THE NEW ORDER : Computer Industry Seeks Open Systems

Once upon a time, choosing a computer system was like getting married: It implied a long-term commitment to the chosen partner that could be reversed only at great cost. Computer companies didn’t sell products, they sold information lifestyles that included the machine itself, software, peripherals, cables, maintenance--and plenty of profit was built in all along the way.

Ten years ago, International Business Machine Corp. began to change all that with its personal computer. The PC was a standard product, so it was possible to mix and match equipment from IBM and the various “clone” vendors such as Compaq Computer Corp. that began building IBM-compatible machines from off-the-shelf components. Independent software developers created thousands of programs that would work on any of these machines.

Still, choosing an IBM-compatible solution meant committing to a basic software system from Microsoft Corp. and a computer-on-a-chip from Intel Corp. It specifically excluded any flirtation with Apple Computer Inc., which was resolutely building its own, self-contained version of information technology bliss. And larger, more powerful computer systems were still based on vendor-specific technology.

With innovation continuing to burst from every electronic seam, however, even the remaining monogamous relations between computer vendors and customers have become impossible to maintain. No company, not mighty IBM or often-arrogant Apple, can satisfy all computing needs anymore. Even these once mortal enemies--champions of tightly guarded technology and go-it-alone marketing strategies--are now talking with one another about technology link-ups.

Computer companies of all stripes, indeed, are frantically building alliances and belatedly trumpeting their allegiance to so-called “open systems.” By adhering to basic hardware and software standards, open systems aim to offer enough commonality among machines from different manufacturers to allow them to run the same applications software--such as word processors or spreadsheets--and communicate with one another.

The idea is to repeat the IBM PC phenomenon--which Intel Chief Executive Andrew S. Grove calls “the most important event in the history of information processing”--throughout the industry.

All this is supposed to spur innovation and provide customers with a far greater range of choice in selecting computer equipment. But the rush to open systems is also producing a great deal of pain at some companies, an overload of misleading rhetoric at others, and a level of confusion and conflict among customers that would elicit a knowing smile from a divorcee.

Just when you thought you knew what DOS meant, a whole new barrage of acronyms has been unleashed--ACE, OSF, EISA, and RISC. Simple words like “open” and “standard” and “compatible” no longer mean what they say. A recent Apple advertisement touted the Macintosh as the best DOS-compatible (that is, IBM-compatible) computer around, even though most would consider the Mac to be the very definition of DOS-incompatibility.

The very complexity of modern computing guarantees some confusion. A certain hardware design might be open in the sense that any manufacturer can produce it. But that in itself doesn’t do anyone much good, because a software system for controlling the basic functions of the machine--such as Microsoft’s DOS--is needed before it can do anything.

And even if this basic software operating system is freely available, vendors often add their own proprietary features that make it incompatible with similar machines. An operating system called Unix is often considered the quintessential open system, but there are many different variants of Unix, and an applications program that’s written for one type won’t work on another.

“Open is almost a meaningless term,” says Esther Dyson, publisher of the industry newsletter Release 1.0. “Open to what? The same with alliance--how allied are you? Being open is an attempt to make an alliance with everyone. But the user is always going to be locked in--you’re just locked in to different things.”

Dyson is hardly alone in her skepticism. And some take the criticism one step further, arguing not only that open systems don’t really exist, but that they’re not even desirable.

“It’s almost a religion that open is good, but open also means not controlled--things don’t work very well together,” says Scott Cook, president of Intuit Software in Menlo Park. “What you really want is a benevolent dictator who guards the standard and makes things work together. The Macintosh is a closed system, but for almost any application it’s better than a PC, because the pieces work well together.”

Steve Jobs, co-founder of Apple and now chief executive of Next Computer, a vendor of advanced desktop machines, argues that true open systems would be a disaster for everyone. “Ultimately, the people who are doing the research and development, creating new products and new markets, need to make more money then the clone vendors who do no R&D; (research and development) and just buy the stuff off the shelf. Otherwise there will be no innovation.”

Still, it’s clear that a greater degree of openness, at least, is the wave of the future. A Macintosh today can read an IBM PC floppy disk, or talk to a PC over a network. The old-line companies that have relied on closely held hardware and software technology--Digital Equipment Corp., Unisys Corp., Groupe Bull SA, Wang Laboratories Inc., Data General Corp.--are struggling to transform themselves, embracing the Unix operating system and touting the ability of their machines to link with others.

Even IBM, considered as recently as the mid-1980s to be a dangerously powerful force that threatened to monopolize and stifle the entire industry, has been staggered by the changes. Battered by clone makers, outsmarted by one-time partner Microsoft, and too bureaucratic to keep up with the frantic pace of change, Big Blue resembles a fallen dictator trying to find a place in the new democratic order.

Some big computer buyers are suffering as well. “The established large-systems users have this giant installed infrastructure--thousands of programmers, millions of lines of computer code written for IBM machines,” said Charles Ferguson, a technology consultant and research fellow at the Massachusetts Institute of Technology.

“And now it has to co-exist with all this other stuff. In certain information-intensive industries, there’s a generational transformation where a set of existing companies is getting stuck with obsolete technology. These new technologies will be maximally useful for new, rapidly-growing companies.”

