Summary

  • The 1989 proposal addressed CERN’s practical problem of finding and preserving information across changing teams and incompatible systems; it was not yet a running public Web.
  • Berners-Lee sketched linked records and a staged investigation under the working name “Mesh.” CERN’s later 1990 project proposal, coauthored by Robert Cailliau, was a separate step.

A proposal about institutional memory

The Web’s origin story is often compressed into a date and a device: a physicist at CERN invented a global network of pages. The surviving March 1989 document is more modest, and more revealing. Its title is “Information Management: A Proposal.” Tim Berners-Lee’s opening problem is how a large research organization can remember what it has learned while its staff, experiments and computer systems keep changing. The proposal’s subject is CERN’s accelerators and experiments, not a public publishing service. Read the W3C transcription; CERN preserves the archival scan as well. Open the original document record.

The institutional difficulty was not a shortage of facts. It was that useful facts lived in many places, were organized for different purposes, and depended on people who might leave. A hierarchical index can make one collection navigable, but it presumes that someone can settle where each item belongs. Research knowledge often crosses those boundaries: an instrument, a measurement, a software package and an experiment may matter to several groups at once. When systems evolve, a record’s original path can become less useful even though the record remains valuable.

Berners-Lee proposed a flexible information pool in which documents or concepts could be connected through references. A node could point to another node without requiring both to sit beneath one fixed branch in a universal filing tree. Links could cross the boundaries of separately managed systems. The point was not simply to digitize a filing cabinet; it was to make relationships survive changes in organizational structure and software.

That proposal is recognizably close to hypertext, but it should not be read backward as a complete specification of the Web. The 1989 text does not describe the later public service in all its parts, nor does it claim a deployed browser-server system. It lays out a direction for information management at CERN. Its architectural sketch and its later implementation belong to different moments.

“Mesh” was the working name

A retrospective note attached to the W3C transcription says Berners-Lee called the project “Mesh” at this stage. He chose “World Wide Web” while writing code in 1990. The 1989 paper uses “web” in ordinary descriptive language too; that does not make World Wide Web the project’s proper name in March 1989. Names matter because they can make a proposal look more finished than it was. “Mesh” captures the linked, many-directional idea before the later name gathered a public identity around it.

The proposal also sketches work rather than reporting an accomplishment. It suggests a practical investigation: understand the requirements, survey available systems, and build toward a portable, extensible approach. It imagines a later programming phase. The first phase is estimated at two people for six to twelve months. That is a planning estimate in a proposal, not evidence of an approved budget, staffing order or completed timetable.

CERN’s preserved chronology says the first proposal circulated for comment in March 1989. In October 1990 it was reformulated as a project proposal with encouragement from CN/ECP management; by November, Berners-Lee and Robert Cailliau had produced the more formal proposal, followed by a prototype. The November 1990 document names both authors and sets out a more concrete project. The CERN chronology places the first prototype and demonstrations later in 1990. Keeping those documents separate prevents the later project plan from being projected onto the earlier sketch.

CERN’s later browser history associates the phrase “Vague, but exciting” with Mike Sendall’s annotation on the proposal. It is a memorable line, but it cannot carry the institutional record by itself. It does not establish formal rejection, approval, funding or delegated authority. The original proposal and the later chronology support a more careful account: an idea was circulated, then reframed and developed into a project. CERN’s retrospective supplies the annotation’s attribution; the surviving project chronology supplies the later sequence.

The proposal goes further than a metaphor for connected pages. It treats people, software modules, groups, projects, concepts, documents and hardware as nodes, with relationships such as “depends on,” “is part of,” “made,” “refers to” and “uses.” That vocabulary was meant to capture how a change in one part of CERN could matter elsewhere. It also rejects the assumption that every useful record had to be moved into a new central store: a gateway might present a hypertext view of existing data, possibly read-only, and let systems be cross-linked rather than replaced. These are design proposals, not evidence that every gateway was built.

What the proposal changed

The proposal’s significance lies partly in the question it made easier to ask. If knowledge is linked across the systems that produce it, can a researcher find a useful path without first knowing the organization chart, file format or machine that holds each piece? The answer was not yet a finished Web. It was a design claim about how institutions might preserve context as their tools and teams changed.

That claim also leaves governance questions open. Who may add or alter a link? Which record is maintained by which group? How does a reader distinguish a current source from an abandoned reference? A flexible network makes it easier for relationships to cross boundaries, but it does not abolish responsibility for the linked information. The proposal describes a way to connect knowledge; it does not make every connected statement authoritative.

A useful historical boundary is therefore simple: March 1989 was a proposal for managing CERN information, under the working name Mesh. The World Wide Web project was reformulated and built through later work in 1990. The distinction gives Berners-Lee’s early document its proper force. It was neither a fully formed global service nor a vague prophecy. It was a concrete institutional argument that information should remain findable even when the people, systems and categories around it did not.

Sources