New York. There's nothing wrong with the Clinton administration's fascination for building its much-touted ''information superhighway'' -- except it probably won't be a highway and the federal government can't build it and shouldn't try. The real challenge is not to get the federal government into the communication business, but to get it adroitly out.
Both President Clinton and Vice President Gore are fond of using the analogy of the Eisenhower-era federal highway program for the job of building a modern, end-of-century, information infrastructure. But the highway analogy is decidedly unhelpful as a blueprint for the information age.
In the 1950s, people knew how to build highways. Construction technology and road design was a lot better than, say, in 50 BC, but in truth, not that much better. U.S. highways were laid out and built in much the same way as the ones the Romans built 20 centuries earlier. And in both the Roman and American cases, only the central government had the resources and the strategic vantage point to put such a system in place -- hardly the case when it comes to information technology.
The most useful learning from the experience of the federal highway program, aside from its scale and cost, may be that none of its more profound consequences -- like the hollowing out of older cities and proliferation of strip developments -- were anticipated by its planners.
Messrs. Clinton and Gore seemed to hit on the highway analogy from the success of the ''Internet,'' a set of high-capacity data connections linking universities and federal research centers, developed mostly with federal money. Internet use has soared, creating an arcane cyberspace party land for wonks, but also speeding the dissemination of new research.
After spending months mining campaign contributions from Silicon Valley executives, Messrs. Clinton and Gore both became impressed, correctly, that the next wave of computer technology would involve connecting computers rather than just making them run faster. The country with the lead in communications technology would be well positioned to take the lead in the global high-tech sweepstakes. America had a good shot at cementing a winning position, and the federal government should help. Again, correct.
The logical, but probably incorrect, conclusion was that the brave new information infrastructure would just be an extended Internet. The conventional wisdom also had it that only the federal government had the financial capacity to pay for the thousands of miles of high-capacity fiber-optic cable that most people expect to be the backbone of the system.
Confounding expectations, however, business is smelling gold in high-tech ''multimedia'' entertainment and business products, and is laying down an information infrastructure faster than anyone imagined just a short time ago. TCI and Time-Warner, which together serve more than a third of cable households, are spending some $7 billion over the next several years on fiber optics. Telephone companies are spending even more.
At this point, it's not obvious that the final information ''highway'' will be a highway at all. Current broadcast technology uses hardly 1 percent of the available electromagnetic spectrum. New technology -- at microwave frequencies and above -- can rival fiber optics in capacity, clarity and freedom from interference. New cableless ''cable'' offerings from Hughes and the newest versions of microwave telephone service under development by Motorola and others could, in theory at least, make much of the frenzied cable laying obsolete before it's ever used.
An ''information ocean'' may be closer to the future reality than a highway. Information might be transmitted by launching tiny packets of specially coded information out into space. Your computer would pick up only those specifically intended for you and assemble them into a coherent message, which could be a personal letter or a hit movie -- no cables, no central switching system required. Or, instead of being beamed into space, they could be pumped into a ''dumb'' fiber network, a sort of universal river of information, with the movies, or memos, or data bases, being reassembled by computers sitting on the river bank watching the stream zip by.
With so much uncertainty, many, probably most, technology and marketing bets will turn out to be wrong. The driving force may be entertainment -- interactive video, home movies, video games distributed by television channel, or that holy grail of home viewing, virtual-reality pornography. Or it may be business -- portable data networks, video conferencing, electronic books and catalogs, home shopping. Or some unexpected combination the two.