(FORTUNE Magazine) – JOHN SCULLEY, chairman of Apple Computer, turns to a VCR in his gadget-crammed office and pops in a cassette. ''Let me show you how the Macintosh will work in a large corporate network,'' he says. On the videotape, a manager flicks on his computer and sees a diary entry for the day reminding him that one of his employees is due for a salary review. The manager types a password that links his desktop Mac to the company's mainframe computer, which handles personnel records. The Macintosh retrieves the files, runs a program that compares the employee's salary with those of his peers, and displays the results as a bar chart on the manager's screen. The manager decides how big a raise he wants to give and alerts the personnel and payroll departments electronically. Elapsed time: three minutes. Declares Sculley: ''That's a perfect example of where we think the industry is going. We're trying to create the links that will bring the desktop Macintosh into the mainstream of corporate computing.'' What is remarkable about Sculley's demonstration is not how innovative it is but how overdue. Connecting different types of computer equipment seems so sensible that many customers have never understood why manufacturers didn't do it long ago. Bowing to pressure from companies that have accumulated a hodgepodge of often incompatible computers with brand names from Apollo to % Zenith, the manufacturers are finally coming around. The results will stand the $150-billion-a-year information-processing business on its ear. At age 40 the computer industry -- now the third largest in the U.S., after only automobiles and oil -- is undergoing a mid-life crisis brought on by two developments long in the making: -- Desktop computers have become so powerful that they can perform many of the functions of their bigger minicomputer and even mainframe brethren. -- Customers have become downright adamant that manufacturers figure out how to make their systems work together. All the major U.S. computer makers, along with many of their Japanese and European competitors, are racing to develop systems that allow machines of different sizes and makes to communicate with each other effortlessly through vast networks. The companies are even cooperating with each other in the process -- unthinkable only a few years back. The industry that has transformed the way most people work is about to be transformed by the way most people want to work. The biggest beneficiaries: companies with strong positions in the desktop market, like Apple and IBM, and companies that link a variety of systems together, like Digital Equipment (DEC) and Hewlett-Packard. Computer makers that fail to adapt are likely to be taken over by larger, stronger competitors. (The boxes accompanying this article summarize the status of five key companies in the industry.) Customers too will profit by getting a fatter return on their investments in hardware and software. As in earlier upheavals that redrew the map of the computer industry -- the advent of minicomputers in the 1960s and of personal computers a decade later -- the immediate catalyst for the turmoil is new technology. The jet engine opened a whole new era of aircraft design; similarly an advanced generation of lightning-fast microprocessors is revolutionizing desktop computing. Already reaching the market are powerful microcomputers that cost as little as $5,000 to $10,000 but that can handle large, complex tasks once reserved for machines 20 times more expensive. Five years ago, right after the start of the personal computer revolution, industry experts observed that if the automobile business had developed like the computer business, a Rolls-Royce would cost $2.75 and go three million miles on a gallon of gasoline. Although today's microcomputers cost no more than those of 1983, their far more capable microprocessors offer the , equivalent of ten times better gas mileage while allowing drivers to cruise in comfort at 250 mph. Even more astonishing, says Michael Homer, manager of business development at Apple: ''The rate of improvement in microprocessor technology keeps increasing.'' Breakthroughs in performance are opening up new markets for microcomputers. Many customers who have bought the souped-up new PCs to speed the work they have traditionally used PCs for -- financial spreadsheets and word processing, for example -- are discovering that desktop machines can handle all sorts of additional computing tasks, from engineering design to newsletter publishing. Says Homer: ''As personal computers get more horsepower, more of the traditional functions of larger computers will keep falling into our price band, which is $10,000 and under. In fact, PCs are already a workable alternative to minicomputers for many companies.'' EQUALLY impressive is the increasing versatility of communications networks. Says John Young, chief executive of Hewlett-Packard: ''The central reason customers haven't been getting as much value out of their computers as they would like is simply the inability to tie them together. That barrier is about to disappear.'' Up-to-the-nanosecond networks like the one built by engineers at Apollo Computer of