I once remarked in a discussion with a friend: “The internet’s output is data, but its product is freedom.”
He replied: “There are assertions of verifiable fact, and then there are invocations of shared values. Don’t mix them up.”
I had intended my statement as an assertion of technical fact. However, on reflection, I realized that compressing the technical argument had flattened it into what sounded like simple flag-waving. To clarify what I meant, it is worth looking at the technical foundations of how the internet was designed.
The End-to-End Principle
The internet’s essential operation is to encode and transmit data from a sender to a receiver. But in 1969, this was not a new capability. We had networks that did this: the telegraph, a telephone network that was nearly a hundred years old, and specialized wire networks for telexes and photo transmission.
To explain the rapid spread and success of the internet, we have to look at how it differed from these older, established networks.
The principal difference lies in the end-to-end principle. Roughly stated:
“The best way to design a network is to allow the sender and receiver to decide what the data means, without asking the intervening network to interpret or validate the data.”
The seminal expression of this idea is found in the 1984 paper by J.H. Saltzer, D.P. Reed, and D.D. Clark, End-to-End Arguments in System Design. The same argument is explored in different terms in David Isenberg’s The Rise of the Stupid Network and Doc Searls and David Weinberger’s World of Ends.
Stupid Networks and Smart Nodes
In older telephone networks, the network itself was “smart.” The switches and central systems were designed to understand, route, and manage the voice signals. If you wanted to invent a new service—such as three-way calling or caller ID—you had to convince the phone company to rebuild the central network to support it.
The internet inverted this model. It was designed as a “stupid” network. Its only job is to move data packets from point A to point B without knowing or caring what those packets represent. Whether the packets contain a line of text, a voice call, an image, or a completely new protocol, the routers in the middle handle them exactly the same way.
Because the intelligence is pushed entirely to the “ends” (the computers sending and receiving the data), the network does not need to be modified to support new applications.
Permissionless Innovation
What made the internet worth adopting in a world already well-provisioned with other communication systems was that the sender and receiver did not have to ask for either help or permission before inventing a new kind of message.
When Tim Berners-Lee invented the World Wide Web, he did not need to ask internet service providers to upgrade their routers. He simply wrote a browser and a server that talked HTTP, and because the network was designed on the end-to-end principle, the packets moved without friction.
The core virtue of the internet is a massive increase in the technical freedom of all its participating nodes. This technical freedom directly translates into productive and intellectual freedoms for its users.
As a friend of mine put it: The internet means you don’t have to convince anyone else that something is a good idea before trying it.
The technical reality is that the internet’s output is data. But by design, its product is freedom.