The Browser Wars: The Fight for the Web's Front Door

Zusammenfassung

This article traces the history of the web browser — from the first graphical browser in 1993, through the brutal commercial war between Netscape and Microsoft that nearly ended in monopoly, to the open-source revival that produced Firefox and the eventual dominance of Google Chrome. It is a story about how control of the software through which people experience the web is equivalent to control of the web itself — and about the recurring tension between open standards and proprietary extensions that has never been fully resolved.

The First Window onto the Web

The World Wide Web existed before anyone could see it comfortably. Tim Berners-Lee’s original browser, WorldWideWeb (1990), ran only on NeXT workstations and displayed text with minimal formatting. The web’s early adopters — physicists, academics, programmers — navigated it through text-based interfaces that required knowing exactly what URL to type. The barrier to entry was high; the audience was tiny.

Marc Andreessen was a twenty-one-year-old undergraduate working at the National Center for Supercomputing Applications (NCSA) at the University of Illinois when he began thinking about this problem. In late 1992 and early 1993, working with programmer Eric Bina, he wrote NCSA Mosaic — a graphical browser that displayed images inline with text on the same page, supported point-and-click navigation, and ran on Unix, Windows, and Mac from the same codebase.

Mosaic was not the first graphical browser. It was the first one that ordinary people could install and use. Released in January 1993 and made freely available, it spread through universities and research institutions at a rate that surprised even its creators. Within a year, Mosaic had been downloaded over a million times. The web had a face.

Andreessen graduated, moved to California, and in April 1994 co-founded Mosaic Communications Corporation with Jim Clark (founder of Silicon Graphics). The name was quickly changed to Netscape Communications after NCSA objected. They hired most of the Mosaic team and began building a new browser from scratch: Netscape Navigator.

Navigator shipped in December 1994. It was faster than Mosaic, more reliable, and full of features its predecessor lacked — persistent cookies, SSL encryption for secure transactions, JavaScript (a scripting language commissioned from Brendan Eich in ten days to add interactivity to web pages). By mid-1995, Netscape Navigator held approximately 90% of the browser market. The company charged $39 for commercial licenses; students and educators could use it free.

On August 9, 1995, Netscape went public. The shares were priced at $28; the first trade came at $71, the stock reached as high as $74.75, and it closed at $58.25 — one of the largest first-day gains in Wall Street history for a company that had never turned a profit. Netscape was worth $2.9 billion by end of day. The dot-com era had a starting gun.

The Tidal Wave Memo

Bill Gates had been slow to recognize the internet’s commercial significance. Microsoft’s dominant products — Windows, Office, the forthcoming Windows 95 — were built around local applications and local files. The network was for file sharing and email; it was not, as Gates initially saw it, a platform.

He changed his mind with characteristic decisiveness. On May 26, 1995, Gates circulated an internal memo to Microsoft’s senior leadership titled “The Internet Tidal Wave.” Its opening paragraph:

“I have gone back and forth on the Internet and how important it is. I believe we are now entering a new phase where it has become clear that this is the most important development since the introduction of the PC.”

The memo identified Netscape as Microsoft’s most dangerous competitor and called for redirecting company resources toward the internet. Within months, Microsoft had licensed the Mosaic source code from Spyglass (a company that had commercialized NCSA’s technology) — under a deal pairing a quarterly base fee with a royalty on Internet Explorer revenue — and began work on Internet Explorer. Because Microsoft bundled IE for free and thus paid only the minimum quarterly fee, Spyglass collected almost nothing on the royalty; in 1997 it threatened a contract audit and Microsoft settled for $8 million.

Internet Explorer 1.0 shipped in August 1995, bundled at no charge with the Windows 95 Plus! Pack. IE 2.0 arrived in November. IE 3.0, released in 1996, was genuinely competitive with Navigator. IE 4.0 (1997) exceeded it on several measures.

The commercial strategy was straightforward and lethal: Netscape charged for Navigator; Microsoft bundled IE with Windows and gave it away. Every PC sold with Windows came with a browser already installed. Netscape’s revenue model — selling browser licenses — collapsed.

