I don't remember paying to build the road in front of my house. I don't even remember paying for the massive hydroelectric projects that provide me with electricity, or the large distribution of pipes that supply me water. I don't own the roads or power plants or pipes, because I didn't pay to build them. Honestly, I couldn't afford it! I do pay, of course, but only to use all these facilities. Isn't it odd that I can't do the same with information: pay for using it, but not for building and owning the infrastructure?

In most countries I'd pay a public or privately owned utility company to use services over infrastructure that is paid for by the utilities themselves. This is a convenient system where the costs are distributed over a vast number of people, ensuring both profitability to service providers (due to the economies of scale) as well as affordable access to consumers. Information services don't work like that. I have to buy software because I have to own it to use it. I have to buy hardware. I have to pay to 'build the road in front of my house' -- the high-speed optical fibre linking me to service providers. And since I own the software, it's my responsibility to keep upgrading it.

There is absolutely no reason why we cannot have information utilities just as we have those for water or power. Software does not have to be bought to be used -- that's like paying for a whole lifetime's usage. An information utility would charge for what is actually being used -- the information, software, services actually being distributed over the utility's infrastructure.

A utility provides access to a wide range of basic services to a large area or population. A utility requires massive investment to build infrastructure, because it does not expect consumers to pay for it. A utility keeps prices low, depending on volume for earnings. A utility can be very profitable despite being affordable, because the basic nature of its services means that all higher- level activity generates income. For instance, whether you read at night, use a dishwasher or watch TV, you pay for power. Similarly, an information utility could charge you for access to databases, gossip or word-processing.

What is the experience with information utilities so far? Not very good, actually. Minitel, the highly popular public access information system has made the small grey terminals ubiquitious in France. They are provided free, and consumers pay only for the services themselves -- from electronic mail to government databases. Minitel's technology is hopelessly outdated, and doesn't interface with the growing Internet or any of the new 'infobahn' systems. This is not a result of being a utility, but the result of government subsidies and inefficiencies, just as state-owned power utilities in India face enormous theft and losses.

An information utility has to be carefully conceived. The nature of information requires swift responses to change, and utilities have to upgrade their computers more frequently than water utilities change their purification methods. An information utility can't survive without an open, extensible, decentralized architecture. It would be metered, charging for the use of all services. This would eliminate the need for user-purchased software, while enhancing the profitability of software producers.

Information utilities , while paying less for infrastructure than conventional utilities, also have an endless supply of services -- software doesn't get consumed as water does, it is replicated. Software companies can be sure of receiving royalties from utilities, reducing the risks of piracy. Consumers pay less, and everybody wins.

A utility is in fact the only practical model for widespread, leave alone universal, access to information services. It is the only model that makes it practical to provide access to rural, poor, or remote areas. It is a model that can bring the world, and not just cash-rich multinationals, into this new era.