The videotape that you bought in Paris, France, won't play on your VCR in Paris, Tex. The novel you started on your old Apple IIc is stranded on disks that will never be read again. And on your new computer, some deep-buried incompatibility has locked the CD-ROM drive in a death struggle with Windows. Life would be a lot easier if these things were built to work together. How to close the "interoperability" gap has become one of the toughest and most controversial questions of the high-tech era. Everyone agrees that standards can save time and money. The billion-dollar question is: How should standards be determined?
Good-government tradition says technology conflicts should be smoothed over by due process, with formal deliberations of committees and subcommittees of engineers. "Standard-setting is a political process--what you want is a level playing field for all the participants," says Sergio Mazza, president of the American National Standards Institute, which coordinates the U.S. voluntary standards system. Embracing that philosophy last year, the Senate developed legislation that would have allowed the Federal Communications Commission to set telecom standards if the industry couldn't agree.
But a growing number of impatient techies say that formal standard-setting efforts under the auspices of groups like ANSI or the FCC simply can't keep pace with the rapid development of technology. They prefer a Wild West free-market approach to sort things out, even if that means a few big or well-placed companies are able to cobble together a system on their own and dictate terms to the rest of the industry. Their voices are beginning to be heard: The telecom bill in this year's Republican-controlled Senate drops all mention of involving the FCC in promoting or setting new standards. That's fine with Richard S. Friedland, president of General Instrument Corp., which single-handedly controls the cable-television scrambling standard: "In most instances, standard-setting slows down the market rather than helps it."
There's no question that formal standards bodies have made huge contributions in everything from batteries and floppy disks to fax machines and phone calls. Even many libertarians concede that some standards--in telephone service, perhaps--require consensus. But in general, the free-marketeers favor chucking Robert's Rules of Order--arguing that a make-do standard now is better than a formal one that emerges years from now, encrusted with dozens of extraneous specifications. Their model is the Internet software protocol, which was invented by the Defense Dept. in the 1970s and has grown steadily in sophistication and usefulness along with the Internet itself--directed only by a small corps of techies called the Internet Engineering Task Force.
As an example of committee-ism, critics point to the Electronic Industries Assn.'s 10-year effort to create a standard for communication between devices in a home such as lights and appliances. The CEBus specification fills two binders and is supported by AT&T Corp. and others. But it has yet to make an impact on the market. Meanwhile, millions of dollars' worth of products have been sold using a proprietary standard, LONWORKS, that was quickly developed by Echelon Corp. in cooperation with Toshiba. Michael Coffey, executive director of the Indianapolis-based CEBus Industry Council, says his group's process had to be slow to be democratic: "Any company that had any interest in this whatsoever was invited to participate. If you showed up at a meeting, you could raise your hand and vote."
CELLULAR SNAFU. Don't count on Microsoft to ask for a show of hands from competitors on its next major operating-system decision. Rivals claim Microsoft has illegally used its position in operating systems to dominate other software. Indeed, Scott McNealy of Sun Microsystems Inc. wants a key Microsoft standard--the interface that programmers use to write software for Windows--placed under industry control. But Microsoft's defenders say customers have only benefited from its market-grabbing ways. Gartner Group Inc. analyst Bernd Harzog wrote last year that an industrywide group "would focus more on balancing the competing interests of the vendors involved than it would on competing for customers' dollars in the open marketplace."
Europeans and Japanese view such unbridled competition with horror. They're particularly shocked by the snafu over digital cellular. The U.S. industry formally united behind one standard, called time division multiple access (TDMA). But almost before the ink dried, several major players defected to a standard called code division multiple access (CDMA). Partly as a result, the unified European standard, GSM, is sweeping the globe. Japanese or Europeans may split over consumer electronics formats such as VCRs or videodisks, but rarely over core communications technology. "The U.S. has had this cowboy attitude that does it a great disservice," says Frenchman Alexandre Balkanski, chief operating officer at C-Cube Microsystems Inc. in Milpitas, Calif.
Yet even the Europeans acknowledge that American-style flexibility has advantages. Jozef Cornu, acting chief of French telecom-gear giant Alcatel, says that integrated services digital network (ISDN), the international standard for digital telecommunications, is too detailed: "If you specify interfaces in a rigid way, you don't leave room for innovation."
PACE PUSHER. The future of standard-setting may belong to fast movers such as LSI Logic Corp., which designs multifunction chips for such things as TV settop boxes and video-game machines. Sitting at the nexus of many components and companies, LSI can push the pace of decisions. "The shakeout among competing companies will come earlier on in the process, and the strong will get stronger," says LSI Executive Vice-President Brian L. Halla.
In coming years, cheap computing power should make it easy to achieve compatibility with software, lessening the need for companies to agree beforehand on every jot and tittle of a standard. General Motors Corp.'s Hughes Electronics figured that out when it built its digital satellite-TV system while the compression technology, MPEG 2, was still being debated. The resulting incompatibility was easy to fix by putting two kinds of software into decoders, one for the official standard and one for the Hughes version. Why the rush? "It's simple: We had a business to start," says L. William Butterworth, executive vice-president for Hughes unit DirecTV.
Jonathan Swift foreshadowed standards wars in Gulliver's Travels when he imagined an interminable fight between the Big-Endians and Little-Endians--kingdoms whose only difference is in where to crack open a hard-boiled egg. Today, Macintoshes use a "big endian" addressing scheme, while IBM-compatible PCs are "little endian." The good news is that technology is beginning to close the great rift in desktop computing. The new Apple/IBM/Motorola PowerPC chip is, in engineering lingo, "bi-endian." Such patches are likely to become increasingly common--producing compatibility without committees. That bodes well for the cowboy approach. The message: Innovate now; clean up afterward.
In the Middle Ages, a "standard" was a banner marking a rallying point in
a battle. Come to think of it, very little has changed. To wit:
Microsoft has gained enormous power by controlling personal-computer standards. Rivals, crying monopoly, want a key standard called the "application programming interface" placed under industrywide control. Microsoft refuses, saying it's playing tough but fair.
Hoping to thwart a Japanese effort to sell high-definition TVs in the U.S., American industry devised an incompatible and better U.S. version. American digital HDTV isn't available yet, but broadcasters are already cool to it. Still, the gambit did hamper Japan.
U.S. cellular operators in 1990 selected a digital standard for airwave transmission. But several have since defected to a rival format. The splintering has slowed the rollout of digital cellular in the U.S.--and allowed a European standard to catch on around the world.
Standard-setting bodies have devoted years of effort to an elegant, seven-layer framework for computer networking called Mpen Systems Interconnection. But they've been blindsided by companies selling Internet-style networking, which grew up in an ad hoc manner.