The DOM Wars and Incompatible Standards

As Netscape and Microsoft competed for dominance, both introduced proprietary HTML extensions and JavaScript APIs that only their own browser supported. Netscape introduced the <blink> tag; Microsoft introduced <marquee>. Each browser implemented the Document Object Model (DOM) — the programmatic interface through which JavaScript manipulated web pages — differently. Web developers faced an impossible choice: write code that worked in both browsers (enormously complicated), write separate codebases for each (expensive), or pick one and exclude the users of the other. The phrase “Best viewed in Netscape Navigator” (or “Best viewed in Internet Explorer”) appeared on millions of websites. The web’s promise of universal accessibility was fracturing along commercial lines.

The Antitrust Case

The Department of Justice had been watching Microsoft’s competitive practices for years. In May 1998, the U.S. government and twenty state attorneys general filed United States v. Microsoft Corporation — one of the most consequential antitrust cases in American technology history.

The central allegation was that Microsoft had illegally maintained its Windows monopoly by bundling Internet Explorer with Windows and pressuring PC manufacturers not to distribute competing browsers. The government presented extensive evidence, including internal emails in which Microsoft executives discussed “cutting off Netscape’s air supply.”

Judge Thomas Penfield Jackson issued his Findings of Fact against Microsoft in November 1999 and ruled in April 2000 that the company had violated the Sherman Antitrust Act. His proposed remedy — breaking Microsoft into two companies, one for operating systems and one for applications — was overturned on appeal. The final settlement in 2001 required Microsoft to share its APIs with third-party companies but imposed no structural remedy.

By then, it barely mattered. Internet Explorer’s market share had reached 96% in 2002. Netscape was dead — AOL had acquired the company in 1999 for $4.2 billion and effectively shelved it. Microsoft had won the first browser war completely.

The Monopoly’s Cost

Winning the browser war had a consequence Microsoft did not anticipate: having destroyed the competition, it had no reason to keep developing the browser. Internet Explorer 6 shipped in 2001. The next major version — IE 7 — did not arrive until 2006. For five years, the dominant browser on the planet received essentially no significant updates.

The web did not stand still. Developers wanted CSS support, better JavaScript performance, richer interactive capabilities. IE 6 provided none of these well. Its JavaScript engine was slow, its CSS implementation incomplete and inconsistent, its security architecture riddled with vulnerabilities that had accumulated from years of neglect. ActiveX — a technology that allowed web pages to run native Windows code directly — was a persistent attack vector that enabled a decade of malware infections.

Corporate IT departments locked to IE 6 by internal applications written specifically for its quirks became trapped. Upgrading the browser broke the applications; rewriting the applications required budget and time. IE 6 support lingered in enterprise environments well into the 2010s. Microsoft’s own engineers, years later, publicly begged users to stop using it. China’s corporate sector, where the browser had become entrenched in government and banking software, was still running IE 6 in significant numbers in 2014.

Firefox: The Open-Source Comeback

Netscape had made one consequential decision before dying: in January 1998, facing certain defeat, the company announced it would release the Navigator source code as open source. The project became Mozilla — a name derived from the internal codename for Navigator (Mosaic Killer, shortened to Mozilla).

The Mozilla codebase was large, tangled, and difficult to work with. Years of commercial development under deadline pressure had left it full of technical debt. A group within the Mozilla project concluded that the right approach was not to polish the existing code but to start over. In 2002, they released Phoenix — a lightweight browser stripped of Mozilla’s integrated email client, chat, and web page editor. Phoenix became Firebird, then Firefox.

Firefox 1.0 shipped on November 9, 2004. Within a year, it had been downloaded 100 million times.

Firefox’s appeal was concrete:

Tabbed browsing: multiple web pages in one window, switchable by clicking tabs — a feature IE still lacked
Extensions: a plugin architecture that allowed third-party developers to add arbitrary functionality, creating an ecosystem of tools (ad blockers, developer tools, password managers) that made power users deeply loyal
Standards compliance: Firefox implemented web standards correctly and consistently, giving developers a browser they could trust
Security: no ActiveX, tighter permissions, faster patch cycles

By 2010, Firefox had recaptured approximately 30% of the browser market. The IE monopoly was broken. The web had a competitive browser again, and web standards development — which had stagnated through the IE 6 years — accelerated.

