See also A history of web development.
This page contains a short history of browsers from 1991 to the present day. History before 1998 is rather sketchy since I didn't live through it myself. I got my own first browser, Netscape 3, in 1997, and only from that time on I started following the browser market and its fluctuations.
The WWW was invented in 1991 and at first was rather geeky. The first browsers, invented simultaneously, was text-only. Then Mosaic 1 (Nov '93) hit the market, and it supported images. This was such a huge step forward that it quickly took over the (rather small) browser market of its day.
Since I didn't live through this Era I can't really tell the story.
After Mosaic came Netscape 1 (Dec. '94). Netscape took over the (still rather small) browser market at one stroke.
It was a much improved version of Mosaic, supporting multiple TCP/IP connections, cookies and the
Its mother company, Netscape Communications Corporation, asked money for the use of its browser but gave it away for free to carefully selected target audiences, like students and teachers. This helped immeasurably to spread the word and to ensure Netscape dominance.
And then the Internet went BOOM.
Microsoft hadn't been seriously interested in the Internet up until this time. The beginning Internet boom showed it the error of its ways and it made up for lost time quite fast.
Microsoft, too, needed a browser. It started out with another Mosaic version and produced Internet Explorer. It's best to draw a merciful veil over Explorer 1 and 2 (Aug. and Nov. '95), but with Explorer 3 (Aug. '96) Microsoft produced its first decent browser. It wasn't quite up to Netscape 3 standards (mouseovers, especially, were lacking), but to make up for it Explorer 3 was the very first browser to support CSS, until then an almost unknown technique.
Having a browser of its own meant nothing without marketing, though. Microsoft's first move was to make Explorer freely available to anyone, unlike Netscape, which still charged a modest sum for the use of its browser.
Explorer's market share grew, but it didn't even come closes to Netscape's. The main reason was that the Internet was still moderately geeky at that point. The average user knew a little bit about browsers, and had often heard the name Netscape. Added to its technical superiority, this ensured continuing Netscape dominance for the moment.
It was clear that Netscape and Microsoft were heading for a clash, and they needed new technologies to support their marketing efforts and give their browser a decisive advantage over the other. At the very tail end of the Second Era W3C, then largely an unknown factor, had published its CSS 1 specification, in which it laid the groundwork for separating content from presentation. This was interesting.
In general Microsoft made much better decisions than Netscape. First of all it rewrote its browser from scratch, so that it wasn't encumbered by legacy code any more. Conversely, Netscape tried to add the new features on top of Netscape 3's code engine, a decision that was to have grave consequences.
Secondly, Microsoft's DHTML implementation was targeted at web designers and developers knowing little of programming, while Netscape's implementation was more geared towards geeky programmers who wanted complicated structures for the sake of complicated structures.
Besides, Explorer supported one important feature that Netscape didn't (and couldn't) support: Explorer could reflow the page. When you use DHTML to, for instance, remove an HTML element from the page, all subsequent elements should 'flow up' to occupy the space left by the removed element. The lack of reflowing ability was the ultimate cause of Netscape's bag of tricks being far more limited than Explorer's.
So in hindsight Explorer was a much better browser than Netscape. It supported CSS1 far better than Netscape, and its DOM implementation was distinctly superior, too. Remember there were no DOM standards in those days, it was every browser for itself and God for them all (or, more accurately, for none of them).
Back in 1997, though, few people saw this. Netscape 4 (June '97) was released slightly earlier than Explorer 4 (Oct. '97) and its documentation has always been far better than Microsoft's. Besides, Netscape had a large and very vocal community of defenders (among which myself, in those days), while initially Explorer hadn't.
So right at the beginning of the Browser Wars the honours were roughly equal, though Netscape still held the market in thrall. That was about to change, though.
Microsoft had one huge advantage over Netscape: it did more than just making browsers. For instance, it made operating systems, too, and rather profitable ones. Not only could Microsoft easily pay for the development of a free browser, but the popularity of its OS made sure that it could easily distribute its browser, too, by simply including it in the OS.
And that's exactly what happened. Nonetheless, this fact alone doesn't explain the huge growth of Explorer's market share. If the WWW had remained confined to the semi-geeky 1995 community, many users would have laughingly trashed Explorer and used Netscape on their new computers.
However, when new users, attracted by the hype, bought a new computer, they wanted 'the Internet installed on it', no hassle, no downloads. Of course just double-clicking on an icon and starting to surf is much easier than downloading another program and installing it.
