Smoke, Mirrors and Silence: The Browser Wars Reignite
Date: May 28, 2004
A new web browser war is brewing. The first sortie, a mini-war over browser tools, is already upon us. This war is different from the hysterical browser war of the 1990s, though: This time, it's for keeps.
In the 1990s, the browser war was a fight for consumer loyalty through novel browser features. If you manufactured the user's web browser, you were a media company. Rhetoric about new browser features was a key strategy back then; this time around, that kind of discussion is not as meaningful.
In this decade, the new battle is for the survival of the web itself. Much of the fight takes place in the background, away from consumers. The choices you might perceive you have are really only the tip of the iceberg. To ordinary consumers, this browser war is as inexplicable as trying to understand why Betamax, an early standard for video recorders, was driven into obscurity and failure by VHS. Examined objectively, VHS just wasn't as good as Betamax. Why on earth did it succeed? The new web browser war is like that.
An Update on Browsers
Superficially, the new browser war is still about web browsers. On one side is the aging but popular Microsoft Internet Explorer 6.0. On the other side is a range of highly polished but less frequently used tools. These tools are effectively led by the Mozilla Foundation, whose technology now appears in many different browser and non-browser products. Among browsers, the Mozilla Foundation offers the Mozilla Application Suite, Firefox, and Camino. Beyond the Foundation are many other Mozilla-enabled browsers such as K-Meleon. In fact, Mozilla has built a niche on just about all platforms—from mobile phones and PDAs to mainframes. Non–Mozilla browsers such as Safari and Opera ensure that the web has not yet been reduced to a two-horse race between Microsoft and Mozilla. It's still a multivendor environment.
Mozilla is responsible for most of the early shots in this new war:
The Mozilla browser is technically better than IE. That is plain fact.
Mozilla has innovative and polished user-centric features—that IE doesn't have—such as tab-based browsing. Microsoft has roused from its browser slumber long enough to suggest that the same polished features are coming to IE sometime soon. That alone has raised a few eyebrows.
The recent announcement by AOL Time Warner that they will update their own browsers with the latest Mozilla technology also gives Mozilla technology a tick.
Mozilla technology just won't go away: Its popularity grows slowly but surely, like tree roots crumbling a rock. As of April 2004, independent statistics show Mozilla to be 4% or more of the global market, but for web developers, use may be as high as 10%. In some markets, such as Germany, use may now be as high as 19%. On some platforms, such as Linux, Mozilla is now the dominant player, and probably well over 50%.
Popularity Isn't All It's Cracked Up To Be
Reducing browser analysis to a popularity contest, however, seems to have a magical effect: It turns off the brains of most commentators. Popularity is an ogre that prevents any real analysis of industry trends. As a result, we're all less informed than we should be. Normally perceptive pundits such as Jon Udell are reduced to "I like it" (in his review of Mozilla's Firefox). Microsoft mavens like Mike Langberg respond "I don't like it" (Firefox again; free registration required to access this MercuryNews.com review).
Overall, the votes stack up in favor of Firefox (Kim Komando, Rob Pegoraro). But even those in favor are sometimes deflated by the sense that perhaps, for the average person, the quality work that better features imply just isn't that important. Homer Simpson, they imagine, would stick to Internet Explorer and not know the difference.
Such remarks highlight the "consumer lifestyle" aspects of browser use and the importance of design, but say nothing about the web from an architectural or global-commons point of view. If web browsers are reduced to consumer lifestyle options, there can hardly be an intelligent debate about the infrastructure they represent. Or can there? Surely browsers are more than yo-yos or Frisbees. Weren't the browser wars (round 1) about something bigger than digital lifestyle options? Wasn't there something actually at stake?
Today, few commentators are willing to dig deeper. Dan Gillmor is willing to apply some strategic analysis: Land all the lawsuits you want; if IE is available on Windows by default, that's what people will use. It's an old saw, unfortunately. That logic may explain IE's popularity, but mere popularity is no source of alarm. Why, then, did we care so much in the past whether IE was embraced? Alison Diana has perhaps a better recent summary: It's the security and standards holes in IE that continue to be problematic to those who care about the state of the global Internet.
