PC vs. NC:
Could you find a stronger proponent of the Network Computer than James E. Felton, president of Network Computer News Service? Or a more worthy adversary than WINDOWS Magazine editor Mike Elgan? We played fly on the wall as the two debated the benefits and shortcomings of the NC.
On Cost:James Felton: The Network Computer will replace the PC as the next-generation information appliance because it offers a dual cost advantage that will increase over time. It will be cheaper to buy an NC because you'll be buying less hardware, less often. And it will be cheaper to own an NC because technical people can handle the administrative headaches much more effectively at the server end.
Both these advantages will increase as the NC becomes more widely deployed. A computer that's cheaper to own and operate than a PC is sure to appeal to a larger audience. And, if it's also easier to use, it will appeal to a still larger audience. Add to this the fact that almost anyone can build and sell NCs, and write and sell network programs, and you have the ultimate low-cost information appliance.
Mike Elgan: If PC users wanted to buy less hardware less often, they would. You get what you pay for. Cheap hardware equals low quality. If real people wanted cheap and slow, low-quality computing, we'd all still be happy with our 25MHz 386 machines. No, thanks!
NCs might be easier to administer (it's easy to attach claims to a platform that doesn't exist yet), but you also have to consider the price of unhappy employees. Reader: Ask yourself if you'd stay in your current job if your employer replaced your PC with a slow, low-quality, centrally controlled NC with no local storage. (And without local storage, how are you even going to safely work on your resume?)
On Bandwidth:JF: Soon, we'll be able to send and receive high-quality multimedia rapidly through the network. Bandwidth isn't likely to be a problem for very long. By any measure--nodes, total bandwidth, traffic--the Internet has been doubling every year since 1970. Over the next decade, it will likely expand a thousandfold. While a few areas may be experiencing occasional slowdowns, and a few service providers may have temporarily sold more service than they're able to deliver, bandwidth as a whole is increasing much faster than usage. I've been monitoring three local service providers and three online services for most of the past two years, and I haven't noticed any degradation of performance (except recently with America Online). In fact, most of the six perform much better than they did at any time in the past.
And, if a fraction of the money currently spent on Windows PCs was spent on infrastructure, bandwidth would multiply even faster. Here's some simple math: Sources say 35 million PCs have sold with Windows 95 preinstalled. At an average of $2,000 each, that's $70 billion. If those people had bought WebTVs or something similar for $329, we'd have roughly $58 billion left to spend on infrastructure. Call me stupid, but it seems if we spent $58 billion on infrastructure, digital multimedia would literally fly down the pipeline, and Windows would find its rightful place beside the abacus.
ME: Hmmm. My experience has been that the Internet is getting slower. But even with high-speed bandwidth, the computing experience is going to be poor with $400 worth of hardware. A good monitor alone costs more than that. No one is going to want to listen to tinny, low-range speakers, look bleary-eyed at TV-quality graphics and run stripped-down, bare-bones applications with a slow processor in 8MB of RAM when ultra-powerful Windows PCs are available for just a couple thousand dollars. And since you're talking about the future of NCs, let's talk about the future of PCs. Assuming Moore's law continues unabated, those $2,000 PCs in three years will be 800MHz screamers, enabling text and voice recognition, real-time virtual reality interfaces and gigs of storage. Any company standardizing on NCs will be missing the most exciting era in computing history. There will never be anything the NC can do that the PC can't. PC users will get all the benefits of NCs (server-based applications, and Web and intranet content) and all the benefits of PCs (speed, power, quality, personal storage, flexibility, customizability and extensibility).
On Ease of Use:JF: As we move the complexity to the server, computing can become much more efficient and cost-effective. You won't need to know what a DLL or a printer driver is in order to communicate. You won't need to concern yourself with the technical aspects of computing, so you'll be able to concentrate on communicating. Do you know (or care) what processor your bank's ATM uses? Do you care how much RAM your answering machine has? Most people don't want to be concerned with such technical matters. They just want well-designed products that do the job they were designed to do at a reasonable cost.
ME: If NCs are easy to use, it will be because of the development of new technologies and languages such as Java and ActiveX, and the improvement of older Internet technologies and protocols. Microsoft, Intel and a lot of other companies are working on implementing these and other technologies that will bring radical improvements in ease of use, and the ability to administer applications and configure workstations centrally. In fact, most of the new features in the next versions of both Windows and Windows NT are aimed at the central administration, installation and configuration of workstations. The difference is, with Windows, you'll still have the flexibility to do a custom rollout and determine how much power and control to put in the hands of each user. NCs won't give you that flexibility.
True, ATMs and answering machines are very easy to use, but they don't do much. It's a perfect analogy to the NC.
I think the NC will succeed as a computing paradigm, but it will always be the low road. NCs will replace dumb terminals and ballpoint pens. But NCs won't replace many PCs.
Back to Top
Back to Top