The History of the World Wide Web

By Victor Hoang · Updated Oct 24, 2026

In March 1989, the European particle physics laboratory CERN was a constantly changing web of projects. Staff turnover was high. Tim Berners-Lee's proposal says knowledge about past projects was often lost or took detective work to recover. A change in one corner could break something in another. Paper records went stale fast.

The World Wide Web began as that proposal, which Berners-Lee wrote at CERN, near Geneva, in March 1989. The first server went live in 1990. On 30 April 1993 CERN released the code to everyone with no fees. The throughline: the Web spread because linking was simple and free to use. Nearly every later fight over it has been about who gets to put walls around the links.

Several thousand people at CERN were working toward shared goals. The Web was his answer to a filing problem inside one lab. That it became the way most people meet the Internet was not the plan.

A note on terms. The Web is not the Internet. The Internet is the network, and the Web is a system of linked documents that runs on it. That split is told in Internet vs World Wide Web: Two Different Origins, so this article stays on the Web's own story.

Timeline at a Glance

The Dream Years (1895-1968)

Linked knowledge is imagined long before any machine can serve it

Picture a clerk in Brussels in the 1920s, hunting through drawers of index cards to answer a question that arrived by post. That was Paul Otlet's world. Britannica records that in 1895 Otlet and Henri La Fontaine established the International Institute of Bibliography, a fee-based service where researchers sent queries by mail or telegraph and staff replied with copies of catalog entries. By 1927 the catalog held 13 million cards, and it peaked at 15.6 million in 1934. In his 1930s writings, Otlet pictured a world where, in his words, "anyone from his armchair would be able to contemplate the whole of creation." Britannica says his vision presaged the Web by more than 50 years. But the Belgian government shuttered the project in 1934, and part of the collection was destroyed after Germany invaded in 1940. Otlet's network was paper, clerks and the post office. It was a precursor, not the Web.

The next dream came from an American. In July 1945, Vannevar Bush wrote in The Atlantic that there was "a growing mountain of research" and that specialists could not keep up with their own field. He proposed the memex, a desk-sized device for storing a person's books, records and communications, and called it "an enlarged intimate supplement to his memory." His key point was how people think: the mind, he wrote, "operates by association," so a machine should let you follow a trail from one item to the next. Bush described a device for one person's own records, not a network of shared ones.

Then the idea got a name and a demonstration. The W3C's history records that Ted Nelson coined the word "hypertext" in 1965, and that Doug Engelbart prototyped a hypertext system called NLS in the 1960s. On 9 December 1968, at the Fall Joint Computer Conference in San Francisco, Engelbart drove NLS from a console on stage while the computer itself sat at SRI about 30 miles away and colleagues joined him by video link. It showed that linked documents could work. It still ran on one lab's machine.

The new problem at the end of this era was reach. Every dream so far lived inside one institution, one desk or one computer. The question for the 1980s was whether a link could cross a boundary.

What this means for teams today: the ideas behind your tools are usually older than the tools. If a good idea keeps reappearing without taking hold, look for what is missing around it, here a shared network and a free, simple format.

The CERN Proposal (1980-1993)

A lab's filing problem becomes a free public tool

The turning point is a memo with a one-line verdict. Berners-Lee had already tried the idea once. From June to December 1980, while consulting at CERN, he wrote ENQUIRE, a notebook program that linked arbitrary nodes. In March 1989 he proposed something larger: a distributed system of linked information for the whole lab. The proposal asks for a "universal linked information system" and, at that point, had no name for it beyond "Mesh". His boss, Mike Sendall, wrote "Vague, but exciting" on it.

The green light came later. Per the W3C's history, Sendall approved the purchase of a NeXT computer in September 1990 and told Berners-Lee to go ahead and write a global hypertext system. In October 1990 he named the browser and editor "WorldWideWeb" and the project "World Wide Web". In November 1990 the first server, nxoc01.cern.ch (later info.cern.ch), went live with the first page. Berners-Lee himself dated it a month later in a 2010 essay, saying the Web went live in December 1990, which is the kind of variant to expect when the first page was never announced. The project page still online describes the project as an initiative "to give universal access to a large universe of documents."

Releases came in steps:

• March 1991: a line-mode browser goes to a limited audience, and on 17 May the Web is generally released on CERN's central machines.
• August 1991: the files are posted by FTP and announced on Usenet newsgroups, according to the W3C. Britannica instead dates the text browser's general release to January 1992.
• 12 December 1991: Paul Kunz installs the first Web server outside Europe, at SLAC.
• 3 November 1992: a project-page snapshot lists 26 reasonably reliable servers.

