IT WAS THE LATE 1970S. Former French presidents Charles de Gaulle and Georges Pompidou had recently died. The Arab oil embargo had quadrupled energy prices for a while. Marseille was still plagued by drug lords. And France had to face the fact that its telephone network was one of the worst in the industrialized world. Less than 7 million phone lines served 47 million French citizens, and the country’s elite felt that the dominance of American firms in telephone equipment, computers, databases, and information networks threatened its national sovereignty. Or, at the very least, was infringing on their cultural pride.
In an influential 1978 report to President Valéry Giscard d’Estaing entitled “The Computerization of Society,” government researchers Simon Nora and Alain Mink argued that the solution to France’s telecommunications problems lay in the field of “telematics”-a combination of telecommunications and informatics. They laid out a plan to digitize the telephone network, add a layer of interactive teletext video technology, and provide entrepreneurs with an open platform for innovation.
Taking Nora and Mink’s vision to heart, the country’s leaders began laying the groundwork for France’s computerized future. In 1983, under presidential orders, computer engineers at the Ministry of Posts, Telegraphs and Telephones (PTT) began rolling out a telematics system called Minitel throughout France. It allowed ordinary people to receive and share information online, ushering the country into the digital age and putting it more than a decade ahead of the United States.
The story of Minitel’s emergence is a fascinating but largely forgotten one. To the extent that it is remembered today, Minitel is portrayed as a closed, centralized system burdened by government bureaucracy that couldn’t change with the times. But in 1983, the system was unlike any that had gone before, eventually growing to more than 20,000 online services before the World Wide Web even existed.
Today in Silicon Valley, those who lived through the Minitel era tend to view it as the epitome of how online systems should not be built and operated: They believe that letting the government design and operate it would lead to disaster. In fact, Minitel was never fully controlled by the government. It was a hybrid system – a government platform for private innovation. And it worked pretty well.
To establish a connection, a user manually dialed a local gateway number using a handset. The call, transmitted over the public switched telephone network, was answered by software installed on the switchboard – typically a CIT-Alcatel E-10 – that played the operator’s audible tone over the line. Upon hearing this tone, the user would place the handset back on the stand and begin using the Minitel terminal, which would perform a special handshake protocol with the switchboard.
The gateways, known as aspoints d’accès videotex, or PAVIs, provided access to a catalog of well-known Minitel services designated by short mnemonic codes. For example, rail travelers bought tickets at 3615 SNCF (SNCF is an acronym for Société Nationale des Chemins de Fer, the French railroad), news buffs gathered at 3615 LEMONDE (Le Monde, the leading Parisian newspaper), and dudes (mecs in French) browsed personal ads at 3615 MEC. Like URLs today, these codes were printed in magazines, shown in TV ads, and pasted on the sides of buses.
Once the user entered the desired destination, the switch created a virtual circuit on the public data network known as Transpac, and data began flowing from the client terminal to the main server and back again. These virtual circuits used the X.25 network protocol, a paradigmatic packet-switching technology developed primarily by researchers at France’s Center for Communist Studies in Television and Telecommunications.
Early on, however, Minitel proponents ran into a chicken-and-egg problem. Why would anyone implement a system if they wouldn’t be interested in working with it? And how do you convince entrepreneurs to create services if the platform doesn’t already have users? Somehow Minitel needed to attract both users and service providers at the same time.
To start the process, PTT ordered millions of Minitel terminals (made by French manufacturers such as Telic-Alcatel and Matra) and provided them free of charge to everyone in the country with a telephone line. Anyone who was curious about the new system advertised on television could simply go to the post office and come home with a shiny new Minitel box.
Minitel’s developers made the system completely plug and play: All you had to do was plug the terminal into the wall, dial the local gateway, and voila – you were transported into cyberspace. Whereas in the United States, would-be cybernats wanting to go online had to buy expensive computer hardware, install confusing software, pay huge long-distance telephone bills, and pay a separate subscription for each service provider they wanted to use.
The first service available on Minitel was an electronic telephone directory, or annuaire électronique. Equipped with a natural-language search interface, this frequently used resource was an easy way to get to know Minitel for free. Later, the government began requiring people to use Minitel for certain administrative tasks, such as university registration. These modest public services spurred the introduction of Minitel into France’s rapidly expanding telephone network. By the late 1980s, every adult in France had access to the network while at home, at work, or at a public terminal on the street.