Chelmsford, Massachusetts, let thousands of Apollo employees in far-flung locations share data and software programs regardless of the basic compatibility of the machines they operate (see diagram below). With the network, a division manager using an Apollo workstation to prepare quarterly reports can instantly obtain sales data stored on Wang Laboratories systems in branch offices or get inventory figures from a Hewlett-Packard minicomputer at a manufacturing site. For most companies the era of network computing is not yet a reality, but it is arriving with stunning speed. A strong synergy is emerging between the new desktop machines and modern computer networks. Each makes the other more useful and thus more attractive. Computer industry analyst Mark Stahlman of the Wall Street investment firm of Sanford C. Bernstein estimates that in just four years sales of high-powered microcomputers, along with other components of highly integrated networks, have grown to constitute 10% of the computer business. What's more, they are increasing three times faster than the industry as a whole. ''By the early 1990s these systems will account for the majority of growth in the industry,'' he says. THE IMPLICATIONS for computer makers are clear. The spread of network computing has already benefited companies at the high-performance end of the desktop market such as Apollo and Sun Microsystems, both of which make powerful workstations used mainly by scientists and engineers. Now the technology revolution is toppling the longstanding business assumptions of dozens of larger manufacturers. Most important, the new networks will demolish the idea that customers must stay married to a single supplier simply because they have made an enormous investment in hardware, software, and training. When customers can link different brands of equipment with ease, they will be in a far stronger position to demand competitive bids. Customers with advanced networks are also more likely to buy a variety of specialized computers, much as stereo enthusiasts mix amplifiers, tuners, CD players, tape decks, and speakers from different manufacturers to form a complete sound system. The trend has already begun with supercomputers, which are special-purpose machines built to serve large groups of users. More than 300 engineers share a Cray supercomputer at General Motors' research lab in Warren, Michigan, for aerodynamics studies and simulated crash testing. As data links improve, the network idea will spread to other areas of computing -- such as office automation systems, where specialized computers could coordinate the use of a company's printers, say, or route electronic- mail messages to assorted destinations. Says Sculley: ''The single largest earthshaking fact in the industry right now is the shift to multiple suppliers of information systems. Customers are saying they want the best solution for each of the things they want to do.'' None of the major players in the computer industry is perfectly positioned to take advantage of the intertwined trends to integrated networks and ever more powerful micros. Each company approaches the new era with a combination of strengths and weaknesses. Their strategies differ, but to a large degree they add up to a simple formula: Push for an advantage where you are strong, form an alliance where you are weak. This accounts for the bewildering array of recent lawsuits, joint ventures, and newly launched efforts to develop uniform communications standards. A good example of the strategy is found at Digital, which excels in networking but has never been important in the PC market. Hence Digital's joint technology agreements with Apple and Compaq, two standouts in microcomputers. Even mighty IBM, which practically invented proprietary systems, is taking part in consortiums that promote industrywide solutions to problems of computer incompatibility. With desktops emerging as the crown jewels of computerdom, personal computer manufacturers are hurrying to improve their machines, especially by adding networking capabilities. American businesses have bought more than 15 million personal computers. Less than 30% of them are hooked up to networks, but soon the great majority of them may be. ''By 1991, 90% to 95% of our machines will be used in networks,'' says Bill Lyons, IBM's general manager of personal systems merchandising. Indeed a much ballyhooed feature of IBM's new PS/2 series of advanced personal computers, the so-called Micro Channel, helps the machines link up with IBM minicomputers and mainframes. The company contends that very shortly personal computers will not only be connected through local area networks, but also function as terminals for large data- processing systems. From airline reservation desks to the factory floor, more than 20 million ''dumb'' terminals are in use throughout the U.S., serving as remote links to a mainframe computer. They're called dumb because they can't do any processing on their own, and to use them properly requires mastering strict rules and procedures. The latest personal computers, by contrast, use on-board microprocessors to run internal software that displays information and instructions far more simply.

