Summary

  • Zimbabwe’s 1990s connectivity was a set of systems with different users and operating conditions: PTC’s ZimNet X.25 network, university and sector email routed through South Africa, and early commercial TCP/IP.
  • A network name or Internet gateway did not mean every user had an always-on route. In 1992, university email could take about a day because calls ran over an office phone after hours; in 1995, a commercial provider still described its Johannesburg PPP connection as dial-up and non-permanent.

The call that ran after office hours

In August 1992, a Rhodes University network operator described two routes by which email reached the University of Zimbabwe: Fidonet and a UUCP dial-up to a machine in the university Computer Centre. The phone line was an office phone, available for data calls only outside normal working hours. Turnaround for mail on that link was about 24 hours. A message could cross a border, but that did not make the sender continuously connected to the Internet. The account is preserved in the NSRC archive of contemporary network correspondence.

That distinction helps explain why early Zimbabwean connectivity is hard to reduce to one “first Internet” date. A country could have electronic mail, a national data network, an Internet gateway abroad and a commercial IP service at different times, for different users. Each was a working system, but none automatically stood in for the others.

A national packet network, alongside academic email

A 1996 account by University of Zimbabwe contributors says the Posts and Telecommunications Corporation launched ZimNet in 1990 using X.25 packet switching. Banks, building societies and other financial organizations used it on a nationwide basis. That was significant data infrastructure, but the account does not describe ZimNet as an Internet backbone. In 1991, the university obtained an Internet link through the UNINET gateway at Grahamstown, South Africa. In 1992, it added a second host, MANGO, through ESANet, the Eastern and Southern African Network. The National Academies history places these services beside one another rather than treating them as a single network.

A contemporary message from October 1990 shows the university connection taking shape. F. F. Jacot Guillarmod at Rhodes wrote that John Sheppard of the University of Zimbabwe had visited Grahamstown to work out an electronic-mail link. The proposed machine was a 386 computer running Xenix, with email and news software. It is evidence of planning, not proof that the link was already operational. A year later, Sheppard’s application for the .zw domain named the University as administrator and gave a Rhodes-hosted route for his mailbox as it transitioned to zimbix.uz.zw. Naming and email reachability were developing alongside network access, not after a fully unified national system had appeared.

Separate communities shared a mail path

MANGO was not simply another name for the university’s Internet connection. Jim Holland’s May 1995 update described it as a voluntary organization for individuals and groups that needed email, not commercial messaging. It also described HealthNet, for public-health professionals, and ESAnet, for academics, as separate Fidonet systems operating from the university and routing mail through Mango. A later October 1995 network listing says the sector-specific nodes were restricted to their intended users, while a commercial provider, Data Control, offered TCP/IP through PPP.

This was both separation and cooperation. Health workers, academics, NGOs and companies had distinct service purposes and access rules; some nevertheless pooled the transport and operating work needed to exchange messages. The record shows a routed mail path between systems. It does not show that every group shared one address space, service contract or interactive Internet experience.

A commercial route was not yet a permanent circuit

The commercial label also needs qualification. In February 1995, an operator for Data Control wrote that its PPP link to Johannesburg was dial-up rather than permanent, and that most mail and news still came through UUCP. A 9,600-bit-per-second two-wire leased line was expected later that month, but the message does not confirm that it was installed on schedule. The operator described broader web access as constrained by bandwidth and still experimental. The contemporary account of networks and research in Zimbabwe likewise notes high ZimNet X.25 data charges and discusses direct Internet use as a developing service.

The 1996 university account says a commercial Internet gateway existed but was too expensive for most academic users, and the university still hoped to build its own gateway by the end of 1995. “Available in the country” and “affordable to a university department” were therefore different conditions. A service could exist and remain out of reach for a particular community.

The operating boundary was drawn by more than protocol. It included which institution paid for the international hop, whether the telephone could be used at night, whether a system stored and forwarded messages or offered live IP, which users were permitted on a sector network, and who could absorb a high data charge. Zimbabwe’s early connectivity was a patchwork because its routes answered different needs under different constraints. The sources do not establish one date when those constraints disappeared or all these systems became one broadly accessible network.

Sources