Finally, the new users were unencumbered by WWW mythology and did not know Netscape's name. Not only didn't they miss it, they didn't even know there was something to be missed.
Not that they really missed something. Netscape 4 showed more and more odd bugs.
Explorer soared, mainly thanks to new users. Netscape's only answer was to remove the price tag from its browser. Nice, but this action removed its main source of income and turned out not to be the answer. It also announced it would open the source of its browser, of which more later.
First of all its CSS support has always been excellent. Since Opera's CTO Håkon Wium Lie is one of the writers of the CSS specs this is not surprising. Secondly, Opera has always been a very light browser, both to download and to run, and therefore excellently suited for older systems.
During the Browser Wars Opera gained some market share, but stayed around 0.1%, not particularly much. Nonetheless it kept itself in the public eye, which considering its competitors was no mean feat. It didn't exactly thrive, but it wasn't a flash in the pan either.
In the US, but not elsewhere, the specialized WebTV browser also gained some market share. It was meant to browse the WWW from a TV and cruelly suffered from the limitations of this medium. TV screens are far narrower than computer screens, have odd colours and are generally not fit to show a web site designed for a computer. Nonetheless WebTV had some modest success. It was eventually bought by Microsoft, that has continued to improve it.
When Microsoft came with Explorer 5 for Windows (March '99), it became clear Explorer was not only winning the Browser Wars, it deserved to win them. Explorer 5 was the first browser to support large parts of the W3C DOM, boasted CSS support far superior to anything Netscape could offer, and was generally received favourably.
In contrast, the continuous streams of bugs that burst forth from Netscape 4's bowels was too much to bear even for the most orthodox Netscapist. A new version of the browser was mandatory if Netscape were to retain even a sliver of market share. Unfortunately the code engine was already working very badly and it could not possibly handle another major update.
So the ancient Mozilla code engine had to be rewritten. Netscape had already opened the source of its browser (Jan. '98) and called on developers from all over the world to create a new, better browser. The Mozilla Project was born.
This didn't help a bit in the short run. Explorer didn't merely win the Browser Wars, it trampeled its opponent, and then stomped on the bits for good measure.
Thus ended the Heroic Age and came the Fourth Era, which is characterized by little action on the browser front and a boring sort of bourgeois respectability. There are several reasons, the most important being that neither Microsoft nor the Mozilla Project released new browsers that extended technical frontiers.
Sure, Microsoft released Explorer 5.5 (July '00) and 6 (Oct. '01), but these releases didn't offer any radically new functionality, just incremental updates of CSS and the DOM. Nonetheless Microsoft at least showed some activity, something that couldn't be said of the sad remains of its competitor.
Microsoft surprised the web development community by Explorer 5 on Mac (March '00) which had the best CSS support of the moment. Unfortunately this browser was never really improved after its initial release, and nowadays it is distinctly behind the competition.
It took the Mozilla Project nearly two and a half years to reach the vaguely alpha-ish Milestone 17 (Aug. '00), and though the geeks saw that a brave new browser was approaching, it wasn't what you'd call a good piece of software for general consumer usage. It took another year a half to produce Mozilla 1.0 (May '02). The world didn't care quite as much as expected.
After the disastrous version 4 (June '00) Opera seemed to have lost the way. Fortunately it came back to the right track with Opera 5 (Dec. '00), though still behind its competitors. Opera 6 (Nov. '01) was not really a major upgrade, all in all, but version 7 (Jan. '03) has taken another major step.
The surprise of the Fourth Era was newcomer Konqueror (Nov. '00). It was a wholly independent browser, originally for Linux only, that had excellent support for all modern technologies. When Apple created its own browser, it took the Konqueror libraries as a starting point.
The release of Safari (first beta Jan. '03), based on the Konqueror libraries, spells the end of the Fourth and the beginning of the Fifth Era. Microsoft's hegemony on Mac has been broken beyond repair. On Windows it plays a waiting game, but Explorer 7 will only appear in conjunction with the new OS rumoured to be released in 2005 (which probably means 2006 or 2007).
Many web developers are sick of Explorer Windows, which hasn't had a major update in four years and is distinctly the worst modern browser CSS-wise (though not in W3C DOM support, there it takes second place after Mozilla).
I wrote an analysis of the browser market at the beginning of the Fifth Era, Browser Wars II: The Saga Continues on Evolt. It contains some tips for web developers who wish to break Microsoft's hegemony.
What's next? Time will tell.