Apple and other PC makers hope to seize a chunk of the dumb terminal market. Says Sculley: ''There are two major differences between the last generation of personal computers and the models available today. The first is that the new PCs are far easier to use. The second is networking, which is a big part of how these machines will operate. When you get to sophisticated networking, if anything requires more performance than your computer has, you can farm it out to other machines on the network. The personal computer will become the epicenter of the whole industry.'' SCULLEY, the intense ex-president of Pepsi-Cola brought to Apple by company founder Steve Jobs in 1983, is a leading apostle of the new style of network computing. In his autobiography, Odyssey, published last year, Sculley describes how he devised the Pepsi Challenge and adopted the 32-ounce bottle ! in an effort to ''nullify a competitor's strength by changing the ground rules.'' At Apple, Sculley's first forays against IBM were focused on desktop publishing -- an application that took advantage of the Macintosh's advanced graphics. He explains: ''Desktop publishing was our Trojan horse. It got us in the door. We now have infiltrated beyond that with personal productivity applications, like project management software. Now there are hundreds of thousands of dumb terminals that are going to get converted.'' Certainly, with Sculley in charge, Apple is the most marketing-oriented competitor in the desktop business. And the Macintosh -- Steve Jobs's legacy to Apple -- comes closer right now to satisfying the requirements of network computing than any other desktop machine. In fact, the Mac's famed graphic displays and ease of use are widely imitated by other manufacturers -- unlawfully, Sculley claims. Apple is suing Hewlett-Packard and software publisher Microsoft, complaining that they copied the distinctive ''look and feel'' of the Macintosh software in their own products. H-P and Microsoft deny the allegations. Whether Apple can take full advantage of its lead is not yet clear. Four years ago, in an untimely burst of enthusiasm, Sculley and Jobs dangerously overextended Apple by building huge inventories too far in advance of customer demand, and the company flirted with disaster. Now, most industry experts agree, Apple is primed for growth. Smith Barney computer analyst David Readerman predicts that sales will increase 40% this year, to $3.8 billion. Sculley harbors even grander ambitions. ''The next revenue milestone for Apple isn't $5 billion, but $10 billion. We hope to be there in the early 1990s.'' FOR COMPANIES that are already big, networks may be the most exciting growth opportunity. Designing and maintaining large communication systems that permit hundreds, even thousands, of computers to communicate and share data could be an $18-billion-a-year business by 1992, according to the Gartner Group, a market research firm in Stamford, Connecticut. Virtually every large company, from manufacturing concerns to financial service enterprises, will require these systems. Insists Unisys Chairman W. Michael Blumenthal: ''The ability to do well in the industry will not depend on who builds the best machine for the desktop, because it will be a fairly standard machine. The gravy is in connecting everything and being able to service and support the multimachine networks.'' He says Unisys's $310 million acquisition last year of Timeplex, a leading supplier of networking equipment, was a major step in that direction. At rival Digital, President Kenneth Olsen claims: ''Networking is the biggest opportunity for us. We picked that area 15 years ago.'' Digital has indeed been a leader in networking within its own product line. The company's internal network, which links 27,000 machines in 66 nations from the U.S. to Israel to Japan, may well be the most sophisticated in the world. Olsen compares it to an electric-power grid: ''We can plug computers in like generators and send processing power wherever it's needed. When we run out of computing capacity on the East Coast in the morning, our network automatically absorbs more from the West Coast.'' Digital's elegant network suffers from one crucial shortcoming, perhaps the price of pioneering. Many of its most powerful features will not work with computers made by other manufacturers. Unfortunately for Digital, as Hewlett- Packard's Young puts it, ''when you walk in the door, you don't find any customers with nothing. They all have lots of different somethings.'' Communications links, where they exist at all, are fairly crude. In most cases customers send information back and forth awkwardly, like semaphore operators at sea in a storm. Software that works on one make of machine often will not run on another. THE BENEFITS of tightly integrated information networks are many. They eliminate the inefficiencies of maintaining multiple training staffs and software. In the future, say industry experts, networks will function as the brains