As the number of Minitel users grew, entrepreneurs jumped at the opportunity to create new services. These startups have been helped by a new payment system built into the Minitel platform that has lowered the barrier to entry. Named after the newsstands that dot Parisian boulevards, PTT’s Kiosk system does the bookkeeping, collecting money from users at one end, cutting checks for service providers at the other, and keeping a tidbit for itself. In this way, small service providers could build lean information systems that generated revenue without having to manage customer relationships, accept credit cards, or keep track of overdue invoices. Indeed, the app store model used by Apple, Steam, and others today is nothing more than a privatized version of the Minitel kiosk.
Providers were allowed to use any hardware or software, as long as its products complied with rules published by the phone company. As demand for Minitel grew, stiff competition emerged in the server hardware market. Vendors built their systems on any machine capable of running a multi-user operating system, from proprietary mainframes and Unix-enabled minicomputers to Commodore Amigas and IBM PCs.
In addition to the iconic terminal equipment, France hoped to launch domestic production of server hardware. This part of the telematics project did not go as planned: Hacker entrepreneurs demanded more Unix support, but French manufacturers such as Groupe Bull were unable to provide it. As a result, Minitel services were often hosted on machines built by U.S. corporations such as AT&T, Hewlett-Packard, and Texas Instruments, and so, ironically, Minitel expanded rather than reduced the U.S. presence in French telecommunications.
System administrators encouraged service providers by offering high-quality documentation free of charge. Over two decades, France Telecom published dozens of brochures on user interface standards, terminals, PAVI, protocols, and so on. The quarterly newsletter La Lettre de Télétel kept industry participants informed of the latest technical improvements and business experiments.
French companies expanded the Minitel platform with new types of terminals and peripherals. Terminals with built-in memory functions, chip card readers, and high-resolution color displays began to appear on the market. Most Minitel terminals feature a serial port and multiple display modes, allowing the terminal to be connected to a printer, credit card reader or PC. For small business owners, this flexibility turned the Minitel terminal into an inexpensive point-of-sale system. And long before the advent of the Internet of Things, the Minitel was incorporated into various home automation schemes, allowing remote control of heaters, VCRs, burglar alarms and sprinklers.
With this open platform for innovation, telematics electrified the country, making 1980s France a place of huge digital experimentation and excitement. And, unlike the speculative boom and bust of the Silicon Valley dotcoms, the kiosk system provided a robust business model for Minitel entrepreneurs while enriching quite a few service providers. The technical infrastructure of the Minitel ecosystem allowed the French to utilize a plethora of online services at a time when the online landscape in the United States was limited to local BBSs and nascent “vegetable gardens” like CompuServe.
While it was not the only network to utilize X.25 or videotex technology in the 1980s, Minitel was unique in that it allowed multiple service providers to control their own machines. France Telecom controlled only the network, while in most other countries a single organization centrally managed both the network and the servers of the videotex system.
For example, in the UK, the entire contents of Prestel’s videotex system were hosted on an IBM mainframe located at the General Post Office. The BTX system in Germany was similarly organized. In the United States, all the content of The Source, an early private provider of online information, was served from a single computer center in McLean, Virginia. Even 101 Online, a subsidiary of Minitel that briefly operated in the San Francisco Bay Area, stored its data in an office on California Street. This degree of centralization ultimately discouraged innovation, excluding the kinds of startups from garages and college dormitories that made the Internet what it is today.
Minitel gave service providers considerable freedom to manage their systems, which became an integral part of the Internet. Minitel administrators also adhered to an early form of network neutrality. The network did not favor one service over another and did not discriminate otherwise. Occasionally a service could be blocked for breaking the law (e.g., for serving as a marketplace for prostitution), but any such exception had to be subject to due process, and the system’s administrators could be sued if they acted arbitrarily. These guarantees of fairness contrasted sharply with the situation in the United States, where private network operators could exclude content on a whim to suit their business interests.
Of course, there was a price to pay for Minitel’s advantages. The network used a non-standard implementation of the X.25 protocol that did not allow private servers to connect directly to each other. Instead, all connections were routed through a public data network, effectively centralizing communication between nodes. This restriction was necessary to implement the Kiosk system, but it also required each host to be individually approved by the state.
Routing all traffic through a central network also allowed the state to attempt to enforce a policy of censorship on Minitel. Due to active lobbying by existing print publishers, only incumbent publishers were granted access to the Kiosk. However, potential providers soon began to circumvent this bureaucratic obstacle by printing fake newspapers, collectively known as “ghost presses,” which entitled them to recognition by the state. Others bought and sold their access on the secondary market. In most cases, the Minitel administration was happy to plug these enterprising savages while receiving a third of their income.
