And indeed, why not? Cell phones are usually subsidized by the network operators for the text and call traffic revenue they generate. Increasingly, they have considerable local processing power—and, with the built-in camera, substantial local news-gathering ability, too. And the networks are now offering offline storage for trivial amounts.
Wenchi Chen is best known, in my part of the forest, for his mini-ITX range of motherboards which, amazingly, are forming a growing thicket of wireless mesh boxes providing rural broadband links to people who dont have ADSL or cable, and cant afford satellite. But the interesting thing for me about the low-power platform isnt just the wireless application.
Rather, its the discovery that more and more people are using these things as servers. And again, why not? It may take two or three low-power PCs to match the performance of a top-range Xeon, but the power budget is a tiny fraction.
And in a co-location center, they charge you for your heat output. And so smart guys are buying a half-dozen mini-ITX boxes and sticking them in their co-lo corner—and thats the cue for the Third World.
One machine per home may be a rich boys dream. One machine per village, however, with mobile-phone peripheral access, is another matter. You can work out a power budget for 100 people with a solar collector or a wind generator and a battery that would keep a mini-ITX system running 24-7 all year, and that would power a micro-cell for mobile phones that would conference all local farmers into a community network.
Just possibly, there are a million such villages in China, though I find the thought a bit implausible that Negroponte would manage to sell one of his PCs to each one. A million-order minimum would definitely restrict his market to China; if China does have a million villages, I dont think youll find another Third World country that does. And if you find one, I dont think they are going to spend $100 on PCs even if there were no other costs.
And there are other costs. The $100 price point isnt the point! The things that matter are the installation costs, the running costs and the power budget. As Wenchi Chen said, if you give someone a free PC and then ask them to pay $25 a month for a broadband feed, they wont be able to use it. "It wouldnt be a valuable resource—and also, if they cant afford the power budget, it wont be a valuable resource."
The village server is a concept with a future, especially if its a wireless village server. But to work, it has to have virtually free access to the Internet, and very cheap access to the phone network. Otherwise, its like giving a jumbo jet to a forest-dwelling, Stone Age community.