and central nervous system for all the specialized information processing that goes on in large enterprises. The first great wave of computerization in the 1960s and 1970s consisted largely of automating complex manual systems, such as payroll and accounting. Then, as computers improved, individual departments, manufacturing divisions, and branch offices of large companies bought a variety of new computers to perform more specialized jobs -- such as running an assembly line or producing a magazine. In the third wave, advanced networks could join these islands of automation together seamlessly. BUT NOT without some practical problems. For example, many users of today's computer networks are dangerously vulnerable to the latest scourge of the industry: computer ''viruses.'' These are essentially rogue programs designed by hackers and sneaked into other software to wreak havoc on the innards of every computer they infect. The most virulent viruses instruct a machine to erase useful data, or even to crash altogether. Although networks are not the culprits, they can act as innocent carriers, spreading a virus from one machine to another. Industry experts say that future network equipment will have to be equipped with ''vaccine'' programs that detect computer viruses and prevent their transmission. Customers who dream of networking nirvanas sometimes awaken unpleasantly to the complexities of the job. Those who persist can end up paying dearly for workable solutions. Even the most basic plans may take months, or years, to put in place. Last year chemical maker Du Pont, which buys computers from IBM, Digital, and Hewlett-Packard, decided to set up a unified electronic-mail system for its employees. Getting the three unique networks to exchange messages with each other cost more than $1 million in addition to the basic software for the project. One major obstacle has barred the way to establishing truly flexible networks: lack of uniform standards. Manufacturers of telephone-answering devices, fax machines, or even switching systems for whole cities recognize that their products must share a common network. With computers, by contrast, a multitude of products is already in use and the problem is to design a network that can accommodate them. Creating a network to handle the rat's nest of incompatible machinery already out there is both a technological challenge and a diplomatic nightmare. Seemingly neutral decisions can tilt the advantage to one manufacturer over another. For example, computer companies cannot agree on specifications for universal electronic-mail systems largely because each manufacturer favors a design that would minimize the amount of reengineering necessary for its own existing system. Operating systems present an even tougher puzzle. These sets of computer instructions control a variety of machines, from desktop workstations to supercomputers. For the past five years, AT&T has claimed it had the best candidate for a common operating system in Unix, invented by scientists at Bell Labs in the late 1960s. AT&T licensed it freely. The prime attraction of Unix is that it works on lots of different computers, which makes it ideal for networking. (By contrast, the MS-DOS operating system used on most personal computers is inflexible. It functions only on machines built with microprocessors designed by a single manufacturer, Intel.) There is just one hitch: AT&T is no longer on the sidelines of the computer industry. The main reason AT&T agreed to spin off its Bell operating companies in 1984 was to get into the computer business. THE POPULARITY of Unix increased during the early 1980s. Other computer companies did not mind selling AT&T's product, because they considered the reconstituted phone company a feeble competitor. Its first computers, introduced in 1984, flopped. But the scoffing ceased abruptly last year, when AT&T set up a partnership with Sun Microsystems to improve Unix and build a new line of computers based on Sun's high-performance microprocessors. Says Nina Lytton, an analyst for the Yankee Group in Boston: ''Suddenly the industry saw AT&T and Sun as a competitive Janus -- specifying standards with one face and selling them with the other. Worse, it looked as though Sun would be able to exploit its close relationship with AT&T to get its own Unix-based computers to market faster.'' In May seven of AT&T's biggest competitors rebelled. Led by Digital, Hewlett-Packard, and IBM, they contributed $90 million to finance a consortium called the Open Software Foundation, which is developing an alternative to Unix. Now the two factions are at a standoff, each claiming to be more ''open'' than the other. Frustrated customers predict the impasse will drag on for a long time. For decades efforts to develop broad networking standards have been stymied by competitive considerations. There are no fewer than 23 world standards organizations, including the Corporation for Open Systems -- itself a coalition of 17 leading