Then came the decision that shaped everything after. On 30 April 1993, CERN's Director of Research Walter Hoogland and Director of Administration Helmut Weber signed a document, and CERN's account quotes it: "CERN relinquishes all intellectual property rights to this code." Anyone could use, copy and change the software. A 1994 release moved to an open source licence, so CERN kept copyright while anyone could modify it. Remember the dates: proposal in March 1989, first server in late 1990, release at CERN in May 1991, release to the world in April 1993. Which one is "the birth" depends on whether you mean the idea, the software or the public gift.

The problem solved was a lab's lost knowledge. The problem created was an empty Web: a few dozen servers and a text-only browser that few people outside physics wanted to install.

What this means for teams today: a small fix for your own team can turn out to be a platform. And the fee you decide not to charge can matter more than the feature you ship.

The Mosaic Rush (1993-2000)

A graphical browser, a standards body and a stampede

The turning point is a browser people could click. The W3C's timeline shows what happened in 1993 alone: about 50 known servers in January, the first alpha of Marc Andreessen's Mosaic for X at NCSA in February, and the University of Minnesota's announcement of fees for Gopher, a rival system, which pushed many users toward the Web. Web traffic measured 0.1% of NSF backbone traffic in March and 1% by September, when NCSA released Mosaic for X, Windows and Macintosh. By October there were over 200 known servers, and by June 1994 over 1,500. Britannica dates Mosaic's release to September 1993, which matches the all-platform versions rather than the February alpha. The detailed browser story, from Mosaic through Netscape and the browser wars, is told in The History of Web Browsers.

In May 1994, the first International WWW Conference at CERN was so oversubscribed that it became known as the "Woodstock of the Web," per the W3C. Five months later the World Wide Web Consortium opened. The W3C's own history says the decision came at the urging of many companies investing in the Web, and that Berners-Lee founded it in October 1994 at MIT, with CERN, DARPA and the European Commission supporting. He required from the start that it have a global footprint. How the consortium and its rivals shaped markup is covered in The History of HTML and Web Standards.

After that the Web became a business, a stretch told in The Dot-Com Bubble: What Happened and Why. Email had been running on the Internet for years before any of this, as the history of email shows.

What this era solved was access and credibility: a free format, a friendly browser and a neutral standards body made it safe to build on. What it created was a boom, and the question of who would own the pages, the traffic and the readers. Finding anything in that flood became its own industry, traced in The History of Search Engines.

What this means for teams today: growth followed an easy interface on top of a free standard, not a better backend. When you pick tools, ask what a first-time user has to learn before they get value.

The Web 2.0 Years (2000-2007)

Readers become writers, and software moves into the browser

The turning point is a relabeling. Britannica's entry on Web 2.0, written by William L. Hosch, defines it as a term for the post-dotcom era of the Web, with an emphasis on social networking, user-generated content and cloud computing. The term gained popularity after a series of Web conferences organized by Tim O'Reilly in 2004. It describes a shift from a mostly read-only Web to one where, in Britannica's words, everyone in the network can produce, distribute and consume content at once.

Look at what the examples have in common. Social networks, video sites, blogging tools and customer reviews on shopping sites all depend on ordinary visitors adding the content. Britannica names Facebook, Twitter, Instagram, YouTube, Blogger and WordPress among them. The pages were no longer written by webmasters and read by everybody else. They were assembled from what millions of people typed in.

That raised a new problem, one Berners-Lee would spell out in 2010. These sites held your data, your friends and your photos in databases that only their own pages could reach. The Web's links had a place for every document, but people's data did not always have a link.

For businesses, the Web 2.0 turn changed jobs more than technology. Customers could review, share and criticize in public, and software could be delivered through a login instead of an installer, a shift traced in the history of cloud computing.

What this means for teams today: when users create the content, they also create the lock-in. Decide early whether people can take their data out, and whether their pages keep a stable address.

The Mobile and App Years (2007-2016)

The phone gets the Web, and the app store competes with it

The turning point is a keynote. On 9 January 2007, Apple announced the iPhone, combining a phone, an iPod and what it called "a breakthrough Internet communications device" with desktop-class email, web browsing, searching and maps. It went on sale in the US that June. For the first time many people carried a real browser everywhere.

