Summary

  • RFC 1208 carried a guide for Interop ’90 into the RFC archive as an information memo, not an Internet standard.
  • Its editors explained why they left “interoperability” undefined: the word’s meaning was unsettled, and the field changed quickly.
  • The 1993 and 1996 Internet Users’ Glossaries reorganized the audience and some definitions, but they still could not certify what running systems did.

Analysis

In March 1991, a glossary about networking reached the Internet’s document series with one conspicuous blank: it did not define “interoperability.” The word was printed nowhere as an entry. Near the end, the editors explained that “the jury is still out on exactly what it means.” That sentence is a small archival scene, but it catches a large engineering problem. A conference could gather products and protocols that claimed to work together; a pocket guide could help visitors understand their vocabulary. Neither act proved that the systems exchanged useful traffic.

RFC 1208 was titled A Glossary of Networking Terms. Ole Jacobsen and Daniel Lynch were listed with Interop, Inc. The memo says it was adapted from The INTEROP Pocket Glossary of Networking Terms, distributed at Interop ’90. Its audience was practical: people at tutorials, a conference and an exhibition who encountered unfamiliar terms and acronyms. This was not an attempt to write the last word on network architecture. The editors said they had left out common computing terms, kept entries short and recommended textbooks for fuller definitions. New terms appeared faster than a compact guide could keep up, so readers were invited to suggest additions.

That modest scope matters. The document entered the RFC series, but its status paragraph says it supplied information and did not specify an Internet standard. Its appearance in the archive did not promote every entry into a binding definition. Nor does the memo tell us what Interop attendees accepted, what vendors implemented, how many copies circulated or whether one definition became common practice. It records what two editors chose to publish for a particular audience.

The missing word was not the only boundary the glossary had to draw. It distinguished lower-case “internet,” a collection of networks connected by routers, from capital-I “Internet,” the largest such network. It described national backbones such as MILNET and NSFNET alongside regional and campus networks. Then it made a sharper classification: to be on the Internet, a network needed IP connectivity, with Telnet or ping offered as examples of reaching other systems. Email-only connectivity did not qualify under that entry.

That was the glossary’s description, not a standards-track admission rule. It also should not be read as saying that email-only networks could not communicate or interoperate in any sense. The entry separated a particular form of IP reachability from access by mail relay. CSNET, for example, had combined different access paths and services; its history is broader than a glossary’s one-sentence classification. The useful point is narrower: even the word “Internet” needed a boundary, and the RFC captured one editorial way of drawing it.

Other entries show the same work of disambiguation. RFC 1208 calls “gateway” the older Internet term for what was then more precisely an IP router. It separately uses “application gateway” for a system that translates between native formats, such as X.400 mail and RFC 822 messages. The word had moved, but older usages had not disappeared. A reader could not infer from a label alone whether a device forwarded packets or translated one application’s messages into another’s form.

Two years later, the reference had a different frame. RFC 1392, Internet Users’ Glossary, was prepared by the User Glossary Working Group in the IETF’s User Services Area. It focused on terms specific to the Internet, while retaining basic terms needed by other entries. It was still informational and explicitly not an Internet standard. Its definition of “Internet” emphasized a three-level hierarchy of backbone, mid-level and stub networks, and described the Internet as multiprotocol. Selected entries carried source labels, including references to RFC 1208 and other glossaries or authors. Those labels made some editorial lineage visible without turning the glossary into an adjudication of every term.

RFC 1983, published in 1996, obsoleted RFC 1392 and kept the Internet Users’ Glossary title and informational status. Its structure remained recognizable, while some examples changed: the Internet entry, for instance, named different backbone and mid-level networks. That is evidence of revision on the page. It does not establish why the examples changed, how readers used them, or whether the described network matched every operator’s experience.

The sequence is therefore less a march toward one final definition than a record of successive audiences. A trade-show pocket guide tried to make a mixed vendor and protocol landscape legible. An IETF user group later maintained a glossary aimed specifically at Internet readers. Each document helped organize language; none claimed to demonstrate interoperability in operation. A useful definition can tell an engineer what to inspect, but it cannot stand in for an exchange between running systems.

That distinction is the lasting lesson of RFC 1208. The glossary could name a router, separate an IP path from a mail relay and describe what counted as the Internet in its own editorial frame. It left interoperability open because the word reached beyond the page. To establish that two systems worked together, one still had to look at the protocol, the implementations, the configuration and the traffic they actually exchanged. A glossary is a map of terms. It is not a packet trace.

Sources