computer makers and dozens of major customers. All 23 groups demonstrated a number of prototype communications standards in June at a splashy convention in Baltimore called Enterprise '88. But the $50 million event disappointed customers who came to shop for equipment that works right now. They found precious little to buy. Industry experts concede that developing a full range of standards-based products will take years. Some of the largest computer companies are jumping into the gap with their own solutions. IBM has been working for more than two years on a design project called Systems Application Architecture, or SAA. Its twin goals: to provide links to other manufacturers' products and to permit IBM's own diverse computer lines to be integrated for the first time. Says Michael Killen, a Palo Alto, California, computer consultant who has just written a book called IBM: The Making of the Common View: ''The importance of this strategy for IBM cannot be overestimated. It is the company's response to the lead that its competitors have built up in networking. Without SAA, you could write IBM off.'' WITH IT, IBM stands to become much more competitive than it is today. SAA promises to cure IBM's biggest weakness, the incompatibilities among its mainframe, mini, and personal computers. Its highly regarded new generation of products, including the PS/2 personal computers and the AS/400 line of minicomputers unveiled in June to considerable Wall Street applause, are designed to work in large corporate networks and to run the same software. Digital and Hewlett-Packard are also working hard to broaden their networking capabilities to include other companies' products. Says Peter Smith, Digital's vice president of product marketing: ''We can concentrate on integrating other folks' equipment because we don't have to worry about incompatibilities in our own product line.'' Digital has already introduced links to other computer makers' systems, particularly IBM's. Hewlett-Packard offers similar connections for its customers. Says John Young: ''We decided more than five years ago to abandon our proprietary networking schemes. Today we have the widest range of networking capabilities of any company in the world.'' The demand for standards transcends networking. Customers hope that standards will make rival machines essentially interchangeable and pave the way for competitive bidding on large purchases. The seeds of this idea were sown in personal computers when the de facto IBM standard created a huge opportunity for Compaq and other clonemakers to plunge in and compete for a piece of the business. Now the federal government -- which buys $17 billion of computers a year, making it the biggest customer in the world -- and other major users like General Motors and Boeing are calling for ''open'' standards. LAST YEAR, for example, the U.S. Air Force declared that all bidders for a $4.5 billion computer contract had to offer machines that could run the Unix operating system. Infuriated, Digital protested to the Federal Board of Contract Appeals. When the board ruled for the Air Force, Digital withdrew from the competition. Sniffed Olsen: ''There is no sense in bidding on a contract that is filled with arbitrary, useless specifications. People without any job to be done don't make good computer customers. The Air Force will never fight a war so they don't really care.'' Other minicomputer makers, like Prime, Wang, Data General, and Tandem, cannot afford to be so haughty. Their profits have weakened in recent months. They are eager to support Unix -- or any other standard the industry can agree on. Says Data General's Tom West, whose design team was the subject of Tracy Kidder's 1981 best-seller, The Soul of a New Machine: ''We're starting over with a clean slate. At some point you have to break with the past.'' For all the problems companies face in adjusting to revolutionary change, in the future the information-processing market as a whole may grow faster than ever. In a world of networking, desktop computers will be the only general- purpose machines that customers buy. Companies that hope to sell larger or more specialized computers will have to make machines that significantly outperform their competitors'. Inevitably, too, as customers gain greater leverage over their suppliers, computer makers' profit margins will decline even as product development expenses rise. That's a natural situation for industrywide consolidation. The 1986 merger of Burroughs and Sperry to form Unisys marked the opening of a new round of mergers and acquisitions. Others will certainly follow. For the moment Apple is in the best position to turn the forces shaping the industry to its own advantage. Just as the spread of mainframes and minicomputers transformed IBM, Digital, and Hewlett-Packard into towering giants, the new era of network computing could just elevate Apple -- the most innovative personal computer company -- to the top ranks of the industry by the year 2000.

