Summary

  • George Sadowsky helped organise a 1993–2001 series of Internet Society workshops that taught engineers from more than 100 countries how to connect, route, operate and govern networks—not merely how to use them.
  • The programme's strongest evidence of control transfer was reproductive: former students returned as instructors, local partners solved live connectivity problems, and participants kept a cross-border peer network after the classroom closed.

Analysis

At the Internet Society's 1996 workshop, roughly 260 students from 99 countries moved through four English-language tracks and one French-language track. Forty-five instructors supported them. The statistic that matters most, however, is smaller: some of those instructors had previously sat on the student side of the room.

That role change captures the institutional ambition behind the Developing Countries Network Training Workshops more accurately than a count of modems, routers or national links. A visiting team can install a connection. It can leave a functioning configuration and an impressive launch photograph. But when the first unfamiliar fault appears, authority migrates back to whoever understands the system well enough to diagnose it. Sadowsky's workshops tried to move that authority in the other direction.

Sadowsky had spent the years from 1973 to 1986 working for the United Nations on technology transfer, with assignments across more than 50 developing countries, including 20 in Africa. He then led academic computing and networking at Northwestern University and New York University. His official Internet Society biography also credits him and Larry Landweber with formulating the 1994 USAID Leland Initiative for initial connectivity in 20 African countries.

The workshops sat at the junction of those experiences: development programmes needed local technical capacity, while university networks knew that competence is acquired by operating systems under pressure.

The programme was neither a lecture tour nor a certification business. The Internet Hall of Fame describes nine locations between 1993 and 2001 and more than 1,500 participants from over 100 countries. Its account of the curriculum begins with the physical and logical act of connecting a network, then moves through routed domestic networks, useful content, operations, management and governance. These were different layers of the same problem. A link could be technically live while the organisation behind it lacked the ability to budget for it, set routing policy, manage users or explain its value to a ministry or university.

Contemporaneous records show the workshops acquiring a repeatable shape. In Prague in June 1994, 25 instructors ran four tracks for 162 participants from 62 countries. The Internet Society board was already discussing more frequent sessions and a formal training group. A later resolution commended organisers, instructors and students together, naming Sadowsky as chairman. The credit matters. Sadowsky supplied continuity and programme design; volunteer instructors, local hosts and participants supplied the working institution.

Curriculum was itself a control surface. Board minutes from 1995 endorsed more regional workshops while insisting that the Internet Society retain responsibility for the curriculum. Landweber and Sadowsky were assigned to review proposals. That combination—distribution of delivery, custody of standards—anticipated a persistent problem in infrastructure training. A programme that is completely centralised cannot scale or adapt locally. A programme with no common technical core may cease to produce interoperable practice. The workshop series was learning to separate what could travel from what had to stay coherent.

The classroom also had to be a live network. Minutes from the 1997 Kuala Lumpur workshop record connectivity-provisioning problems that MIMOS and Telekom Malaysia helped resolve. The incident is more revealing than a flawless demonstration would have been. Local partners were not stagehands for foreign experts; their knowledge of facilities and carriers was part of the operational system. Troubleshooting the workshop was, in effect, part of the workshop.

Internet Society minutes from 1996 went further, reporting that every new national Internet connection between 1993 and 1995 had involved a former workshop student. That was a board report, not an independent causal study, and it should not be converted into a claim that the workshops created those networks. National connectivity required local institutions, regulators, carriers, capital, equipment and years of work. The report is still important because it indicates where alumni appeared: not on the edge as consumers, but inside the teams making connection decisions.

The programme's economics exposed the limit of its initial design. The 1999 workshop offered English and French tracks, moved roughly 2,000 pounds of books and ran at an estimated loss. By then the board judged the original basic-connectivity objective largely achieved and saw training needs becoming more advanced. Later minutes put the broader Network Training Workshops total at 2,500 trainees and concluded that central delivery was breaking down. Chapters, Sustainable Internet Training Centers and locally rooted regional models were becoming more plausible than repeatedly flying a central faculty around the world.

The two totals should not be merged. Sadowsky later estimated that the Developing Countries series had reached about 1,500 people over ten years; a later board paper used 2,500 for the wider Network Training Workshops programme. The imprecision is a warning against treating attendance as the outcome. One trained person may remain isolated inside an institution that gives them no mandate. Another may train a team, advise a ministry, build an operator or become the colleague whom engineers in several countries call during an outage.

Sadowsky's own retrospective stresses that second kind of effect. He remembered 30 to 40 volunteers making the programme possible each year and described an email-connected international community that lasted beyond the sessions. Participants moved into ministries, nonprofits and businesses. The network between people was not decorative alumni engagement. In markets where documentation, spare equipment and vendor support were uneven, a trusted peer who had encountered a similar routing failure could be an operational asset.

The workshops therefore did not “export the Internet” from one centre to a waiting periphery. That framing erases the engineers who arrived with institutional knowledge, adapted techniques to local constraints and later taught others. What travelled was a deliberately incomplete package: shared protocols, working methods, a curriculum and access to peers. What had to be supplied locally was purpose, legitimacy, maintenance and the decision to make engineering competence a career rather than a temporary project input.

Sadowsky's distinctive role was to organise for his own diminishing necessity. The most successful workshop did not maximise dependence on its chairman or its visiting faculty. It enlarged the group able to teach, question and repair. In infrastructure, that is a harder legacy to photograph than a first connection—and a more consequential one.