itself is an example of a small, successful, emergent system that spread through a persistent increase in links between its many technological and human nodes.
The ultimate problem for the historian is that to make arguments that encompass both human agency and systems theory requires a tremendous amount of detailed scholarship. The study of the expansion of the APC into Brazil and Africa has been relatively simple to conduct, because the APC was not that large an organization and its records are relatively well preserved. Understanding to story of how the Internet was adopted by corporations and individuals during the latter part of the 1990s is a far more challenging task, but one well worth embarking on because, unfortunately, most current explanations of the growth of the Internet are far too simplistic and technologically deterministic.