The framing, and whose it is
The idea that a technology market opens for a period and then closes — that a company succeeds or fails less on the merits of its product than on whether it arrived while the window was open — is James L. Pelkey's, developed across the eighty-one interviews behind A History of Computer Communications, 1968–1988. It is used here by name because it is his analytical contribution rather than a general observation, and because the work it comes from deserves the credit.
What follows applies that frame to the first of three windows.
The trigger was a regulator, not an invention
Nothing was invented in 1968 that made this market possible. The modem already existed; Bell had been selling data sets for years. What changed was permission.
The FCC's Carterfone decision established that the telephone network had to accept attachments that did it no harm. Before it, the decided what could touch its lines and the answer was reliably its own equipment. After it, anybody could build a device, prove it was harmless, and sell it to a telephone subscriber.
That is an unusual way for an industry to begin, and it is why this window has a precise start date when the other two have approximate ones. A regulatory decision has a date; a market need does not.
The window: 1968 to 1972
Roughly four years, and the shape of it is worth noticing. The opening is sharp — a ruling in 1968 — and the close is gradual, as the companies that moved first established positions the later entrants could not dislodge.
The Computer Inquiry proceedings running alongside did the other half of the work: by separating regulated transmission from unregulated data processing, they made it lawful to sell a service that both moved data and computed on it. Between them, the two proceedings defined what a non-carrier was allowed to build and allowed to sell.
What the window produced
Modems, first and most obviously — the device that made a telephone line carry data at all, now purchasable from somebody other than the phone company.
Statistical multiplexers, which are the more interesting product because they answer an economic question rather than a physical one. Terminals are idle most of the time; a fixed time slot per terminal wastes most of a leased line. Buffering and interleaving instead meant fewer lines for the same terminal count, and the sales pitch was arithmetic.
Data sets, line drivers, and the whole interconnect catalogue that a corporate data centre needed once it was permitted to buy from anybody.
And, inevitably, a trade association: the IDCMA existed because a market opened by a regulator stays open only while somebody argues for it in the proceedings that follow.
Why it closed
Not because the products stopped selling — modem revenue kept growing for two decades. The window closed because the questions the market had been formed to answer were answered.
By 1972 the shape of a data communications product was settled, the incumbents were established, and a new entrant had to displace somebody rather than serve an unmet need. That is the definition of a closed window: the market still exists and it is no longer the place where a company can be built from nothing.
The engineers and founders who had done it once did not stop. Several of them appear again at the start of the next window, which opened around 1979 for a different reason and produced a completely different set of companies.
The revenue proves the point about windows
Modem revenue did not peak when the window closed. It peaked fifteen years later: about $33 million in 1970, and still climbing through $993 million in 1987 before turning down.
That gap is the whole argument in one series. The market was at its largest long after it had stopped being a place where a company could be built from nothing — which is why "the market is growing" and "the window is open" are different claims, and only the second one is an invitation to start something.
Figures from Pelkey, Appendix A.1, Product Revenues 1970-1988.