With the 20th anniversary of the web's conception closing in this week, the lesson of its arrival is this: don't set out to invent things because you will invent the wrong thing.

There was a time, in the 1970s, where the question ‘what is the Internet for?' would have seemed absurd. The answer was obvious: message communication between academics with the right equipment (educational computers and a network), more or less extending how the US military had used the same system.

Sir Tim Berners Lee is credited with putting down the first draft of a very different idea of the Internet, the world wide web, in March 1989, but as interesting as what he came up with is the fact that he did so while working under the wing of the one of the world's most prestigious physics labs, CERN. Whatever CERN was invented to do, nobody would have foreseen the most important computer application in recent history coming out of its work.

That's always been the great and ineffable thing about institutions such as CERN. By funding scientists and engineers to do non-applied research in its purest, most expensive and riskiest forms, they are free to come up with answers to questions that were not actually being asked when they started. It's not a form of science that is popular with the tax payer because it sounds like it sounds a lot like a romantic job creation programme for extremely bright minds, but its role in science is inarguable.

The web couldn't be invented by a private company because it would have been beaten to death by commercial rivals before it became useful, or copied by them and mucked up. Rather like the Internet itself, it had to be done in a commerce-free environment, by accident, and without private ownership under patent. The last part isn't an original observation, but it is true nonetheless. Private enterprise isn't an efficient way to make big things quickly.

What Berners Lee invented was significant, but so what where and how and why he invented it.

But let's not write the money men out of history completely: the first ever web server ran on the ‘joke' computer platform know as the NeXT. Invented by then Apple émigré Steve Jobs, like the web, nobody could work out what it was for. If only the world came up with such useless inventions more often.

Next blog: But is the web a good thing? Not everyone thnks so.