Minitel was thus hardly the rigid, static system envisioned by many Internet advocates in the 1990s. Its hybrid architecture, bridging public and private, open and closed, provided a rich platform for innovation and entrepreneurship at a time when online services elsewhere in the world were faltering.
For a generation of French people, Minitel wasn’t about hardware, switches or software. It was the people they interacted with, the services they used, the games they played, and the advertisements for those services they saw in newspapers and on billboards. Many of the services we associate with the Web had predecessors in Minitel. Before Peapod, there was 3615 TMK (Tele-Market), a service that allowed Parisians to order groceries with same-day delivery. Before there was Cortana or Siri, there was Claire and Sophie, services that provided personalized information through natural-language interfaces. Before there was Ticketmaster, there was Billetel. And before there was telebanking, there was Minitel banking.
However, what is most remembered in the popular memory of Minitel are the services that were undoubtedly “pink messengers.” These “pink chat rooms” were sites of flirtatious research that ranged from casual online dating to downright lewd and crude discussions. The “pink minitel” services were not only popular, but also the most profitable. The profitability of these adult-oriented services led to an advertising war between pink providers in print, on television, and on billboards, so that the phenomenon was hard to get rid of even if you had never used a Minitel service. Telematics advocates were both delighted by this enthusiastic embrace of the new technology and concerned about its rainbow hue. One PTT minister lamented, “[I] don’t want telematics to tarnish its image by exclusive use in prodigal communication!”
The emergence of the pink Minitel was the result of both low-tech and high-tech innovations. At the low end were animatrices, a new type of information worker whose job was, in the words of one popular song of the time, to “digitally undress” users. Animatrices were often young people posing as women. Their job was to keep unsuspecting customers online as long as possible. While many animatrices were paid, others were self-described “minitel addicts” who traded their services for free connection time.
The entrepreneurs behind these pink chat rooms, some of whom later came to dominate France’s telematics industry, also developed more sophisticated tools to maximize revenue. Computers equipped with software allowed animatrices to carry on multiple conversations simultaneously. Another practice frowned upon by many in the community, but nonetheless widespread, was the use of bots for online solicitations. Minitel tycoon Xavier Niel employed such automated animatrices, inviting users to “hang out with me in another chat room.”
The rabid popularity of adult services depended on certain privacy protections built into the network itself. Starting at the local gateway, all Minitel connections were anonymized. No usernames or credit card numbers were required, so the chat room providers never knew the real identities of their customers, nor did they need that information to make money. Because billing was handled by PTT, providers received one lump sum per billing cycle rather than dealing with thousands of individual accounts. This payment system, which effortlessly charged the user, was also the reason why Minitel had relatively little advertising.
Privacy and anonymity extended to users as well. Consumers’ phone bills did not indicate which sites they visited. Instead, the phone company aggregated all activity for a billing period into a single bill. That way, an employee assigned work tasks by Minitel could easily sneak in a chat room, pink or otherwise. For some, messenger became the new water cooler (to the dismay of many business owners).
Minitel enthusiasts valued the privacy and anonymity of the network. In late 1984, Minitel engineers added a feature to the terminal that saved the last page visited and allowed the user to restore an interrupted session, much like a browser cookie does today. The public outcry was swift and violent. Editorials in newspapers that (rightly) saw Minitel as a competitor warned that Big Brother had arrived. Some 3,000 terminals were returned in protest. PTT soon abandoned the feature.
Minitel usage peaked in 1993, when users spent more than 90 million hours using various kiosk services at their terminals. In subsequent years, usage declined as home computers and dial-up Internet access proliferated. Loyal users could continue to use Minitel with terminal emulation software, but many others simply moved on. Minitel’s easy-to-use terminal and its simple videotext interface, once so innovative but now proving inflexible, held back further development.
Although hundreds of thousands of users continued to use the system every month throughout the 1990s and beyond, Minitel no longer seemed a shining symbol of France’s telematic future. Rather, it was an unremarkable part of everyday life, no more colorful than a radio or a telephone. In 2012, after almost 30 years of continuous operation, PAVI were closed down and the Minitel era came to an end.
But it would be wrong to consider Minitel a failure. In fact, it offered an intriguing model for fostering innovation without sacrificing the public interest in fairness and privacy. The millions of curious users and risk-taking entrepreneurs who flocked to the platform in the 1980s were among the first to encounter the issues of trust, intimacy, privacy, and civility that characterize life on the Internet today. That grand telematics experiment is over, but it can still teach lessons to the many engineers and computer scientists trying to make the Web better.