Tim Berners-Lee — The Universal Link and the Semantic Web

<!-- hierarchy:nav -->

Path: OIP › Thinker Reference › Thinkers › Tim Berners-Lee — The Universal Link and the Semantic Web

Shelf: Thinkers · Traversal: self-explaining · hierarchical · voxel-ready
Machine root: OIP tree · Registry

Tim Berners-Lee — The Universal Link and the Semantic Web

§SELF — thinker-tim-berners-lee

What this page is: A profile of the computer scientist who invented the World Wide Web and later proposed a machine-readable layer on top of it. What it explains: The URL as the core invention, the Semantic Web vision, why it failed, and why large language models make it possible now. Why read it: You will understand why the URL matters, what the Semantic Web was trying to do, and why its failure was a reader problem, not a data problem.

What Tim Berners-Lee Did

Tim Berners-Lee is a British computer scientist who invented the World Wide Web in 1989 while working at CERN (the European Organization for Nuclear Research). He created three technologies:

• HTTP (HyperText Transfer Protocol): The rules for how a client (usually a web browser) asks a server for a resource and how the server responds. A protocol is a set of agreed-upon rules for communication.
• HTML (HyperText Markup Language): The format for writing documents that can contain links to other documents. Markup means annotations added to text to give it structure (headings, paragraphs, links).
• URL (Uniform Resource Locator): The addressing scheme that gives every piece of information on the web a permanent, unique address.

The Key Idea: The URL

The URL is Berners-Lee's most important invention. Before the URL, there was no universal way to point at a piece of information. A URL like https://example.com/page does three things at once: it names a resource, it says where to find it, and it says how to retrieve it.

This seems obvious now. It was not obvious then. The URL is what makes the web a web — every document can point to every other document, across any server, any organization, any country. A link in a document in Tokyo can point to a document in Toronto. The user clicks it and the browser knows exactly what to do.

The Semantic Web (2006)

In 2006, Berners-Lee proposed the Semantic Web. The idea: extend the web so that machines, not just humans, can understand the information. The core mechanism was to give every entity — every person, article, tool, service, concept — a URI (Uniform Resource Identifier). A URI is like a URL but more general: it identifies a thing without necessarily saying where to retrieve it.

The vision was that machines would traverse a graph of linked data. A graph, in computer science, is a structure made of nodes (things) connected by edges (relationships). If every entity has a URI and relationships are expressed as links between URIs, then a machine can follow links from one entity to another, building an understanding of how things connect.

Meaning would live in the connections, not in the documents. A document says "Alice works at Company X." The Semantic Web would encode this as a triple: Alice (subject) → works-at (predicate) → Company X (object). Each part is a URI. A machine could then traverse this graph and answer questions like "Who works at Company X?" by following the links.

Linked Data Principles

Berners-Lee defined four rules for publishing data on the Semantic Web:

1. Use URIs as names for things.
2. Use HTTP URIs so people can look them up.
3. When someone looks up a URI, provide useful information, using the standards (RDF, SPARQL).
4. Include links to other URIs so they can discover more things.

RDF (Resource Description Framework) is a format for expressing those subject-predicate-object triples. SPARQL is a query language for asking questions of RDF data stores.

What the Semantic Web Got Right

The idea was correct. Structuring knowledge as a graph of linked entities is a powerful way to represent meaning. The four Linked Data principles are sound. If machines could actually read and traverse these graphs, the web would be a database that answers questions instead of just serving pages.

What the Semantic Web Got Wrong or Left Unfinished

It failed because it needed a reader that understood what it was traversing, and no such reader existed.

The Semantic Web assumed that if you encoded facts as RDF triples and linked them together, machines would be able to reason over them. But a machine reading a triple like http://example.org/Alice → http://example.org/works-at → http://example.org/CompanyX does not know what "works-at" means. It knows the URI. It can look it up. But the response it gets is more RDF, more URIs, more things to look up. Without a way to understand the meaning of the predicates, the machine is just chasing pointers.

This is the same problem HATEOAS faced: links are only useful if the reader understands what they mean. Programmers tried to solve this with ontologies — formal definitions of what each URI means. But ontologies are brittle, expensive to build, and never covered enough of the world to be useful. The Semantic Web stalled because the web of data was built, but there was no agent that could read it.

Why This Is Changing Now

A large language model (LLM) is a program trained on text that can infer meaning from context. An LLM encountering the triple above does not need an ontology. It knows what "works-at" means because it has read that phrase billions of times in natural language. It can traverse a graph of linked data and understand the relationships, because the relationships are expressed in terms it already knows.

The LLM is the reader the Semantic Web was waiting for.

Solid Protocol

Berners-Lee's latest project is Solid (Social Linked Data). Solid is a protocol that gives users personal data pods — storage spaces they control, where their data lives. Applications ask permission to access the pod, and the user grants or denies it. The data in the pod uses linked data interfaces. Combined with LLMs that can read and reason over this linked data, Solid could enable a web where users own their data and intelligent agents act on their behalf.

How It Connects to Other Ideas

• Roy Fielding's REST: The Semantic Web's graph of linked URIs is an extension of REST's hypermedia principle. Both assume a client that discovers meaning by following links. Both failed for the same reason until LLMs appeared.
• Object Interface Protocol (OIP): OIP endpoints describe what they offer. An LLM can read an OIP endpoint's description (which uses linked data patterns) and understand how to interact with it, the same way it can read a Semantic Web graph.

Sources

• Berners-Lee, Tim. "Information Management: A Proposal." CERN, 1989. The original web proposal.
• Berners-Lee, Tim, James Hendler, and Ora Lassila. "The Semantic Web." Scientific American, May 2001.
• Berners-Lee, Tim. "Linked Data — Design Issues." W3C, 2006.
• Solid Project: https://solidproject.org

---

Up the tree

• OIP root — protocol root, zero-context entry
• Thinker Reference hub — full hierarchy map
• Thinkers shelf — siblings on this shelf
• Voxel graph article — how pages link as voxels
• Self-describing protocol

Related on this shelf

• Alan Kay — The Big Idea Is Messaging
• Alfred North Whitehead — Process and Reality
• J.L. Austin and John Searle — Speech Acts
• Barbara Liskov — Abstract Data Types and Distributed Consensus
• Bram Cohen — BitTorrent and Content-Addressed Protocol Design
• Butler Lampson — Protection and Access Control
• Carl Hewitt — The Actor Model
• Charles Sanders Peirce — Signs, Abduction, and Pragmatism

Machine surfaces

• Public page: https://miscsubjects.com/a/thinker-tim-berners-lee
• JSON article: https://miscsubjects.com/api/articles/thinker-tim-berners-lee
• OIP ask: https://miscsubjects.com/api/dispatch?ask=Tim%20Berners-Lee%20%E2%80%94%20The%20Universal%20Link%20and%20the%20Semantic%20Web