The net's co-inventor tells BBC Click Online how it all began, when, as an assistant professor of electrical engineering at MIT, he took a leave of absence to brush up on his networking theory.

Dr Kahn never expected to be become part of technology legend

The work that we did was principally on designing what a network would look like.

It was me working alone writing memos on the subject.

I thought, at that time, that this was about as much practical experience as one would really need, to be a good theoretician back in the university.

But it turned out that an agency of the US government, the Defence Advance Research Projects Agency, known as Darpa (it was known as Arpa back then) actually had plans to build a computer network in the country.

At the time, many people didn't think this was a very practical thing to do because it clearly didn't look like a business opportunity and there weren't that many computers around.

But I thought it was an interesting technical challenge, so I was actually the system designer of the Arpanet - the very first computer network.

I called the project 'internetting'

When I got to Darpa, I got involved in the creation of two more nets.

One was using satellites, a kind of Ethernet in the sky, on Intelsat-4, and the other one was a kind of a mobile network where the nodes were packet radios that broadcast to each other, so all the nodes could be in motion, in principle, or they could stay fixed as well.

The whole goal of that effort seemed pretty straight forward at the time: given that you've got the nets, put them together and get the machines on them to work together.

That was the genesis of the project itself.

How to 'internet'?

When I first started the programme I was talking about what we were trying to achieve, which was netting these different computers and networks, so I called the project "internetting".

For computer communications, computers talk in little bursts. They're not continuous like speech.

Vint Cerf was another key player in the creation of the internet Read Click Online's interview with Vint Cerf

So setting up circuits when you're only going to use it for a little bit of time would be about as inefficient as reserving a road from New York to Los Angeles to drive your car, and letting nobody else on that road.

So the idea that you could share it with little bundles of information that were separately addressed was an interesting challenge.

We had a few simple goals.

We had to find a notion of what we called a gateway - today they're known as routers - that would handle IP routing through the net.

[It was] an end-to-end protocol which we called TCP that also had to understand IP kinds of communication that would deal with end-to-end problems, putting information back in order, doing error checking, getting re-transmissions when things didn't arrive, getting rid of duplicates, getting things back in order when they arrived out of order and so forth.

I think we succeeded in those very significantly.

Security issues

In terms of dangers, such as viruses, fraud or identity theft, I don't think we were thinking about that at all when we got started.

We were not really thinking about the dangers and perhaps we should have done

If we had been worried about that, the net might have been better today but we might not have even got there.

Certainly people have asked why we didn't build security in from day one.

We weren't worried about viruses because we were dealing with a very narrow research community that was "colleagular".

They were all friends and colleagues, and many of those systems had no protections on them whatsoever.

It was only many years later when the net became really a public utility of sorts that those things started to show up.

We were not really thinking about the dangers of that and perhaps we should have done.

I wish we had spent more time on that, but again, in the context of what we were doing, we might not have actually got the project off the ground.

We would have spent all our time trying to convince people to let us deal with security technology when we didn't even have the technology that could work without it.

We might not have got there.

Robert Kahn was interviewed for the BBC's Click Online programme.

Click Online is broadcast on BBC News 24: Saturday at 0745, 2030, Sunday at 0430, 0645 and 1630, and on Monday at 0030. It is also shown on BBC Two: Saturday at 0745 and BBC One: Sunday at 0645. Also BBC World.