FYI France

File 10.1994a FYI France Essay :


April, 1994

Baby Bell Minitel?
Internet Competition from the French Connection

by Jack Kessler,

For Connexions: the Interoperability Report,
v. 8 n. 4, April, 1994, p. 2-11, available from the Connexions Archive in elegant .pdf at,


The Telecom Giants

So the telecom giants are converging on the previously-tidy little world of inter-networking. So the Viacoms and QVCs and TCIs and Paramounts and Baby Bells -- with their Martin Davises, Sumner Redstones, Barry Dillers, and John Malones (Microsoft and Bill Gates will be in there somewhere) -- are descending: upon the careful, comfortable, rigorously-standardized "academic testbed" which until recently has brought us "BITNET" and "The Internet" and inter-networking generally. This is the old news, now.

But the new, breaking, news is the decision -- not yet made, quite -- of just what approach to telecommunications is going to be adopted by "the giants," or those which emerge victorious from the current telecom merger bloodbaths.

Will they simply plug into the Internet? That seems unlikely: there is much which they appear to want which the Internet does not yet offer, and much which the Internet offers which the mass marketers who run the telcos and cablecos do not much want. The marketers need multimedia and simplicity, for instance, neither of which the Internet does well yet; and they don't need the bland interfaces and arcane command and indexing structures, which so often succeed only in making the current Internet appear to be "information-overloaded."

But is there an alternative? Does the Internet have any competition? The easy marketing answer is: yes -- always -- there are competitors out there, and there will be more. Monopoly is the dream of every product developer, and it's always an illusion.


The competition: Minitel

Minitel is one Internet competitor. You wouldn't know it, from conversation on "the net." "The net" appears to be synonymous with "The Internet" at least to the 2-PhDs-per-household, 6-figure-income world of current US networking.

But the new networking byword, even in the US now, is "general public": this is the market most interesting to the commercial giants which will run US networking in the next century. It is a world also interesting to certain foreign governments which are trying to leapfrog non-ASCII-American-English-speaking populations into the 21st. The Internet does not address this by-comparison-to-academia relatively-impoverished & uneducated & polyglot world of the "general public", yet.

The French Minitel, though, already is "general public". It has been, since its 1972 introduction. The marketing legend is that the government originally gave terminals away for free, loaded the telephone books online, and then stopped printing the books. No one admits this now, although they do say that the printed fone books were very hard to come by for awhile back then...

Today, 17,000 services are offered, from home shopping to reserving items at the Bibliotheque Nationale to the infamous "sex chat" of "Minitel Rose". All this arrives via 7 million terminals -- no longer free, but still cheap -- and many millions more free diskettes and commercial "V.23" terminal emulation programs for Macs and PCs.



The total number of Minitel users? That depends on assumptions about statistics which are as shaky in the French case as they are when applied to the Internet. Minitel knows that nearly 7 million terminals are "out there", with several million more potential terminals in place via Macs and PCs equipped with emulation programs.

Recent estimates of Internet usage rely on a factor of 10: from 2 million known current Internet "servers" they assume 10 human users per server to reach current figures of "20 million Internet users/' The same logic, applied to Minitel, might yield 100 million Minitel users! Specious, no doubt -- the assumptions cloak all sorts of statistical unknowns, such as the number of unused Internet accounts and Minitel terminals lying around in US academic computer centers and French homes, and the number of uses "typical" of US academic use versus French domestic use.

Still, there are a lot of Minitel users now. Comparing Internet/Minitel total user figures is networking overkill "bean-counting." The point is that both systems are quite comparably-huge and are growing rapidly. France Telecom claims that usage-time, a figure which they are able to tabulate accurately, now is approaching 10 million hours during some months.

The basic technology is simple, and perhaps even crude by current networking standards. The Minitel is videotex: cute little alpha-numeric graphics screens with rigid indexing and command structures, putting out text and images at intolerably low levels of resolution -- 25-line pages only 40 characters wide, like those little "information" stands which populated US airports for a while a decade ago -- and at insufferably slow rates of speed, 1200bps.



Minitel's current low transmission speed is a major problem. Minitel New York's able Philippe Belvin explains the problem faced in France itself. There, France Telecom and Alcatel populated the countryside, a decade ago, with smart PADs -- Packet Assembler Dis-assemblers, known in French as Points d'Acces Videotex / PAVs. These were based on the now-outmoded "V.23" norm, which provides for 1200bps maximum input and only 75bps output from a user's modem.

The idea at the time, Belvin explains, was to provide for normal typing speeds for output: terminal users typing at more than 9 words per second were not foreseen, but neither was uploading of large datafiles from emulation-package-equipped Macs and PCs.

For the latter application -- and for fast typists -- "V.22bis", 2400 / 2400bps, and "V.27ter", 4800 / 4800bps, norms now are being used in the US and elsewhere, and 9600 and 14400 are under development. But at home in France there still is the problem of replacing all those very slow old 1200bps PAVs. They say this will be done by end-1994; but we'll see, and it will be interesting to see who wins in the current raging debate over exactly what to replace the old standard with.

Transmission speed, though, is a universal problem. There is high-speed work under way in France. France Telecom presented -- at INTEROP Europe 93 in Paris last fall -- T3 applications (actually E3, 34Mbps rather than the US T3's 45Mbps, but same idea and problem) including entertainment and news video, a LAN interconnection service for "bandwidth-on-demand", and ATM networks.

France has super-computing, and one would expect their active participation -- as a leading participant, second only to the US -- in whatever develops from the current US gigabit testbed developments, now at 1.2Gbps and soon to move to 2.4Gbps.

But French networking practicality -- this seems a contradiction in terms, to those who can remember the sad state of French telephony in the 1960s -- is more remarkable than their equally-astonishing presence in the forefront of networking research. Imaging applications, including extensive video-conferencing, are being channeled to their already in-place national fiber optics and ISDN infrastructure:

Any French office or home now can have ISDN -- two 64Kbps channels plus one 16Kbps "signaling" channel -- for US$60 per month. This may retard the development of higher bandwidth applications temporarily -- times are hard financially in France as they are in California -- but at least they have the applications off the drawing board and under way. Both the "Numeris" domestic ISDN infrastructure and the Minitel are demonstrations of the French commitment to this networking practicality. They may be the "low end," but they also are the "mass market," and "mass market" is the key word for networking development for the mid-'90s.


Reasons for success

Minitel has solved, in addition, a number of basic problems which still confront the Internet: