Internet history · Part 4 of 10

Tim Berners-Lee and the Web, 1989 to 1993, the four most important years

Jul 7, 20257 min read#networking#history#internet-history

Tim Berners-Lee and the Web, 1989 to 1993, the four most important years

Field note. The web was bolted on top of an already-running internet. Knowing that order changes how you think about why 'web stuff' and 'game stuff' use different tools and protocols.

In 1989, the internet was a tool. Useful for researchers, accessible to anyone with a Unix workstation and a terminal, but technical. Email, FTP, Usenet, Telnet. Power-user oriented.

In 1993, the World Wide Web went public. Within five years, the internet was the dominant medium of information exchange on the planet.

The transformation came from one researcher at CERN, Tim Berners-Lee, and a small team. This article is the story.

CERN and the problem

CERN (the European Organization for Nuclear Research) in Geneva is the world's largest particle physics laboratory. It hosts thousands of researchers from around the world, working on long-running experiments.

The problem: documentation. Experiments lasted years. Researchers came and went. Knowledge was scattered across paper documents, computer files in various formats, on various machines, in various countries. Finding "the latest version of the analysis software's documentation" was genuinely hard.

Tim Berners-Lee, a young British software engineer, had been at CERN since 1980 (with a break in the middle). He'd already built a small experimental system called ENQUIRE in 1980, which let you navigate between documents via "links."

In March 1989, he wrote a proposal: "Information Management: A Proposal." It described a system of interlinked documents accessible from any computer on the network, retrievable by following links from one document to another.

His manager, Mike Sendall, scribbled on the cover: "Vague, but exciting."

Berners-Lee was given the time to try.

What he built

By 1990, Berners-Lee had built:

  1. HTTP (HyperText Transfer Protocol): the application protocol for fetching documents over the internet.
  2. HTML (HyperText Markup Language): a simple language for marking up documents with formatting and, critically, links.
  3. A web browser: software that could fetch documents via HTTP, render HTML, and follow links.
  4. A web server: software that could serve documents via HTTP.

The first web server ran on a NeXT computer at CERN. The address was info.cern.ch. The first web page is preserved and still accessible at https://info.cern.ch/hypertext/WWW/TheProject.html.

Berners-Lee also wrote a name for the system: World Wide Web. Initially he'd called it "TIM" and "Mesh," but WWW stuck.

By December 1990, the first complete HTTP / HTML / browser / server system was working between two computers at CERN. The Web existed.

What was novel

Several systems before the Web had similar ideas:

  • HyperCard (Apple, 1987): linked cards, local only.
  • Gopher (University of Minnesota, 1991): hierarchical document retrieval over the internet, popular for a few years.
  • WAIS (Wide Area Information Servers, 1991): full-text search across distributed databases.
  • Various hypertext systems in academic research.

What made the Web specific:

Internet-native. Worked over the existing internet (TCP/IP). No special infrastructure required.

Universal addressing. Every document had a URL (initially called UDIs, Universal Document Identifiers). Anyone could link to anyone's document, no permission needed.

Open standards. HTTP and HTML were published, free, anyone could implement. No royalty, no licensing.

Simple. HTTP was very simple compared to alternatives. HTML was easy to write. A small team could implement a working web browser. The protocol could be debugged with telnet.

Document-oriented. Focused on linked documents (text, eventually images), not on more abstract data exchange.

These choices proved decisive in the years that followed.

The early growth (1991-1993)

CERN announced the Web publicly in August 1991. Berners-Lee posted to the alt.hypertext newsgroup. Initial reception was niche but enthusiastic.

By 1992, a handful of websites existed. The Stanford Linear Accelerator (SLAC) put up the first significant U.S. website. CERN released the source code for the browser and server. Anyone could deploy.

In November 1992, there were 26 known web servers. By late 1993, there were several hundred. The community was technical and small but growing fast.

Mosaic

The pivotal event: Marc Andreessen and Eric Bina at the National Center for Supercomputing Applications (NCSA, at the University of Illinois) developed a browser called Mosaic in early 1993.

What Mosaic did differently:

  • Graphical. Showed text and images in the same window (most earlier browsers showed images in separate windows).
  • Cross-platform. Versions for Unix, Windows, Mac.
  • Easy to install. A regular user could download and run it.
  • Polished UI. Felt like a real consumer product.

Mosaic for Windows shipped in mid-1993. For most non-Unix users, Mosaic was their first web browser.

The effect was electric. People who had no interest in command-line tools suddenly had a graphical, point-and-click way to browse documents on the internet. Web traffic spiked. New sites appeared by the day.

NCSA released Mosaic for free. Anyone could download it. The internet community started talking about something new.

CERN releases the Web

In April 1993, CERN made a critical decision: it released the Web technology to the public domain. HTTP, HTML, the browsers, and the server software were free for anyone to use, modify, or commercialize.

This was not obvious. Other organizations might have patented their technology, licensed it, sought rent. CERN explicitly chose to give it away.

The decision is sometimes credited as the single most economically consequential act of the late 20th century. The Web's openness enabled the commercial internet's explosion. Anyone could build websites. Anyone could implement browsers. No royalties, no permissions.

Tim Berners-Lee has consistently said this was his preference and CERN's choice was crucial.

Andreessen leaves NCSA

In early 1994, Marc Andreessen graduated and went to Silicon Valley. He met Jim Clark, a former Stanford professor and founder of Silicon Graphics. Together they founded a company called Mosaic Communications (later Netscape Communications) to build a commercial web browser.

The first Netscape Navigator beta shipped in October 1994. It was a much-improved Mosaic, designed by the same people, this time as a commercial product. By 1995, Netscape Navigator was the dominant browser.

Netscape's IPO in August 1995 is sometimes counted as the formal start of the dot-com era. The company had no profits but the IPO was hugely successful, signaling that "the internet" was investment-worthy.

The browser wars

Microsoft, which had ignored the internet through the early 1990s, woke up. In late 1995, Bill Gates issued his famous "Internet Tidal Wave" memo, redirecting much of Microsoft toward web technology.

Microsoft built Internet Explorer (initially licensed from Spyglass, a Mosaic descendant) and bundled it with Windows 95. The first browser war was between Netscape Navigator and Internet Explorer.

Microsoft's leverage (Windows bundling, free distribution, deep pockets) was decisive. By the late 1990s, IE had overtaken Netscape. Netscape was acquired by AOL in 1999 and faded.

Internet Explorer dominated through the 2000s. Then Firefox (a Netscape descendant via Mozilla) ate into its share. Then Chrome (from Google, launched in 2008) eventually overtook everyone.

The browser wars continue today, with Chrome dominant, Safari strong on Apple devices, Firefox and Edge fighting for minority share.

The Web's evolution

What started as "documents with links" has evolved into a full application platform:

  • HTML 1.0 (1993): basic markup, links, images.
  • HTML 4 (1997): tables, forms, scripts.
  • CSS (1996+): styling separate from content.
  • JavaScript (1995): scripting in the browser. Originally for small interactions; now runs entire applications.
  • AJAX (mid-2000s): pattern for dynamic page updates without reload. Gmail and Google Maps demonstrated.
  • HTML5 (2014): video, audio, canvas, offline support.
  • WebRTC (2010s): real-time peer-to-peer in browsers.
  • Web Components, Service Workers, WebAssembly: increasing capability.

The trajectory: from documents to applications to platforms. Modern websites are application runtimes that happen to also display documents.

Berners-Lee's continued role

Tim Berners-Lee remains active. He:

  • Founded the World Wide Web Consortium (W3C) in 1994 to develop web standards. Still its director.
  • Was knighted in 2004.
  • Won the ACM Turing Award in 2016.
  • Founded the Web Foundation, advocating for an open and equitable web.
  • More recently, has launched Solid, a proposal to give users more control over their data.

He's been a vocal advocate for the Web's original ideals: open, accessible, royalty-free, user-controlled. He has also expressed concern about its evolution: surveillance capitalism, platform dominance, misinformation. His current work emphasizes restoring user agency.

What the Web didn't predict

A few things the early Web didn't anticipate:

Commerce. Berners-Lee's design was for research documents, not online shopping. E-commerce was added on top.

Surveillance. Cookies (1994) were originally for simple state, not the cross-site tracking they became.

Search. Early Web had no good search. Yahoo (1994) was a curated directory; AltaVista (1995) was the first major search engine. Google (1998) made search effective enough to be the entry point.

Social media. The Web's link model gave way to social networks' walled gardens by the mid-2000s.

Mobile. The Web was designed for desktops with mice. Mobile changed everything by 2010.

These weren't failures. They were extensions. But the original simplicity gave way to enormous complexity.

Conclusion

Tim Berners-Lee's 1989 proposal solved a specific problem at CERN. The solution turned out to be applicable everywhere. By 1993, the Web was on a trajectory that would transform the internet from a researchers' tool to a global medium.

The key choices that made this possible:

  • Universal naming via URLs.
  • Open standards (HTTP, HTML).
  • Public-domain release by CERN.
  • Free, cross-platform browsers (Mosaic).

Without any one of these, the Web might have been a niche technology like Gopher. With all of them, it became the platform on which the modern internet runs.

Three or four years of work by a small team in Geneva, plus the right combination of decisions, plus a free graphical browser. That's how the Web happened.

Coming up

Next: the browser wars. The 30-year story of which company's software people use to access the Web, and why it matters.


Hosting your game server with AndroHost means we handle most of what's in this post for you automatically: tier sizing, SRV records, off-site backups, DDoS protection.

Browse plans·More posts·Discord