Yet the new order, chaotic though it may be for a time, has created tremendous opportunities, both for shrewd computer buyers and for vendors that can develop successful alliances and pick the right spots for innovation. Sun Microsystems Inc., an upstart manufacturer of high-powered workstations, has become the fastest-growing Fortune 500 company in the country by pushing open systems long before it became trendy.

Sun, founded in 1984, seized on a new type of hardware design called reduced instruction set computing (RISC), and joined forces with American Telephone & Telegraph Co.--inventor of Unix--to create an improved version of that operating system. It then offered to license the technology to any company that might want to make a Sun clone--a radical strategy that was designed to make the design a PC-like standard and thus encourage software developers to write applications.

Sun’s competitors insist, however, that Sun is no more open than anyone else. “You have to distinguish between the reality and the marketing,” said Jobs. “To me there’s only one open system in the world--DOS Windows,” he said, referring to an operating system produced by Microsoft. Sun, he said, maintains tight control over crucial pieces of the operating system software, and Sun clones have yet to make much impact on the market.

Grove of Intel agrees, and considers a lot of the rhetoric about open systems and increased customer choice to be just that--rhetoric. “Fifteen years ago, users had choices too--there was IBM and Burroughs and Univac,” he said. “It’s exactly the same now. What exactly is so open about Sun? They’re a proprietary company looking out for their own commercial interests, as proprietary as they come. Try to operate the Sun software on a Hewlett-Packard machine--you might as well have chosen Burroughs and Univac.”

Sun Chairman Scott McNealy, however, considers such criticism sour grapes. “My reaction to the competitors is ‘grow up and get a life,’ ” he growled. “It’s the Big Lie, an effort to confuse the market.” Openness, he said, only refers to system interfaces, the “hooks” for connecting hardware and software components, and all the Sun interfaces are freely available.

Sun’s success has spurred many competitors to form alliances to promote their own versions of open systems. The Open Software Foundation--which includes IBM, Hewlett-Packard, Digital Equipment, and many other companies--has developed a version of Unix and a so-called graphical user interface (the basic menu that a user sees on the screen) as an alternative standard to Sun.

And Compaq, which exploited the IBM PC standard better than anyone and has shown itself to be a shrewd alliance-builder, is now spearheading one of the most ambitious industry alliances yet: the advanced computing environment or ACE. More than 20 companies, led by Compaq, Digital Equipment, and Microsoft, have agreed to develop computers that use a RISC chip from Mips Computer Systems Inc. and a new software operating system from Microsoft.

The ACE effort provides a good illustration of the confusing dynamics of open systems computing. In theory, ACE is supposed to establish basic standards for a new class of high-performance desktop computers. Customers could purchase an ACE machine from any of a number of vendors with the assurance that it would work with ACE-compatible hardware and software from other vendors.

In fact, ACE actually includes two different hardware options: the RISC chips from Mips and advanced versions of Intel’s widely used line of PC microprocessors. And there will also be two different software systems for controlling the basic functions of the machines: the new operating system from Microsoft, and a variant of the UNIX operating system that’s widely used in technical computing.

Compaq President and Chief Executive Joseph R. (Rod) Canion asserts that ACE nonetheless meets the basic definition of an open system: non-proprietary hardware and software available from a range of different suppliers that can run the same set of applications software (though the software will have to be slightly modified for the different combinations of hardware and software).

“That’s true open systems from the point of view of the user, and it’s going to happen, it’s inevitable” said Canion. Computers buyers have had a taste of it in the PC world, and now they won’t settle for less when they buy more powerful systems. “I liken it to the idea of individual freedom--once it took hold, it was very difficult to stop,” he said.

Alliances, Canion said, are a necessary element in creating standards in today’s heterogeneous computing world. “If it was just Compaq going out and saying ‘Here’s the new standard,’ we would have been laughed at,” he said. Compaq has already led the successful EISA consortium, a group of PC vendors that developed its own technology for shuttling data around inside a computer and thus beat back IBM’s effort to re-insert proprietary technology into the IBM PC.

And Compaq has also shown the way in building another type of alliance--links to component companies and other firms that can fill in the gaps in its technology portfolio. Compaq made an early investment in innovative disk-drive vendor Conner Peripherals, and recently agreed to pour $185 million into Silicon Graphics Inc., a vendor of high-powered engineering workstations and a leader in RISC and graphics technology.

Both types of alliances--those aimed at establishing standards and those designed to provide access to technology--will continue to proliferate. IBM and Apple are believed to be discussing IBM’s possible use of some Apple software, perhaps in exchange for Apple’s gaining the right to use an IBM RISC chip. And in many cases companies that are allied in one area will be opponents in another. Sun and Hewlett-Packard, for example, are cooperating in the development of a new type of programming technology even as they compete in other areas.

Companies that can successfully solve specific compatibility or interconnection problems--such as local area network vendor Novell--should prosper in this complicated new environment. And customers who know what kind of openness they really need--and can cut through the vendor hype and find it--will appreciate all the new choices.

“This is absolutely good for the customer, and it’s also good for the industry,” says James R. Bell of Hewlett-Packard, whose very title--director of integrated information management--is a reflection of the changes in the business. “Growth is limited by the ability to do things for the customer, who was inhibited by the constraints of the old model. In an open systems world we can expand the size of the pie.”