World Wide Web

World Wide Web

A page of hypertext being edited
Invented by	Tim Berners-Lee
Proposed	March 1989, at CERN
Public domain release	30 April 1993
Core technologies	URL, HTTP, HTML

The World Wide Web is a system of documents and applications, identified by URLs, linked to one another by hypertext, and retrieved over the Internet using HTTP. It is an application built on the Internet, and the two are often confused: the Internet is the network, the Web is one thing carried on it.

Tim Berners-Lee, working at CERN, circulated "Information Management: A Proposal" in March 1989.^[1] CERN's problem was mundane: thousands of physicists on short contracts, using incompatible machines, with documentation scattered across systems that each demanded their own conventions. His proposal was a distributed hypertext system with no central index, so no one had to be in charge of it.

By the end of 1990 he had written the first browser, the first server and the first pages. The three inventions that made it work were:

• the URL, a single naming scheme covering any document anywhere;
• HTTP, a stateless request-and-response protocol simple enough to implement in an afternoon;
• HTML, a document format in which a link is an ordinary piece of markup.

On 30 April 1993 CERN placed the Web software in the public domain, with no licence conditions attached. Competing hypertext systems of the period charged server fees. The Web did not, and it won.

The Web's links are one-way and unverified. A page can link to any other without arrangement, notification or permission, and the target need not exist. Earlier hypertext designs maintained two-way links and consistent state, which required central coordination and made systems fragile at scale. Accepting broken links was the concession that let the Web grow without a registry.

Two further properties shaped it. Retrieval is stateless: each request stands alone, so servers can be replicated and load-balanced freely, with sessions layered on afterwards through cookies. And clients are tolerant: browsers render malformed markup rather than refusing it, which lowered the barrier to publishing and, in the process, made strict standards nearly impossible to enforce afterwards.

The graphical browser Mosaic (1993) put images inline with text and brought a general audience. Netscape and Internet Explorer then competed by adding proprietary features, producing a period in which pages worked in one browser and not another; the World Wide Web Consortium, founded by Berners-Lee in 1994, has issued the common standards since.

Client-side scripting turned pages into programs. The addition of asynchronous requests in the mid-2000s let a page update without reloading, and the Web became a delivery mechanism for applications rather than only documents. Later standards added offline storage, video without plug-ins, and typed low-level code for near-native performance.

Berners-Lee's first browser was also an editor: he expected readers to write. The browsers that spread did not keep this, and for a decade the Web was mostly read-only for most people. Wikis restored the missing half by moving editing into the page itself, using an ordinary form and a simplified markup.

• Internet
• Hypertext
• Wiki

1. ^ Berners-Lee, Tim. "Information Management: A Proposal". CERN, 1989.