All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
It's been two years now since Oracle CEO Larry Ellison upstaged Bill "a computer on every desk" Gates at a PC industry conference in Paris with his declaration that the PC was dead and would soon be replaced by the "network computer."
The NC, as imagined by Ellison, was to be a stripped-down, idiot-proof machine that relied on the network for its software - essentially the antithesis to the bloated general-purpose machines that have over the years traded efficiency for untamed expandability and modularity. The kicker? This box would break the price barrier - at US$500, it would be the ultimate affordable computer and would become a simple, standard home appliance like the TV or the toaster.
But two years later and the network computer, as a concept, is nearly as bloated as the Wintel machines it was originally fighting against. While Ellison and his successive followers - Sun Microsystems and its JavaStation most prominently - have plodded along without a unified platform, the once NC-mocking Gates has brought to bear a line of dissimilar, but competing, computers known as Net PCs.
The first thing to keep in mind, however, is that both NC and Net PC architects have, at least for the time being, abandoned uniformly any remaining notions of these network computers as consumer boxes. Until the Net gets a zillion times faster, the concept has necessarily been subsumed by the domain of corporate computing.
But Gates' "group of five" (Microsoft, Intel, Hewlett-Packard, Compaq, and Dell) has seen fit to abandon more than just the "consumer" NC; its spec for the Net PC also discarded the "thin" client model that Ellison proposed, and instead only modified its existing, Windows-bloated PCs to fit the bill. The result, is that the Net PC really isn't a network computer at all.
Regardless, touting a sealed box (no floppy or CD-ROM drives, etc.), software allowing corporate MIS department heads to remotely administrate and control them, a smaller footprint, and an anticipated significant drop in total cost, the Net PCs were supposed to sell like hot cakes to cost- and configuration-conscious corporate IS departments. Right? Not exactly.
As hardware companies are incorporating many of the management features heralded by the Net PC into their existing lines of PCs anyway, the difference between the two has become negligible. This has corporate customers and hardware vendors alike scratching their heads and wondering why an entirely new line of computers meeting strict Wintel-mandated specifications is necessary.
"I said a long time ago that the initial Net PCs were going to be somewhat of a kluge," said Greg Blatnik of Zona Research, a Redwood City, California, research and consulting firm that has been active in the network computer area. "Essentially, they took existing PCs and put a padlock on them." Blatnik added, however, that over time a more elegant approach to the Net PC will emerge.
That some of the concepts behind the Net PC have strong merits to customers - like manageability and security (essentially users can't tinker with their own computers) - hasn't stopped some initially confident hardware manufacturers to back off on releasing product this year. Digital, Micron Electronics, and IBM have all put the brakes on their Net PCs, undermining the Wintel alliance, while others like Hewlett-Packard have said they'll only reluctantly offer the machines this fall.
Jeff Moeser, director of desktop and server products at Micron Electronics, said that despite reaching beta on its Net PC, the company has put its box in a "holding pattern." Moeser claimed that a less-than-thoughtful approach went into the Net PC specs and design process. "If you look at the direct players," he noted, "there's nothing there that's really revolutionary." For Moeser, the simplified $500 network computer is still "a doable thing," but only if it's built from the ground up, not by rehashing the Wintel PC.
Yet some industry watchers see the hold-ups only as a perception problem - everyone knows what PCs are, but what the hell are close-cousin Net PCs?
"Intel blew the launch," said Rob Enderle, senior analyst at Giga Information Group, an industry research firm in Santa Clara, California. "Intel gave potential customers the impression that with the Net PC, you were paying more and getting less. Not the most compelling argument if you're trying to convince somebody to buy a particular new technology."
Still others, like Michael Dertouzous, head of MIT's computer science department, aren't very high on what the underlying Net PC concepts mean for the livelihood of corporations themselves. "For highly centrally managed companies, the network computer offers advantages because they can direct software from a central location that will be used by everybody and they can maintain a degree of control. That is not the direction we're headed in the corporate world or the world in general. [That is] gaining power through giving every member in the organization the freedom and flexibility to do what they want to do to accomplish their job."
That network computers would allow a centralized corporate cognoscenti to control the desktops of employees to a high degree by cutting costs on hard drives - which are cheap and getting cheaper - seems to Dertouzous an unwise tradeoff.
Zona Research's Blatnik labels Dertouzous' opinions an "extreme western philosophy," and points out a dichotomy between what users think they need versus what organizations need. "I think we've gone through such a long period of this kind of personal computing paradigm, that people don't recognize that it's not personal at all. It's the organization's computing." While Blatnik allows that some balance needs to be met between control of the worker and the corporation, he remains certain that in computing, "the shift will definitely be toward the server and away from the desktop."
This doesn't play well to people like Dertouzous, who say the power of computing "lies in empowering the tentacles of the corporation." Nevertheless, cost concerns have become Priority Number One in designing networks, and to the extent that projects like the Net PC fail or succeed, the legacy of the network computer has, for the time being, gone full circle from affordable individual empowerment to corporate efficiency. A revolution? Maybe, but not the type you were expecting.