CHART: IBM TOTAL REVENUE $55.3 BILLION (last four quarters)

HARDWARE SALES MICROS $4.9 BILLION MINIS $3.2 BILLION MAINFRAMES $16.4 BILLION

STOCK PRICE JULY 1, 1987 $164 JULY 1, 1988 $127

OUTLOOK The only player with a strong presence in all three computer markets, IBM is working hard toward making its machines talk to each other and to other companies' computers. CREDIT: ILLUSTRATIONS BY ELIOT BERGMAN INC. CAPTION: See above. DESCRIPTION: Financial data on IBM computers; three color illustrations: Microcomputer, minicomputer, computer mainframes.

^ CHART: DEC TOTAL REVENUE $10.8 BILLION (last four quarters)

HARDWARE SALES MICROS $343 MILLION MINIS $4.8 BILLION MAINFRAMES $447 MILLION

STOCK PRICE JULY 1, 1987 $165 JULY 1, 1988 $114

OUTLOOK The leading supplier of minicomputers wants to build large corporate com puting networks. Its continuing weakness: an inability to crack the personal computer market. CREDIT: ILLUSTRATIONS BY ELIOT BERGMAN INC. CAPTION: See above. DESCRIPTION: Financial data on DEC computers; three color illustrations: Microcomputer, minicomputer, computer mainframes.

CHART: UNISYS TOTAL REVENUE $9.7 BILLION (last four quarters)

HARDWARE SALES MICROS $246 MILLION MINIS $1.1 BILLION MAINFRAMES $2.7 BILLION

STOCK PRICE JULY 1, 1987 $41 JULY 1, 1988 $36

OUTLOOK As sales growth slows for the company's Burroughs and Sperry main frames, Unisys hopes to sell more desktops and networking equipment. The likeliest route: acquisitions. CREDIT: ILLUSTRATIONS BY ELIOT BERGMAN INC. CAPTION: See above. DESCRIPTION: Financial data on Unisys computers; three illustrations: microcomputer, minicomputer, computer mainframes.

CHART: NOT AVAILABLE CREDIT: ILLUSTRATIONS BY ELIOT BERGMAN INC. SOURCE: GARTNER GROUP YARDSTICK CAPTION: THE LAST STRIVES TO BE FIRST IN A BOOMING MARKET Sales of microcomputers by the top 100 U.S. makers are growing faster than any other segment of the industry, and will soon surpass those of minicomputers. IBM dominates mainframes and micros but has lost market share -- leadership -- minis. DESCRIPTION: Three charts: Percentage of total sales for computers by top manufacturers in United States for 1981 and 1987.

CHART: HEWLETT-PACKARD TOTAL REVENUE $9.0 BILLION (last four quarters)

HARDWARE SALES MICROS $625 MILLION MINIS $1.9 BILLION MAINFRAMES None

STOCK PRICE JULY 1, 1987 $60 JULY 1, 1988 $55

OUTLOOK The well-regarded minicomputer maker wants to win big in the hot market for technical computers. But it faces stiff competition from DEC and young rivals like Sun and Apollo. CREDIT: ILLUSTRATIONS BY ELIOT BERGMAN INC. CAPTION: See above. DESCRIPTION: Financial data on Hewlett-Packard computers; three color illustrations: microcomputer, minicomputer, mainframe.

CHART: APPLE TOTAL REVENUE $3.3 BILLION (last four quarters)

HARDWARE SALES MICROS $2.3 BILLION MINIS None MAINFRAMES None

STOCK PRICE JULY 1, 1987 $40 JULY 1, 1988 $47

OUTLOOK Potentially the biggest beneficiary of the trend to desktop computers, its chief product, the company must prove that its machines are useful additions to large corporate networks. CREDIT: ILLUSTRATIONS BY ELIOT BERGMAN INC. CAPTION: See above. DESCRIPTION: Financial data on Apple computers; three color illustrations: microcomputer, minicomputer, mainframe.