The network isn't the PC

Some wise heads are in agreement that Internet access appliances from here on might not depend on Wintel, the Microsoft/Intel duopoly.
There is little dissension that the future of publishing and media will be wired. The World Wide Web is already reaching into many parts of the media equation already and it doesn't take an Einstein to know that there is more ahead.

Electronic media must have its equivalent of paper or celluloid. At the beginning of the Internet Age, it was high powered workstations in front of highly technical people. The torch has quite successfully passed to the ordinary PC in front of fairly ordinary business people.

But PCs are great horizontal devices, capable of doing everything for themselves right there on every desktop. They must be fed and watered, supported, upgraded and debugged. That makes them expensive in terms of the processor power, memory, mass storage and expandability. They also soak up a great deal of management time - because they are so configurable, there is lots more to go wrong. On a network, a PC is an MIS accident waiting to happen.

No less a figure than Larry Ellison, the outspoken founder and chairman of Oracle, is pointing out that the PC, as a concept, is past its sell-by date. He foresees a future with the Internet everywhere and the PC nowhere!

"A PC is a ridiculous device. PCs are not going to be the centre of the universe anymore," said Ellison in Paris last month. He had the temerity to say this in the presence of one William H Gates III. What nerve! Gates was on hand to tell the world about the PC in your pocket that would be your communications gateway for voice and data as well as your wallet.

My vote is with Ellison on this one. I, for one, wouldn't want my wallet to spontaneously reboot itself or demand that I upgrade the software every year.
If the PC is not to be the medium for our digital dreams yet to come, then what is? A functional definition requires a screen, a keyboard, a pointing device and a communications capability. While this sounds like a diskless PC or X terminal, that's not what Ellison has in mind. For the Web and the Internet to have a lock on the future of media, the gateway cost has to hit $500 to create even greater mass appeal than it has already. There are some manufacturers who are lining up to make that price.

It doesn't have to have an Intel or Intel clone processor, it could be anything that is cheap and powerful. There are a lot of unsung processor chips like Cambridge's own ARM that could step in to do the necessary for Ellison's appliance.

Wyse, the world's foremost terminal maker, is getting ready to launch a similar sounding device. While they are careful to keep their eggs in the X terminal/Win terminal camp, their price target and conceptual aims are right in line with Oracle's epiphany.

Wyse, no longer a PC maker, is very familiar with the elasticity of demand for cheap terminals. The cheaper they are, the bigger the market and the less incentive there is for competition to try to grab a slice of the pie. Their Winterm products are set to deliver a similar kind of service for Windows, Windows 95 and NT applications users - saving money on desktop hardware and management complexity. Ellison's network appliance is just a download away from a Wyse Winterm.

Ellison says he will demonstrate a cheap network computer next year that will push personal computers out of the networking limelight. The networking appliance is being developed with Oracle and unnamed manufacturers (Wyse? Clues are to be found at to sell for about $500. Ellison claims that it will be a "powerful" device offering fast access to the World Wide Web of the Internet along with electronic mail and multimedia capabilities in a meeting with analysts at the end of September.

"They will be able to do a lot of the common functions that are done with the personal computer a lot more cost effectively, and a lot more reliably, and they'll do some things you can't do with a PC at all," sayeth the oracle at Oracle. The network computer could only be used when connected to a host via the Internet or a similar network to the one planned by Novell and AT&T.

Simplicity is the key to keeping the purchase price and the operational price low. Ellison says that it will have only two cables coming out the back: one you plug into the wall for electrons the other you plug into t he wall for bits," he said. Presumably, Oracle will earn a bit of money by providing the software guts for a remote server to store files and applications software. Oracle will also be the originator of the operating software for the new computer, not Microsoft.

Ellison reiterated reports that a single personal computer costs a business about $4,000 to $5,000 a year in maintenance and depreciation, making them too expensive and complicated for many users including homes, school s and small businesses.

But swapping an Oracle network appliance for a PC will have other ramifications for the digital media of the next decade. Content, as ever, will be king. And the applications software that has made Microsoft (and indeed Oracle) millions in the past will change drastically.

Why pay out hundreds of pounds for a word processor like Microsoft Word when (1) you don't use half the features (2) you don't use it more than a few hours a week (3) it's buggy and (4) it will be old hat within a year or 18 months. Metered software use could also save a lot of money for business.

The same could be said of many components in a PC. It would be far more cost effective (and greener, too) to just rent CPU cycles when you need them. Oracle's scheme would put computing power in concentrations where it could be used efficiently. In a way, this is heading back in the direction that computing originally came from - centralised computing with mainframes.

With the sweeping changes that the Internet is to bring to many commercial endeavours, it would be a mistake to imagine that it has to be delivered solely on a PC.