How the World Got Wired · The First Message

What Happened Next

The ARPANET grew. By 1971 it had fifteen nodes. By 1973 it had connected its first computers outside the United States — one in Norway, then one in London, through a satellite link and an undersea cable. Researchers and graduate students at the connected universities started using it for things its designers had not really planned for. They sent each other notes. They argued about software. They shared files. They played games.

The first thing that became wildly popular on the ARPANET was something the engineers had thrown together almost as an afterthought, in 1971. A man at BBN named Ray Tomlinson wrote a small program that let one user on the network send a short text message to another user on a different computer. He needed a way to separate the user’s name from the computer’s name in the address, so he reached for an unusual character on his keyboard — one that wouldn’t be confused with any normal letter or number. He picked @.

He had just invented email. And he had just decided, more or less on a whim, that every email address in the world would forever have an @ sign in the middle of it.

By the mid-1970s, more than half of all the traffic on the ARPANET was email. The network that had been designed to share expensive computers was instead being used, mostly, by scientists writing each other letters. The technology was carrying something its designers had not predicted: human conversation.

But there was a problem.

The ARPANET worked beautifully. It used Paul Baran’s packet switching, Donald Davies’s word for it, Larry Roberts’s design, BBN’s hardware, and a set of rules — protocols — that the universities had worked out together for how their computers should talk to each other.

The rules, though, were the rules of the ARPANET, and only the ARPANET. They worked when every computer on the network was the same kind of node, connected through the same kind of line. They did not work for networks that were built differently — packet radio networks the military was experimenting with, satellite networks linking ships at sea, small networks growing inside other companies and other countries.

By the early 1970s, it was clear that the ARPANET was not going to be the only network in the world. There were going to be many networks. Some would be in the air. Some would be under the ocean. Some would be inside a single university. Some would span the planet.

How could all these different networks ever talk to each other?

Two men sat down to answer that question in 1973. One of them was Bob Kahn, the engineer who had built the IMP. The other was a young Stanford professor named Vint Cerf, who had been a graduate student at UCLA on the very night Charley Kline typed lo.

What they invented is the reason any computer in the world can talk to any other computer in the world today. It is the reason your grandmother’s video, leaving her phone in Vietnam, can pass through dozens of completely different networks owned by dozens of different companies in dozens of different countries, and still arrive, intact, on your screen.

The thing they invented is called TCP/IP, and it is the closest thing the internet has to a constitution.

That’s Chapter 5.

All 24 chapters of How the World Got Wired · Library · All courses