on top of yesterday's $3000 box, and suddenly you've got a $4000 budget-buster.
Not only is computing mor
e expensive in parts, it's more expensive in time. Sure, you can get bare-bones functionality as soon as you unpack the thing, but after a few months of serious use, any Windows PC will be clogged with dead files and creeping toward a maintenance aneurysm.
The Web PC is supposed to change that, mainly by using the Internet (or whatever the Internet evolves into) as the repository for the bulk of its code -- OS, applications, and data. Things won't get messed up as quickly, because end users won't be able to clog their own computing arteries. Software will be distributed over the 'net. OSes will be upgraded at some central office. And you'll be able to access your data from anywhere -- like you do TV programming.
For some home users and a lot of office workers, the concept works. Say you support 100 workers in your office. Why should you have to support 100 instances of a platform that contains an OS, applications, utilities, and complex, convoluted hardware, when all most users do is word proces
sing and E-mail? Why not, instead, move toward a variant of terminal-based computing, where professionals maintain the central repository of power and software, and meter it out as a resource to those who need it?
That's the image that proponents of the Web PC are trying to sell us. Like I said, it works for a large group of home users and office workers. But there are problems with the model. First, it's unworkable today. Second, there are many people for whom the model will never work. Third, there's the huge cultural addiction to the private ownership of technology, which I won't even begin to dissect here.
Technically, here's why the model doesn't work today: In a word, it's bandwidth. Moving an OS and large applications across the Internet just isn't feasible with today's phone lines or even with residential ISDN. However, once cable companies start to offer high-speed cable modems, this problem may be solved. And inside a company, the terminal model does work: Even basic, cheap, 10-Mbps Et
hernet is fast enough for the majority of everyday corporate applications.
The real issue, though, is the cost of bandwidth. Even when the infrastructure is in place to support the Web PC model, people will still have to pay for connectivity. It may simply be less expensive to load down a high-powered PC with the code you need, instead of loading it over the 'net each time you want to use it. As complicated as a PC is, it works pretty well for a lot of people. Warts and all, a PC or a Mac is an efficient repository for advanced software.
Moreover, a lot of people need to work while untethered from any network. Look at the Laptop Brigade on your next airplane trip. Until each airliner gets a high-speed link to the 'net, portable-computing users will remain tied to their notebooks. In that case, and in many similar situations, the current PC model looks like the only solution.
The whole concept of the Web PC is so crucial -- and is evolving so fast -- that we've changed our editorial calend
ar to address it. We're working on a feature about the topic that will run next month. This month, we're covering an equally important topic: the growing use of the Internet to supplant broadcasting. The two trends are closely related, of course. So pay attention. The fabric of computing is changing as we watch.
Raphael Needleman, Editor in Chief, (
rafe@well.com
)
Not Till It Flies!
