World Wide Web

What the World Wide Web is

The World Wide Web is a system of linked documents and other resources that people reach over the internet. A program called a web browser asks a web server for a resource by its address, shows it on screen, and lets the reader follow links from one resource to the next.¹

Tim Berners-Lee, a British scientist working at CERN, the European particle physics laboratory near Geneva, proposed the web in March 1989. By the end of 1990 the first web server and browser were running at CERN.²

The idea at its center is hypertext: text that contains links to other text. The 1990 project proposal described hypertext as a way to link and reach many kinds of information “as a web of nodes” that a reader can move through at will.³

Three technical pieces make the system work, and all three are still in use:

• URL (Uniform Resource Locator): the address of a resource. It names the host computer first, then the page on that computer after the first slash.¹
• HTTP (Hypertext Transfer Protocol): the rules a browser and a server use to request and deliver pages. It was designed to be stateless, meaning each request stands on its own, and it runs over the internet’s existing TCP/IP connections.¹
• HTML (Hypertext Markup Language): the text format that marks the structure of a page, such as headings, paragraphs and links. It applied an existing document standard called SGML.¹

Reading a web address

Take the address of the restored first web page: http://info.cern.ch/hypertext/WWW/TheProject.html. The part before the colon, “http,” is the scheme, which tells the browser which protocol to use. The part after the two slashes, info.cern.ch, is the host, the computer that holds the page. Everything after the next slash is the path to the page on that host.¹⁴

The general rules for these addresses are set out in RFC 3986, an Internet Standard published in January 2005 and written by Berners-Lee, Roy Fielding and Larry Masinter. It uses the broader term URI (Uniform Resource Identifier), which covers URLs.⁴

The name

Berners-Lee chose the name to stress that the system has no center and that anything can link to anything, which in mathematics forms a graph, or web. He considered and rejected other names, including “Information Mesh,” and he writes the name as three separate words with no hyphen.⁵

The web and the internet are not the same thing

People often use “the web” and “the internet” as if they meant the same thing. They do not. In Berners-Lee’s own explanation, the internet is a network of networks: computers and cables that pass small packets of data from one machine to another. The web is an information space built on top of it, made of documents connected by hypertext links and served by programs that talk to each other over the internet.⁵

The internet is older by about two decades. The first node of ARPANET, the research network that grew into the internet, went live at UCLA in September 1969. Vint Cerf and Bob Kahn published a refined version of their TCP design in 1974, and ARPANET switched over to the TCP/IP protocols on January 1, 1983.⁶ The web arrived on that network almost eight years later.

A useful way to keep them apart: email, file transfer and video calls also travel over the internet, but they are not the web. The web is the part you open in a browser, where pages have URLs and link to other pages.

The difference shows up when you run a website. Your domain name and the address system that points it at a server are internet infrastructure. The pages, links and HTTP requests are the web layer that sits on top.

The 1989 proposal and the 1990 revision

The web started as an answer to a practical problem at CERN. In his 1989 proposal, Berners-Lee wrote that the laboratory kept losing track of its own knowledge: staff typically stayed about two years, projects changed constantly, and information about how systems worked was hard to find or lost when people left.⁷

The proposal, titled “Information Management: A Proposal” and dated March 1989, argued that the usual ways of organizing information did not fit. A tree (a strict hierarchy of folders) could not show how things really related to each other, and keyword searches failed when two people used different words for the same thing.⁷

Instead it proposed a web of notes joined by links. It called for a client/server design, where the program that stores information is separate from the program that displays it, so that many kinds of computers could take part. It also called for no central control and for gateway servers that would present existing data, such as Usenet news and CERN’s phone directory, as hypertext. For a first phase it estimated two people working for 6 to 12 months.⁷

Ideas the web built on

Hypertext was not new in 1989. W3C’s history of the web traces the idea back to 1945, when Vannevar Bush described the Memex, an imagined desk that would store a person’s documents and the trails between them, in The Atlantic Monthly.⁸

In the 1960s, Douglas Engelbart built a working prototype system called NLS and invented the computer mouse, and Ted Nelson coined the word “hypertext” in 1965. In 1967, Andy van Dam and colleagues built the Hypertext Editing System and later FRESS.⁸

Berners-Lee had also tried the idea himself. During a consulting stint at CERN from June to December 1980 he wrote a program called ENQUIRE. It was never published, but his biography calls it the conceptual basis for the web.⁸⁹ His proposal went further by calling for links between nodes in different databases, on different kinds of computers across the network.⁷

From proposal to project

Berners-Lee’s manager, Mike Sendall, wrote “Vague, but exciting” on the cover of the proposal and let him keep working on it.¹⁰ The document was sent around again in May 1990 with only the date changed.⁷

Sendall also backed the purchase of a NeXT workstation and suggested Berners-Lee use it to try out the global hypertext idea.¹¹ Sources give slightly different dates for that approval. The ACM’s Turing Award biography says Sendall authorized the work in May 1990, framed as a test of the NeXT machine, while W3C’s history places Sendall’s approval of the NeXT purchase in September 1990.¹⁸

On November 12, 1990, Berners-Lee and Robert Cailliau, a Belgian systems engineer at CERN, put forward a more formal plan: “WorldWideWeb: Proposal for a HyperText Project.”³² It named the system WorldWideWeb, or W3, and aimed to give the CERN and high-energy physics community a single way to reach reports, databases, computer documentation and online help through existing CERN servers.³

The 1990 plan set modest goals. It asked for a simple common protocol for requesting readable information over a network, automatic format negotiation between browser and server, and text-based browsing that would work on most of CERN’s screens, with graphics left for later. It preferred free and public-domain software, and it left out multimedia research and complex network security. The first phase was planned to take about three months. The summary asked for four software engineers and a programmer, including people to write browsers for the NeXT, the Macintosh, X Window terminals and CERN’s IBM mainframe.³

Building the first server, browser and website

Berners-Lee began writing the software in the autumn of 1990. His press FAQ says he started coding in September 1990, while his W3C biography dates the start of the first server and client to October 1990.⁵⁹ He wrote the first web server, called httpd, and the first client, a browser and editor called WorldWideWeb.⁹

He built WorldWideWeb in the Objective-C language on a NeXT computer. He credits the NeXTStep system’s built-in tools, such as its interface builder and ready-made text components, with letting him finish in a couple of months what would otherwise have taken about a year.¹²

The first browser was also an editor. It was a WYSIWYG (“what you see is what you get”) editor, so a page looked on screen much as it would when browsed. It could read pages over HTTP and also read Usenet news, FTP sites and local files. It could only save edits to local files, because the HTTP command for uploading a page (PUT) did not exist yet. To avoid confusing the program with the information space, he later renamed the program Nexus.¹²

Some features that later browsers made standard were already there. A Navigate menu offered back, next and previous, a Link menu let the user mark a page and link to it, and a style menu supported style sheets. He notes that a screenshot he made in 1993 differs from the first edition, which was grayscale, because NeXT screens were, and showed inline images in separate windows.¹²

The first web page served over HTTP dates from the end of 1990 and was hosted on a CERN machine. Before that, he says, the first pages lived as local files on his own machine and were opened with “file:” addresses. His NeXT was both the first browser and the first server.⁵ W3C’s history puts the first server and first page online in November 1990. Berners-Lee dated the first edition of the browser December 25, 1990, though he says he was not working that day.⁸¹²

The NeXT that ran the first server carried a handwritten label in red ink telling people not to switch it off.² The machine is still at CERN.¹³

The first website was about the web project itself. CERN’s restored copy describes WorldWideWeb as a “wide-area hypermedia information retrieval initiative” and links to sections on what was available online, help, software, technical details, the people involved and how to get the code.¹⁴ CERN started a project in 2013 to restore that site, which had stopped being available at its original address.¹⁵ The restored version is online today at info.cern.ch.¹⁶

Going public, 1991 to 1993

A browser that ran only on an expensive NeXT workstation could not reach many people. Nicola Pellow wrote a line-mode browser, a text-only program that worked on simple terminals, and it was released in 1991 with the server software and a code library for developers.²

W3C’s timeline lists a limited release of the line-mode browser in March 1991 and a general release of the web software on CERN’s central computers on May 17, 1991.⁸ CERN describes the line-mode browser, launched more widely in 1992, as the first browser that ordinary users could easily get.¹⁷

