Confine the set to a profession and two is more likely, while within a country it rarely exceeds three. This might seem extraordinary and counter-intuitive, but put it to the test. On a recent flight my wife sat next to an Indian who turned out to work for a company I consult with. I picked only one name in that company and it was his manager. Distance one.

While waiting for breakfast in an American hotel I met a pharmaceutical specialist who knew someone I had been meeting the day before. Distance one. In a swimming pool in south-east Asia a man from mainland China knew someone I knew. Distance one.

Might it be that this people distance has an equivalence in the information world? In recent experiments with directories I discovered that COCHFORE is a unique designation of me in a population of more than a million. That is, the first four letters of my name, followed by the first four letters of my street, identify me alone. Perhaps we should not be surprised as the number of possibilities we are looking at is 26 to the power of eight (not quite, because of language limitations).

But ignoring the constraints of language, there are enough characters to provide a unique address for everyone on the planet. So in principle it ought to be a lot easier to find people and information than it is.

My physical address has 56 characters, my telephone 12, my e-mail 25, home page 38, bank account 22 and so on. Also, I personally have more than 30 different addresses spanning home, office, telephones, fax, computer, banks, passport, medical, insurance, car œ But I actually need only one core address of nine characters for everything.

What of documents - do they have an association distance? Computers usually have a filing system at least three folders deep. This takes the people-document distance out to 16 or more. Then there is the subject matter relating them in a more complex and tenuous manner. Dogs, hot-dogs and food for example. Now we have potentially enormous document distances.

Soon we will have more things communicating than people, and they will all have addresses. Because a lot of them will be mobile we cannot associate addressing with physical location. So we need to cut the problem in a different direction.

Ideally, we would be able to remember addresses with ease. And it is within our grasp to create hyperspace addressing to get us to our destination in a very short distance. For example: female, XYZ Company, accountant, educated at Leeds. Or alternatively: a young woman who wrote a Telegraph article on accounting futures about a year ago. Both routes ought to be sufficient to locate the individual, and the distance is less than six, and computers could do it better than us.

Unfortunately, the future of electronic addressing seems to be ever-growing strings of meaningless characters. Machines may be our only hope of finding a human-scale solution, provided we get out of the loop that is.

Peter Cochrane holds the Collier Chair for the Public Understanding of Science & Technology at the University of Bristol. His home page is:
http://cochrane.org.uk