Eighteen months later the phone grew a second front door. On 10 July 2008, Apple said more than 500 native applications would be available in its App Store when the iPhone 3G went on sale the next day, and described the phone as a new software platform for developers. An app did not need a link, a URL or a browser.

The argument that followed ran in public. In August 2010, Chris Anderson and Michael Wolff wrote in Wired that the Web had been eclipsed by apps and other services and was in decline, under the headline "The Web Is Dead. Long Live the Internet." Months later, Berners-Lee published his own answer in Scientific American. He argued the Web was "certainly not dead," but under threat: large social sites walling off what their users post, where, in his words, "Each site is a silo", and magazine apps that put material where you cannot bookmark it, email a link to it or tweet it. His defense rested on three design principles: universal links, decentralization (no central approval to add a page) and open, royalty-free standards. He also noted that the Web had run over the Internet without any change to the Internet itself, even as connection speeds rose from 300 bits per second to 300 Mbps.

The count settled one half of the debate. Britannica says that by 2016 smartphones accounted for more than half of Web browsing. The phone had not killed the Web, it had become the Web's main window. But the app stores and social silos stayed.

What this era solved was reach: the Web went with everyone, all day. What it created was a second, closed way to publish, with a gatekeeper that took a seat at the door.

What this means for teams today: build for the open web first, and treat each app store as a distribution channel you rent. A page that has an address can be linked, shared and found. An app screen often cannot.

The Open Web Question (2010s-present)

Most of the world is online, and the argument is about access, privacy and control

The turning point is the scale. The ITU's latest estimates, labeled end-2025, say that almost three-quarters of the world's population is now online, and that 2.2 billion people remain offline, most of them in low- and middle-income countries. These are Internet figures, not Web figures, but the Web is how most of those users reach the network.

The debate Berners-Lee raised in 2010 returned as a policy project. The Contract for the Web, run through the Web Foundation, set out principles for governments, companies and citizens covering access, openness, privacy and data rights. It was drafted over ten months by more than 80 signatories, informed by a public consultation of more than 600 people. Its core group included the governments of France and Germany and companies such as Google and Microsoft.

Underneath, the foundation is the one CERN gave away. The W3C, which Berners-Lee founded in 1994 to coordinate web standards, still does that work, and the HTML specification is now a single living document, as covered in the standards article. The Web has changed shape many times and stayed one Web.

What this means for teams today: the open parts of the stack are what make everything else replaceable. When you evaluate a vendor, ask if your data and your pages have addresses you control. For the broader network story, see the history of the Internet.

Key People

Key Facts

Key Facts: Web History

• CERN's directors signed the document releasing the Web code on 30 April 1993, stating that CERN relinquishes all intellectual property rights to it.
• Known Web servers grew from about 50 in January 1993 to over 200 in October 1993 and over 1,500 by June 1994, per the W3C.
• Otlet's world card catalog reached 15.6 million cards in 1934, long before any network could link them.
• Berners-Lee wrote that the Web ran over the Internet without change as speeds rose from 300 bits per second to 300 Mbps.
• Smartphones passed half of Web browsing in 2016.

The Pattern Behind the History

We call the sequence the Gift and Wall Cycle. It is our synthesis of the history above, not a research finding.

1. The idea arrives early. Otlet, Bush, Nelson and Engelbart each described linked knowledge decades before a shared network could carry it.
2. A local problem forces a simple tool. CERN's lost project knowledge produced a plain system of linked pages, not a grand design.
3. The owner gives it away. The 1993 release meant nobody needed permission, and a friendly browser did the rest.
4. The commons attracts the platforms. Social sites and app stores grew on top of an open Web, then built walls around their own pieces of it.
5. The open core gets defended, again. Berners-Lee's 2010 essay and the Contract for the Web are the same move: restate the principles when the walls go up.

What It Teaches Teams Today

• Solve your own filing problem simply. The Web began as a fix for lost knowledge in one lab.
• Giving away the standard can be the growth strategy. The fee CERN waived in 1993 is part of why anyone could build on it.
• Interface beats infrastructure. Mosaic made the Web usable before the network got any faster.
• Own your addresses. Pages with URLs can be linked, cited and found, while data inside a closed app often cannot.
• Know which date you mean. Proposal, first server, CERN release and public release are four different milestones. For the full dated list, see the Internet timeline, and for who gets the credit, who invented the Internet.