On August 6, 1991, Berners-Lee described the project in a post to the Usenet newsgroup alt.hypertext. He offered the NeXT browser-editor, the line-mode browser and the server code by anonymous FTP from the machine info.cern.ch. He explained that the server could deliver ordinary files and could also build hypertext pages on the fly from databases or live data, and noted that the code was copyright CERN but that free distribution and use were generally not a problem.¹⁸ His summary of the project was one sentence long:

“The WorldWideWeb (WWW) project aims to allow links to be made to any information anywhere.”¹⁸

The web soon spread beyond Europe. In December 1991, Paul Kunz and Louise Addis set up the first web server in the United States, at the Stanford Linear Accelerator Center (SLAC). W3C dates it to December 12, 1991.²⁸

Other groups began writing browsers. By January 1993, browsers called Midas, Erwise and Viola were available for Unix computers running the X Window System.⁸ In February 1993, the National Center for Supercomputing Applications (NCSA) at the University of Illinois released the first test version of Mosaic for X, and later versions followed for PCs and Macintosh computers.⁸²

Public money followed. At the end of 1993 the European Commission approved its first web project, called WISE, with CERN as a partner.²

By late 1993 there were more than 500 known web servers, and the web carried about 1% of all internet traffic.²

Free for everyone: April 30, 1993

The step that let anyone build on the web without permission came on April 30, 1993. On that day CERN released the World Wide Web software royalty-free, so that anyone could use, copy, change and redistribute it.¹⁹²

The release took the form of a short internal document addressed to anyone it might concern. It was signed by Walter Hoogland, CERN’s Director of Research, and Helmut Weber, its Director of Administration.¹⁹ Its key sentence read:

“CERN relinquishes all intellectual property rights to this code, both source and binary form.”¹⁹

CERN’s own pages describe this as putting the web software into the public domain. The next version, in 1994, was released under an open source license instead, which let CERN keep the copyright while still allowing anyone to use and modify the code.¹⁷¹⁹

Getting there took effort. Berners-Lee says he pressed CERN’s management for about 18 months to release the technology without royalties. He has written that if the web had been proprietary and under his total control, it probably would not have taken off, and that an open system was needed for it to be universal.⁵

That choice drew notice later. When Berners-Lee received the first Millennium Technology Prize in 2004, MIT reported that the prize committee praised his decision not to commercialize or patent the technology.²⁰

W3C and the standards that hold the web together

As the web grew, many companies were writing browsers and servers, and they needed shared rules. The World Wide Web Consortium (W3C) was founded on October 1, 1994, at the MIT Laboratory for Computer Science, in collaboration with CERN and with support from the US research agency DARPA and the European Commission.⁸²¹ Berners-Lee left CERN for MIT that year to lead it.²

The first W3C members’ meeting took place at MIT on December 14, 1994. Two days later, CERN’s governing council approved the Large Hadron Collider, and CERN stopped its own web development and passed its web work to the French research institute INRIA.⁸

W3C’s model was to fund paid staff through corporate memberships.¹ For almost three decades, it was run by host institutions. MIT was joined by INRIA in April 1995, Keio University in Japan in 1996, ERCIM (which took over the European role from INRIA) in 2003, and Beihang University in China in 2013.²¹ At the start of 2023, W3C became a public-interest non-profit organization incorporated in the United States, with more than 400 member organizations at that time.²² Its About page now lists more than 330 member organizations and a staff of about 50.²³

W3C is not the only body behind web standards. The Internet Engineering Task Force (IETF) publishes many of the core protocols as numbered documents called RFCs, and since 2019 the WHATWG has maintained HTML itself.²⁴²⁵

Berners-Lee wrote or co-wrote the early ones. RFC 1738 lists him with Larry Masinter and Mark McCahill, RFC 1866 with Dan Connolly, and RFC 1945 with Roy Fielding and Henrik Frystyk.²⁶²⁷²⁸

For HTML, W3C and the WHATWG spent years publishing competing versions. On May 28, 2019, they agreed to work on a single version: the WHATWG maintains the HTML and DOM Living Standards, and W3C takes periodic snapshots of them through its own review process to publish as W3C Recommendations.²⁵

Behind the standards sits a set of design ideas Berners-Lee wrote down in 1998. They include simplicity, modular design, tolerance (accepting imperfect input while producing correct output), decentralization (no single point whose failure breaks everything) and the principle of least power, which favors the simplest language that can do a job.³⁰

