The Reluctant Telegraph: On the Unanticipated Structure of the First Web Addresses

When Tim Berners-Lee wrote the blueprint for the World Wide Web in 1989, his document, "Information Management: A Proposal," was primarily concerned with creating a "large hypertext database with typed links." It was a vision of connection, of trails blazed through knowledge. But it said remarkably little about the roads themselves—the addresses. The URL, as we know it, was an afterthought, a practical necessity that emerged not from grand design, but from the pragmatic need to tell one piece of the nascent web from another.

Before there were Uniform Resource Locators, there were simply… numbers. The very first website, hosted on Tim’s NeXT computer at CERN, was accessible internally by its IP address. To share a document, you would have had to tell someone, "Go to this string of digits." It was the digital equivalent of giving someone a set of geographic coordinates instead of a street name. It worked, but it was fragile, inhuman, and entirely unsustainable for a system meant for global, public use. The need for a more durable, memorable naming convention became immediately apparent.

The structure they landed upon was not invented from whole cloth. It borrowed, almost telegraphically, from an existing hierarchical filing system. The forward slashes (/) that now define every URL’s path were a direct import from the Unix operating system’s method of navigating directories. The protocol prefix (http:) was a nod to other internet protocols like ftp: and news:. This was not the work of an architect drafting a utopian city plan, but of a tinkerer assembling a functional tool from the best parts available in the workshop.

This ad-hoc origin story explains so much about our contemporary struggles with URLs. The first addresses were descriptive labels for the file structure of a single server. A document’s path literally described its location in a folder on a specific machine. The creators were building a system for their colleagues at a physics lab, not for a global commercial and social ecosystem. They could not have anticipated the sprawling, dynamic, database-driven sites of today, where a URL like `/blog/post/12345` points not to a file but to a query result.

The profound legacy of this reluctant beginning is the inherent tension between a URL’s technical function and its human function. The system was designed to locate a file on a network. It was not originally designed to be a permanent, meaningful, and shareable title for a piece of content. The concept of a "permalink"—a URL intended to last forever—was a cultural adaptation layered onto a technical specification that had no such innate guarantee. The entire practice of redirects, canonical tags, and thoughtful site structure is our ongoing, collective effort to retrofit permanence and meaning onto an addressing system that was built for a much smaller, simpler world. We are, in a sense, constantly trying to teach an old, pragmatic telegraph new, eloquent tricks.

Notes & further reading

A few pages I came back to while writing this: