The proposal that became the world wide web
Sometime in March 1989, Mike Sendall — head of CERN’s computing and electronics group — read an eleven-page memo from one of his software engineers and wrote four words across the cover: “Vague, but exciting.” The memo, titled “Information Management: A Proposal” and dated 12 March 1989, had been circulated with almost no fanfare. In retrospect, it is the most consequential job memo in computing history.
The engineer was Tim Berners-Lee, a 33-year-old Oxford-educated physicist-turned-programmer who had arrived at CERN’s sprawling campus on the Franco-Swiss border near Geneva in 1984. CERN employed thousands of researchers on short contracts — particle physicists, engineers, computer scientists — who cycled through, scattered knowledge across incompatible machines, and left. The institutional memory of one of the world’s largest scientific organisations lived in the heads of people who had already gone home. Berners-Lee’s proposal was, at bottom, a solution to a filing problem.
What he proposed was combining two things that already existed: hypertext (the idea of linked documents) and the internet (the network already connecting computers worldwide). Existing hypertext systems — including Ted Nelson’s long-dreamed Xanadu, in development since 1960 — required bidirectional links: to reference a document, you needed its author’s cooperation. Berners-Lee dispensed with that. His links were unidirectional: anyone could point to anyone else’s documents without permission. It sounds like a technical footnote; it is the reason the web grew faster than any communication medium in history.
Working on a NeXT workstation at CERN, he built three things: HTTP, the protocol for transferring documents across a network; HTML, the markup language for writing them; and the URL, a universal address for anything published online. By 20 December 1990 he had a working browser — named WorldWideWeb, one word — and the world’s first web server running at the lab. The inaugural website, at info.cern.ch, described the project itself. A page about a web, published on the web.
A Belgian engineer named Robert Cailliau, who had independently proposed a hypertext project at CERN, became Berners-Lee’s most useful internal ally. On 6 August 1991, Berners-Lee posted a short summary to the Usenet newsgroup alt.hypertext, inviting collaborators. CERN then faced a decision: patent it, license it, or give it away. On 30 April 1993, the laboratory released the World Wide Web software into the public domain, royalty-free, forever. That act — quiet, unrepeated, and almost certainly unique among organisations with anything worth patenting — meant anyone could build on it without asking.
Within a year there were dozens of browsers. Within five years, tens of millions of users. The web consumed the existing internet — Gopher, FTP, Archie — the way a faster current absorbs a slower one: not instantly, but without reversing.
Sendall’s four words were exactly right. It was vague. The vagueness was the point.
Sources
- History of the World Wide Web — Wikipedia — key dates (March 12, 1989; December 20, 1990; August 6, 1991), the three core technologies (HTTP, HTML, URL), Mike Sendall’s role, Robert Cailliau’s contribution, and the public announcement on alt.hypertext.
- WorldWideWeb History — CERN — the “Vague, but exciting” annotation and the NeXT workstation on which Berners-Lee built the first browser.
- History of the Web — World Wide Web Foundation — the institutional knowledge problem at CERN, Berners-Lee’s insight combining hypertext and the internet, and CERN’s 1993 royalty-free release.