Google Chrome and the Performance War

On September 2, 2008, Google released Chrome — accompanied by a comic book explaining its architecture, drawn by Scott McCloud. The technical choices were deliberate and significant.

Multi-process architecture: each browser tab ran as a separate operating system process. If one tab crashed or was compromised, it could not affect other tabs or the browser itself. IE and Firefox ran all tabs in a single process; a single bad web page could take down the entire browser.

V8 JavaScript engine: Chrome included a new JavaScript engine, written from scratch, that compiled JavaScript to native machine code rather than interpreting it. Web pages that took seconds to respond in existing browsers ran in milliseconds in Chrome. Google had financial reasons to want fast JavaScript: its products — Gmail, Google Docs, Google Maps — were web applications that ran entirely in the browser, and their performance depended on JavaScript speed.

Minimalist interface: Chrome’s UI was stripped to the minimum — a combined address bar and search field (the “Omnibox”), a row of tabs, no menu bar. The browser was designed to disappear, leaving only the web page.

Chrome’s market share growth was swift and then decisive. It reached 10% within a year, 25% by 2011, 50% by 2016. By 2023, Chrome and Chromium-based browsers collectively accounted for approximately 65% of global browser usage.

The competitive pressure Chrome created forced Firefox to accelerate its own JavaScript performance dramatically — to the benefit of web developers and users everywhere. And the underlying engine — Blink (Google’s fork of WebKit, itself derived from KHTML, originally developed for the KDE Linux desktop) — became the rendering engine of an expanding family of browsers.

Dead End: Internet Explorer

Internet Explorer did not die quickly. Its enterprise entrenchment meant it retained significant usage among corporate users long after Chrome and Firefox had displaced it everywhere else. Microsoft released IE 8 (2009), IE 9 (2011), IE 10 (2012), and IE 11 (2013), each better than the last and each still behind its competitors on performance and standards compliance.

In 2015, Microsoft acknowledged that the IE brand was irreparably damaged and launched Microsoft Edge — a new browser with a new rendering engine (EdgeHTML) and a clean codebase, bundled with Windows 10. Edge was a competent browser that arrived into a market already decided.

The ActiveX Legacy

IE’s lasting harm was not its market share but its technical decisions. ActiveX — the technology that allowed web pages to run arbitrary Windows code — enabled a decade of malware distribution. It was exploited in attacks that cost organizations billions of dollars in cleanup. The corporate applications written specifically for IE 6 and IE 7 trapped entire industries in outdated browsers for fifteen years. When Microsoft finally retired Internet Explorer on June 15, 2022, twenty-seven years after its launch, some Japanese local governments requested an extension because they still had internal systems that required it. The browser was gone; its consequences were not.

In 2020, Microsoft made the decision that acknowledged Chrome’s dominance completely: it rebuilt Edge from scratch using Chromium — Google’s open-source browser engine. Microsoft, which had launched the original browser war by cloning Mosaic, ended it by adopting its competitor’s engine. Edge Chromium launched in January 2020 and became the default browser in Windows 11.

The Chromium Monoculture

By the early 2020s, the browser market had reached a new kind of concentration. Chrome, Edge, Opera, Brave, Vivaldi, Samsung Internet, and dozens of other browsers all ran on the same underlying engine: Chromium/Blink. Only Mozilla Firefox and Apple’s Safari (based on WebKit, Chromium’s ancestor) remained as independent rendering engines.

This monoculture created a new version of the 1990s problem. When Blink implemented a feature — or declined to — that became the standard, regardless of what the W3C specified. Google’s decisions about what Chrome would support effectively set the web’s technical direction. A bug in Blink was a bug on 65% of the web.

The web standards process — maintained by the W3C and the WHATWG — continued to function, but its practical output was shaped heavily by what the dominant browser engine would implement. The open web had survived the IE monopoly; whether it could maintain genuine openness under a de facto single-engine regime remained contested.

For the JavaScript language that browsers run, see The Evolution of Language. For the World Wide Web that browsers navigate, see The Connected World. For the open-source projects — Firefox, Chromium — that kept competition alive, see The Open Source Revolution.