Robert Kavner, American Telephone & Telegraph Co.'s top computer executive and architect of its hostile takeover move on NCR Corp., was clearly blindsided by the question.

A technology analyst, doubting the wisdom of AT&T's wanting to combine its computer business with the unwilling NCR, asked Kavner if the proposed marriage could be a success. In fact, Kavner was asked, could he name any high-technology merger -- even a friendly one -- that had been successful?

Before a room crowded with computer executives and Wall Street analysts, the embarrassed -- and still a mite annoyed -- AT&T executive mumbled that he couldn't.

It's not surprising. Compared with other industries, mergers between technology companies have been rare. Of those that have occurred, the largest and most notable have fallen far short of their original goals -- or, even worse, failed.

Unisys Corp., formed from the merger of Sperry Corp. and Burroughs Corp., is fighting for survival. International Business Machines Corp. acquired Rolm Corp. but subsequently sold it for less than it paid. Hewlett-Packard Co.'s purchase of Apollo Computer Inc. has failed to help H-P gain market share in computer workstations. Advanced Micro Devices Inc. acquired Monolithic Memories Inc., but the merger of the two semiconductor firms has added to AMD's disappointing earnings. France's Groupe Bull SA acquired Honeywell Inc. but continues to lose market share in computers.

Such a dismal track record raises questions about the wisdom of AT&T's hostile bid for NCR -- a deal that, if successful, would be the largest U.S. technology industry merger ever. Problems in merging technology firms have been daunting enough -- even for much-smaller firms.

Such disappointments, in a nutshell, are one of the primary arguments that NCR Chairman Charles E. Exley Jr. has used in his vocal opposition to AT&T's bid since it was launched Dec. 3.

Among the reasons most cited for the poor results are the difficulties of blending incompatible technologies and the brilliant -- but often temperamental -- engineers responsible for them. But underlying it all, experts say, is the Herculean task of stitching together two large, separate operations, while still keeping pace with rapid-fire advancements in technology.

In the end, some experts now contend, technology is one industry where two is not necessarily stronger, smarter or better than one.

"It sounds like hyperbole, but no one I know can think of a single example of where a large high-technology merger has been really successful," said Richard Shaffer, the New York technology analyst who put Kavner on the hot seat. "And it's hard to see how AT&T's play for NCR would be any different."

Indeed, some companies seeking partnerships have specifically avoided mergers. For example, Japan's Fujitsu Ltd. acquired just under 50 percent of Amdahl Corp., a mainframe computer maker. Rather than attempt to buy a larger interest and merge the firms, Fujitsu has left Amdahl and its management virtually alone.

Software an Exception Software publishers have fared better than hardware firms in merging operations. Computer Associates International Inc., one of the nation's largest publishers of mainframe and other business software, was built through a series of acquisitions and mergers. Other smaller publishers, including personal computer software leaders Borland International Inc. and Ashton-Tate Corp., have expanded operations successfully by purchasing smaller companies.

But software firms don't face the same obstacles to a successful combination as do computer makers. They don't have to merge into a single entity huge investments in manufacturing equipment and plants. They don't have the same entrenched installed base of equipment and users who must be protected and served as technology advances. And software programs can be accumulated into a library of offerings, while hardware manufacturers often believe that they must choose between competing products and select just one to survive the merger.

"History would say that mergers between computer makers are not a good idea," said Ulric Weil, a technology analyst in Washington.

Unisys is the favorite post-merger whipping boy of the high-tech industry. Formed in 1986 with Burroughs's $4.8 billion friendly purchase of Sperry, Unisys was supposed to breathe new life into two large, but technologically lagging, old-time computer makers.

"The Power of Two" was its slogan. And for a while it looked as though it might work. Within the first two years, stock in the merged company was trading at $50 per share, giving the company a market value of $7 billion.

But today, Unisys is under siege. Integrating the product lines of the two computer makers, neither of which was compatible with the other, proved far more difficult than originally imagined.

Furthermore, the merger came just as the computer industry was undergoing fundamental changes in technology that reduced the importance of the mainframe models made by the two original companies and increased the importance of personal computers, workstations and other "open systems" models that can easily be connected with one another. At the time of the merger, neither company had yet to fully integrate these systems into their product lines.

After continuing losses that have exceeded $1 billion and repeated rounds of layoffs, the company finally introduced its first integrated open systems product line in October, four years after the merger. Meanwhile, its share of the mainframe market had already shrunk to about 4.7 percent as of October from 8.2 percent in 1986, and it was starting almost from scratch in other markets.

"We still believe the concept behind the merger is a correct one," one company insider said. "But our problem has clearly been implementing that strategy. ...

"We'd have to agree that the technical differences were more consuming than we initially realized and they did take time away from other activities, including keeping up with technological advances." Today, Unisys shares are trading in the neighborhood of $2 to $3 each and the company has a market value of about $450 million.

"The power of two. ... Who would have ever thought that {number} would be its stock price?" scoffed Scott McNealy, the often tart-tongued chairman of Sun Microsystems Inc., the Silicon Valley workstation maker.

Hewlett-Packard, which bought Apollo Computer for more than $500 million last year, has run into similar problems trying to blend the two workstation manufacturing operations.

Incompatible technology held up introduction of a unified product line, while bi-coastal corporate culture clashes heightened tensions. Although the merged company did finally field a next-generation workstation, uncertainty among customers about which company's technology and approach would ultimately win out sparked defections, primarily to Sun Microsystems.

Hewlett-Packard, whose primary goal for the merger was to become the world's largest workstation manufacturer, has seen its market share slip to 23 percent from 26 percent, and the company remains No. 2, behind Sun.

Analysts say there are dozens of similar stories, involving primarily smaller companies. One large company that tried the acquisition route -- and failed -- was IBM, the world's largest computer maker.

In 1984, IBM acquired Rolm, a maker of telephone equipment, for $1.25 billion. Its goal was virtually identical to that of AT&T in pursuing NCR: to become a powerhouse in the emerging blend of computer and telecommunications technologies.

But nearly four years later, IBM sold Rolm to the German technology giant Siemens AG for about $100 million less than it paid. Merging the firms' technologies and corporate cultures was a problem. Also, after the merger, IBM decided that linking with software firms may help it achieve its goals better than combining with hardware companies.

"After we bought it, we discovered there were other options to protect our flanks from the convergence of telecommunications and computers," explained Robert Ripp, IBM's treasurer. "Our position now is to become equity partners in the businesses we're interested in. ... It's better than just gobbling up a large number of smaller companies."

Ripp said experience has taught IBM that not only can its money be stretched further by making smaller investments in a wider variety of companies, but it can avoid the difficulties some companies have blending with IBM's unique corporate culture.

The Smaller, the Better Although big-name, big-bucks mergers have not fared well, some analysts and technology entrepreneurs argue that combinations of smaller companies can have a better shot at success.

In these instances, the purchaser gets much-needed technology while the seller, often the company founder, is freed from administrative chores to pursue research, usually on behalf of the new company. Often, too, these mergers allow the seller to get out from under financial burdens that have slowed the company's operations.