(FORTUNE Magazine) – Except for the urge to throw a shoe at it from time to time, I'm not moved to interact much with a television set. Most people aren't, which is why so many breathless visions of interactive TV over the years have turned into expensive disappointments. Today the interactive TV dream has resurfaced again, what with Michael Armstrong and John Malone trying to justify AT&T's $48 billion plan to buy Tele-Communications Inc.

We are witnessing the resurgence of convergence--the latest incarnation of the old notion that computers, communications, and consumer electronics are merging into one all-digital delivery medium. This year's version dares to dream that we'll change the way we buy. I read in Time that soon we'll be kissing our malls good-bye in a rush to the online shopping paradise, mousing and clicking through Yahoo on our way to the New Economy.

Not so fast, I say. New technology merely presents opportunities for economic and social change; even though our imaginations leap forward to what might be, real change is almost always gradual. It's seldom clear in advance which innovations will ultimately make the greatest difference for people. All you can bet on is the short term, where it's often wisest to apply new technologies to proven markets--as unexciting as they might seem.

History bears this out. Consider that this year is the 50th anniversary not only of the transistor but also of the hologram. Both seemed promising in 1948, but one invention has given rise to a multibillion-dollar industry while the other remains largely a curiosity. It required a decade for the CD to overtake the cassette tape as the most popular medium for recorded music. The daily newspaper, which dates from the second half of the 18th century, still attracts as much advertising as television, which was supposed to supplant it. And then there's the lesson of Sylvan Goldman. He never got the rock-star media treatment that today's wizards of electronic commerce enjoy, but he had an economic impact that many of them, as yet, can only hope for. An Oklahoma City grocer, he made possible the supermarket in 1936 by inventing the lowly shopping cart, wobbly wheels and all.

What does all this mean for AT&T and TCI? Well, the pizzazz in the deal lies in the possibility that the combination will bring about higher-speed access to the Internet and increased access to digital video. Those services, made possible through cable modems and a new generation of set-top boxes, could be big growth businesses for the new company. But that's the unknowable future. Less noticed in the merger daydreaming is a proven, workaday application that will be bigger sooner--cable telephony, the ability to make voice calls over TV cable.

One day the Internet may become a medium for the masses. For now, it's more like television before Milton Berle--a novelty that most people don't yet find compelling. We may be entering the Internet era (almost a third of all U.S. households own modems), but the average American spends about 30 hours a year online, compared with about 1,552 hours watching television. Online shopping accounts for less than 1% of retail sales.

Everyone, however, uses the telephone. Annually, Americans spend twice as much for local phone service--an estimated $50 billion--as they do for basic cable television service. They spend 25 times more on basic cable than they do on pay-per-view movies, today's closest approximation to the video on demand that futurists insist is coming soon. Put another way, cable operators have to carry only about 2% of local telephone service to generate more revenue than they do selling pay-per-view.

As it turns out, cable operators are already carrying more than 2% of local phone traffic--in some cases, much more. Just in the past 18 months, such providers as Cablevision, Cox Communications, MediaOne, TCI, and Time Warner have begun offering telephone service to subscribers in Atlanta, Hartford, Los Angeles, Rochester, N.Y., and elsewhere. The various services, which to an end-user seem just like regular phone service, have fewer than 50,000 subscribers nationwide. But, depending on location, these small numbers of subscribers represent between 10% to 20% of those to whom the service is available. In Orange County, Calif., Cox Communications has managed to get 17% penetration in the households it serves. Abroad, where telephone service often is worse and competition weak, cable operators routinely divert as much as a fourth of the residential voice business, sometimes more than a third. In the U.K., more people use cable for phone calls than use it for TV. By contrast, after a selling effort lasting almost two years, the largest provider of Internet service over cable networks, @Home Network, has signed up only 147,000 subscribers, or less than 2% of its total available market.

Cable telephony is also compelling because it's cheap. The additional cost of providing telephone capacity on modern cable networks is extremely small. What's more, revenues from telephony help cable operators pay to upgrade their networks further to handle glitzier services.

While the business is lucrative for now, the profits may not last--at least not for the service providers. Telephone companies are fighting to protect their core voice business. The short-term result for the cable operators and the telephone companies is likely to be rising costs (as they scramble to upgrade and outdo each other's new service offerings), falling prices (as they undercut each other to woo customers), and, eventually, disappearing profits.

The outlook is brighter for the companies that supply hardware for cable telephony, such as ADC Telecommunications, Antec, Motorola, Newbridge Networks, Northern Telecom, and Tellabs. At present such equipment relies on the technology that ordinary telephones use, known as circuit switching, but it's rapidly being adapted for the so-called packet-switching technology that will let cable operators deliver voice calls over the Internet, which is expected to be big business one day (again, a long time before Yahoo wipes out the malls). ADC, for example, recently demonstrated Internet telephony on its Homeworx cable telephony system. CableLabs, the cable-industry research consortium, is also testing Internet telephony solutions, as is Com21, which supplies modem systems to the cable industry. (I am an investor in Com21, @Home Network, and Tellabs.)

Cable telephony creates sales opportunities for companies that make the so-called gateways that connect telephony systems to the Internet, such as Clarent, of Redwood City, Calif.; eFusion, of Beaverton, Ore.; Vienna Systems, an affiliate of Newbridge Networks, Kanata, Ontario; and VocalTec, of Northvale, N.J. Lucent Technologies has also recently entered this market.

There are obstacles, of course. One is power. Today it's not too difficult or expensive to ensure that cable telephones work even during electric utility outages, just as ordinary telephones do. But that's not the case with Internet telephony. Net phones also present reliability problems. For example, unless a line is busy or damaged, an ordinary telephone will connect with the number being called. The Internet, however, is a best-efforts technology; depending on the amount of network traffic, it might or might not be able to complete a call to 911.

Reliability is a big marketing issue. The other day I called my cable operator (Time Warner, which also owns the publisher of this magazine). A recording informed me that seven cable channels were experiencing technical difficulties, which, it was hoped, would be repaired as soon as possible. Would I trust this company with my telephone business? I was tempted to throw a shoe.

RICHARD A. SHAFFER is founder of Technologic Partners, an information company focused on emerging technology. Except as noted, Shaffer has no financial interest in the companies mentioned. For an expanded version of Watch this Space, visit www.tpsite.com/tp/fortune/ If you have comments, please send them to shaffer@technologicp.com