Growth by the numbers

The web grew fast once browsers and free server software were available. CERN calls 1994 the “Year of the Web.” Its first international web conference, held at CERN in May 1994, drew 380 people, and a second conference in the United States that October drew 1,300.²

Counts of “sites” depend on what is counted. Netcraft’s survey counts hostnames, so one company can have many. Its July 2026 survey found 1,494,915,628 sites, but only 305,348,459 distinct domains and 14,772,048 web-facing computers.³²

The number of people online has grown as well. The International Telecommunication Union (ITU), the UN agency for digital technology, estimates that about 6 billion people used the internet in 2025, or 74% of the world’s population, up from 60% in 2020. About 2.2 billion people were still offline.³³ Two years earlier, CERN had cited ITU figures of more than 5 billion regular users, about two-thirds of the world.¹⁹

The standards have scaled with that growth. When W3C relaunched as a non-profit in January 2023, it said its community had produced nearly five hundred open standards, which it credited with powering about 2 billion websites.²²

The web today

The web of 2026 runs on the same three ideas as the web of 1990: addresses, a request protocol and a markup format. What has changed is the layers added on top.

Most websites now use encryption. HTTPS, the encrypted form of HTTP, is the default protocol on 90.3% of websites, according to W3Techs in October 2026.³⁴ Pages are styled with CSS and made interactive with JavaScript, but they are still delivered as HTML over HTTP.

Most sites now run on a content management system (CMS). As of October 22, 2026, W3Techs finds that WordPress runs 39.6% of all websites it surveys, Shopify 5.3%, Wix 4.2% and Squarespace 2.4%, while 32.1% use none of the content management systems it tracks.³⁵

The server software has changed too. In Netcraft’s July 2026 survey, nginx served 20.37% of all sites, Cloudflare 16.67% and Apache 11.50%.³² Every one of these programs still does the job CERN httpd did in 1990: it answers HTTP requests for URLs.

The web’s inventor has spent recent years arguing that it needs repair. In a 2019 letter for the web’s 30th anniversary, Berners-Lee named three sources of harm: deliberate malicious intent, business incentives that reward clickbait and misinformation, and unintended effects of well-meant design, such as outraged and polarized discussion.³⁶

Timeline

1. July 2026Netcraft's survey counts 1,494,915,628 sites and 305,348,459 domains³²
2. January 31, 2023W3C relaunches as a public-interest non-profit organization²²
3. June 2022HTTP Semantics (RFC 9110) published as Internet Standard STD 97²⁴
4. May 28, 2019W3C and the WHATWG agree to maintain a single HTML standard²⁵
5. April 30, 2013CERN starts a project to restore the first website¹⁵
6. January 2005RFC 3986 sets the generic syntax for URIs (STD 66)⁴
7. May 1996HTTP/1.0 documented as RFC 1945²⁸
8. November 1995HTML 2.0 published as RFC 1866²⁷
9. August 1995Netcraft's first web server survey finds 10,835 sites³¹
10. December 1994URLs specified in RFC 1738²⁶
11. October 1, 1994World Wide Web Consortium (W3C) founded at MIT⁸
12. May 1994First international World Wide Web conference held at CERN, with 380 attendees²
13. April 30, 1993CERN releases the web software royalty-free¹⁹
14. February 1993NCSA releases the first test version of Mosaic for X⁸
15. December 12, 1991First web server outside Europe goes online at SLAC⁸
16. August 6, 1991Berners-Lee announces the project on the alt.hypertext newsgroup¹⁸
17. December 1990First web server and browser running at CERN on a NeXT computer²
18. November 12, 1990Berners-Lee and Robert Cailliau submit the WorldWideWeb hypertext project proposal³
19. March 1989Berners-Lee writes "Information Management: A Proposal" at CERN⁷

What it means for your website

See also

• Internet (coming soon)
• Tim Berners-Lee
• HTTP
• HTML
• The first website (1991)
• How does a website work
• Web hosting

