Why a trade show is worth an article

Because in 1988 the industry's central argument — OSI or — was still formally open, and the two answers were being demonstrated in ways that could be compared by anybody who walked the floor.

Specifications did not settle it. Procurement mandates did not settle it. What made the outcome legible, before any committee conceded anything, was the difference between an event where equipment from competing vendors had to work together in public and an event where it did not.

Interop was not invented for the occasion

This is the part usually missed. was founded in 1986, by Dan Lynch, growing out of the first TCP/IP Vendors Workshop in Monterey, California. Large corporations — IBM and DEC among them — were attending from the beginning.

So by 1988 it was not a new idea being tried. It was two years of accumulated practice at getting rival implementations onto one network and finding out, in front of customers, whose stack was wrong.

That accumulation is the thing worth taking from this. A demonstration is a claim you can check; a specification is a claim you have to trust. The second is cheaper to produce and the first is what a buyer remembers.

The proposition that made it uncomfortable

Connect your equipment to a shared live network, on a show floor, alongside your competitors, where it either works or visibly does not.

No enjoys that. A controlled demonstration on your own stand can be prepared; a shared network cannot, because the other end of every conversation belongs to somebody else. Interoperability stops being a marketing adjective the moment it has to happen in a room with a deadline.

Which is precisely why it was persuasive. OSINET and COS existed because conformance to a document and interoperation with another vendor's product are different properties — and the show floor tested the second one continuously, in front of the people writing the cheques.

The other event, and its source

Pelkey's history records an OSI counterpart in the same year: the Enterprise Network Event, or , intended to show enterprise buyers a coherent OSI ecosystem at the moment GOSIP and the other procurement mandates were at their most confident.

That account rests on his history, which devotes a numbered section to each show: Enterprise Network Event (OSI), June and Interop (TCP/IP) Trade Show, September. The months are in the section titles, which is where the three-month gap in this comparison comes from.

Unlike Interop, ENE has left little other independently accessible record — itself a small piece of evidence about how the two are remembered.

What his sources describe is an ecosystem more described than shipped: standards near-complete, implementations in progress, conformance regimes being constructed, and a buyer being asked to plan around all of it.

What a buyer could actually see

Set the two side by side and the asymmetry does not need interpreting.

One demonstrated a working network between rivals. The other demonstrated an intention.

Nothing about that comparison required knowing which stack was technically superior — and the technical case for OSI was serious, made by capable people, with institutional weight behind it. It required only noticing which one was running.

The lesson that outlived both

Every argument since has had the same shape, and the resolution has usually arrived the same way.

Running code is not an argument that beats specifications on merit. It is an argument that arrives earlier, and arriving earlier compounds: implementations attract implementers, deployment produces operational experience, and operational experience produces the next generation of engineers who consider the thing normal.

By the time a standards process has finished being thorough, the question it was answering has often been settled by the people who did not wait.