Long before Sun's ads started telling the world that "We put the dot in dot-com" -- whatever that bit of punctuational bravado means -- the company's slogan was "The network is the computer." The provocative line suggested the direction Sun's engineers thought computing ought to be headed: The company was evangelizing for the Internet when the rise of the Internet was by no means a universally recognized slam-dunk.
But "the network is the computer" always left me a little uncomfortable, too -- because, in the daily grind that constitutes office life, the network wasn't the computer. The computer was the thing that sat on your desk, stored all your stuff and tended to stay just the way you wanted it, at least as long as someone else didn't mess it up while you were out to lunch. The network was the thing that you only vaguely understood and that some support guys in the back room ultimately controlled. The computer was yours; the network was not. That little fact is one of the things that (if you'll pardon the expression) put the "personal" into personal computing.
Now, a couple of different forces are trying to push the industry away from the personal-computing model and toward a more centralized approach. While Sun peddles its latest edition of the hoary network computer or
"network appliance" concept, a bevy of Web start-up companies are promoting a
new era of "Web-based computing," in which we'll do all our work -- scheduling, e-mail, even word processing -- inside browsers and store all our data across the Web itself.
Sun's plan, unveiled this week, centers on a new product called the Sun Ray, a stripped-down workstation consisting of not much more than a monitor and keyboard, like the old "dumb terminals" that used to plug into mainframe computers -- plus a smart-card reader for authentication (think of logging in to a Sun Ray as something like using your bank's automated teller machine). With Sun Rays spread throughout a company's network, Sun imagines that you'll sit down at any desk anywhere and pick up your work from the server wherever you left off -- using a software package called Star Office that Sun recently acquired, and that provides tools similar to Microsoft Office.
Sun's strategy is aimed at the corporate world, whereas the Web-based computing companies that are trying to get us to move more of our work and data into our browsers are targeting the mass market. But both fly a similar banner of economy, convenience and simplicity. And if either really takes off, then the network really will be the computer: Our desktops will lose their independence and become vassals of a central authority, either in the corporate IS department or at the main office of some media conglomerate that has wired your neighborhood.
To be sure, independence has always come at a price. As individual users we've all had to take on more responsibility for protecting our data, learning the intricacies of our systems and dealing with crashes than many of us would probably like. No doubt there's some comfort to be had in handing all this responsibility back to a geek priesthood.
But before we rush to embrace the brave new world of network-centric computing, let's think harder about what we're giving up -- and to whom.
It's easy to see why the Sun Ray might appeal to corporate managers: They want systems that are cheaper and easier to administer. Previous "network computer" schemes -- like the Oracle NC, which Larry Ellison was promoting four years ago -- floundered because they didn't save enough money: PCs kept getting cheaper, and NCs just seemed like less bang for not a whole lot less buck. This time around, if Sun can deliver on its promises and provide relief from the "Microsoft tax" -- the per-seat license fees that companies typically pay for using Microsoft's operating systems and Office applications -- it stands a chance of converting lots of firms to the Sun Ray cause.
But some of the same traits that make the network-centric model so appealing to corporations raise red flags when you transfer them to home use. Corporate managers are no doubt delighted that network computers like the Sun Ray are incapable of playing popular PC games, since their employees aren't supposed to be squandering company time on such diversions anyway. But in the home, the hardware limitations that prevent most network computers from handling complex games are likely to turn off a lot of customers, who will favor low-end PCs instead.
Of course, proponents of Web-based computing don't care whether you're using a Sun Ray, a $400 PC or a "screaming" Pentium III: Their idea is to completely divorce your work from your hardware. As long as your browser is connected to the Net, you're fine, and you can access your calendar, address book, e-mail and projects.
That puts a heavy burden on the browser -- and in case you hadn't noticed, browser development, once the supersonic engine of Web change, has been largely dead in the water for a couple of years now. One reason I don't want to move my work life onto the Web is that I don't want to spend every minute of my computing life inside my browser: It's not fast enough, it's not adaptable enough, it doesn't provide as useful a palette of editing tools as my stand-alone applications and it still crashes more often than I'd like.
The other big problem with the Web-based computing approach is a matter of trust. Do I trust the Web site offering an application like Web-based e-mail or calendaring to keep my information to itself? If the service is free, sooner or later the application provider will itch to take whatever information it has about me and earn some money with it somehow.
Even if I'm not worried about privacy, do I trust the provider not to lose or screw up the data that I depend on? Sure, few of us are very diligent about backing up our own computers -- but at least when we lose data ourselves we know who to blame. When some little Web start-up loses your entire address book, good luck getting someone to complain to on the phone. The recent Hotmail fiasco -- during which anyone could access anyone else's Hotmail e-mail through a simple Web back-door -- does not inspire confidence in the Web-based approach, even when it's being delivered by a behemoth like Microsoft (which owns Hotmail).
Microsoft itself has made it clear that if users are unhappy with what they get for free on the Web, it will be happy to transform its familiar shrink-wrapped software tools into fee-driven network-based services. In other words, instead of buying Office 2000, you'd pay Microsoft a fee to deliver use of it across the Net for a certain time period. Such metering schemes delight the software sellers with visions of endless streams of subscription revenue and an end to piracy problems. But they'll be a hard sell to consumers who still think of software as a buy-once product rather than a pay-as-you-use service.
With all of these different strategies vying for dollars and attention, you can count on one principle to govern what succeeds and what's left in the ever-growing trash pile of technological dead ends: simplicity. Whoever devises a system that is truly easier to use can count on some kind of market.
Sun has a lot of experience creating powerful, reliable, expensive servers that require highly trained personnel to run, but it has yet to show the world that it knows how to build intuitive interfaces for the everyday user. The Web-based computing companies are only beginning to figure out how to make their services as easy to use as the stand-alone desktop applications they're designed to replace -- and those applications aren't exactly models of simplicity to begin with.
With the right design, either of these approaches could take off and make "the network is the computer" a fact rather than an aspiration. In the end the distinction between network and computer hardly matters: All that counts is whether it's possible to do your work quickly and reliably. But for now, I'm not giving up my personal-computer independence -- not until Sun, or whoever, can prove that doing so will pay off.