References

1. Tim Berners-Lee, A.M. Turing Award laureate (biography by Thomas Haigh). Association for Computing Machinery. accessed October 22, 2026.
2. A short history of the Web. CERN. accessed October 22, 2026.
3. WorldWideWeb: Proposal for a HyperText Project. Tim Berners-Lee and Robert Cailliau, CERN (copy hosted by W3C). 1990-11-12. accessed October 22, 2026.
4. RFC 3986: Uniform Resource Identifier (URI): Generic Syntax. RFC Editor / IETF. 2005-01. accessed October 22, 2026.
5. Frequently asked questions by the Press. Tim Berners-Lee, W3C. accessed October 22, 2026.
6. A Brief History of the Internet (Leiner, Cerf, Clark, Kahn, Kleinrock, Lynch, Postel, Roberts, Wolff). Internet Society. 1997. accessed October 22, 2026.
7. Information Management: A Proposal. Tim Berners-Lee, CERN (HTML copy hosted by W3C). 1989-03. accessed October 22, 2026.
8. A Little History of the World Wide Web (created by Robert Cailliau, maintained by Dan Connolly). W3C. accessed October 22, 2026.
9. Sir Timothy Berners-Lee OM, KBE, FRS, FREng, FRSA (longer biography). W3C. accessed October 22, 2026.
10. CERN celebrates 20th anniversary of World Wide Web (press release). CERN. 2009-03-13. accessed October 22, 2026.
11. D.M. Sendall. Tim Berners-Lee, W3C. 1999-07. accessed October 22, 2026.
12. The WorldWideWeb browser. Tim Berners-Lee, W3C. accessed October 22, 2026.
13. NeXT server (Computer Centre visit point display). CERN IT Department. accessed October 22, 2026.
14. The World Wide Web project (restored copy of the first website). CERN. accessed October 22, 2026.
15. CERN celebrates 20 years of a free, open web (press release). CERN. 2013-04-30. accessed October 22, 2026.
16. info.cern.ch: home of the first website. CERN. accessed October 22, 2026.
17. The birth of the Web. CERN. accessed October 22, 2026.
18. Re: Qualifiers on Hypertext links… (alt.hypertext post, archived). Tim Berners-Lee, via W3C. 1991-08-06. accessed October 22, 2026.
19. 30 years of a free and open Web (Naomi Dinmore). CERN. 2023-04-30. accessed October 22, 2026.
20. Berners-Lee wins $1.2 M award. MIT News. 2004-04-28. accessed October 22, 2026.
21. W3C history. World Wide Web Consortium. accessed October 22, 2026.
22. W3C re-launched as a public-interest non-profit organization (press release). World Wide Web Consortium. 2023-01-31. accessed October 22, 2026.
23. About us. World Wide Web Consortium. accessed October 22, 2026.
24. RFC 9110: HTTP Semantics. RFC Editor / IETF. 2022-06. accessed October 22, 2026.
25. W3C and WHATWG to work together to advance the open Web platform (Jeff Jaffe). W3C blog. 2019-05-28. accessed October 22, 2026.
26. RFC 1738: Uniform Resource Locators (URL). RFC Editor / IETF. 1994-12. accessed October 22, 2026.
27. RFC 1866: Hypertext Markup Language – 2.0. RFC Editor / IETF. 1995-11. accessed October 22, 2026.
28. RFC 1945: Hypertext Transfer Protocol — HTTP/1.0. RFC Editor / IETF. 1996-05. accessed October 22, 2026.
29. RFC 9114: HTTP/3. RFC Editor / IETF. 2022-06-06. accessed October 22, 2026.
30. Principles of Design (Design Issues). Tim Berners-Lee, W3C. 1998. accessed October 22, 2026.
31. The first year: August 1995 to August 1996. Netcraft. 1996. accessed October 22, 2026.
32. July 2026 Web Server Survey. Netcraft. 2026-07-27. accessed October 22, 2026.
33. Statistics: ICT Facts and Figures 2025. International Telecommunication Union. 2025-11. accessed October 22, 2026.
34. Usage statistics of Default protocol https for websites, October 2026. W3Techs. 2026-10. accessed October 22, 2026.
35. Usage statistics of content management systems, October 2026. W3Techs. 2026-10-22. accessed October 22, 2026.
36. 30 years on, what's next #ForTheWeb?. Tim Berners-Lee, World Wide Web Foundation. 2019-03-12. accessed October 22, 2026.
37. Cool URIs don't change. Tim Berners-Lee, W3C. 1998. accessed October 22, 2026.

Web Building Encyclopedia. First published October 22, 2026. Last reviewed October 22, 2026. Every fact on this page is linked to a source we opened. Spotted an error? Write to [email protected].