Set Up for Failure by Security Systems
I must say that I agree with Alison. Lately I care about security very much. I'm forced to maintain multiple browsers on my equipment because it happens to be my personal problem that I spend a lot of time with browser technology. For two years or more I've barely touched Internet Explorer's security features because of the leaky, tricky, buggy mess I long ago concluded that they represent. Alas, last week I did some brief security testing and forgot to reinstate high security afterward. It took just one idle hour of web surfing on low security before some pathological web site designer leapt down the throat of my PC through the Swiss cheese that is IE. He totally messed up my computer. I'm willing to admit that it was my fault, but somehow that's cold comfort when you're on crutches afterward.
I happen to be technical enough to go through Microsoft Windows with a fine-toothed comb, twiddling bits over here and checking for secret compartments over there, but in this case even that strategy didn't fix the intrusion. I had to scrub the box back to the bare metal and reinstall two years of software, all merely to get Mr. Psychotic Web Marketer off my case, and my PC stabilized. Trust me, that's a lot of software application preferences to change, just to get my workflow back up to speed. "Not happy, Jan," as we say down here. Thank you IE, I don't think. I ought to use a disk imaging package, but I really don't want to spend my life preparing for every stupid accidental security breach. A computer shouldn't require the management of a nuclear power plant.
The first reason we still really need a browser war, then, is just so that we can have products that let us surf the web in relative safety, with some protection against our own temptations, absentmindedness, and vulnerability. I don't mean safe from computer vulnerability—I mean "safe at any speed," as in Ralph Nadar on cars. Thank you, everyone except IE. I just don't need those time bombs hanging around.
History Repeating Itself: The Value of Standard Data
Even security features, however, come down to a popularity vote. It's Alison's other point—standards—that comes closest to the heart of the new browser war. The new war isn't about popularity, digital accessorizing, or even consumers. It isn't directly about standards either—most of the W3C's web standards are now well settled and fine-tuned to the smallest degree. The new war is about the data that makes up the bytes of text in all those HTML document on the web.
Once before, we had a big fight about data. That was in the early days of relational databases. When Oracle was releasing version 2.0 of their flagship product (there never was a version 1.0), few could imagine a practical use for a relational database management system (RDBMS). Compared to the then-existing network and hierarchical databases, the performance of SQL engines was abysmal. Like today's web-browser popularity arguments, all you has to do was say "performance" and that was apparently the end of the argument for SQL.
Of course, SQL not only survived and thrived, it eventually dominated. Relational data is highly standardized, highly portable, highly interoperable, highly queryable, and even human-readable sometimes. Early RDBMSs were slow, but the idea of manageably open data propelled the technology forward anyway, until the performance problems faded away. The systematic benefits of manageable data drove RDBMSs and SQL into the spotlight. SQL was never about performance. It was about open, manageable data. Performance issues were nothing more than a hall of mirrors, or a teething problem.
At the time, competition between RDBMS vendors was fierce, even after Oracle gained the edge. The emergence of the SQL92 standard represented a crisis for all vendors. If SQL92 was widely applied, RDBMS servers would be reduced to a marginally profitable commodity. The database vendors knew that open data was a customer issue, and to make money they had to show that they were dedicated to addressing the issue. There followed an extensive period of rhetoric about striving for standards compliance. In reality, though, no vendor wanted standards compliance. For a long time, there was a "go slow" on actual standards implementation from all vendors.
Another significant example was portability of CASE tools metadata. By the time the hashed-out CASE standards were supported by implementations, no one really cared. It no longer mattered whether your data could be moved from one CASE repository to another.
In summary, the strategic value of standard data support in the world of SQL and RDBMSs was blocked by the vendors. This was done despite both public and customer interest in standard data.
Everything Old Is New Again
The new browser war is the same landscape, with a few modern twists. The bits that are the same are obvious:
We have the standards for the data. HTML, XML, XHTML, CSS, and a plethora of other standards commoditize web browsers and various bits of server infrastructure used for the web.
We have the same profile for the data. In the best of all possible worlds, it's highly portable, highly manageable, highly interoperable, highly queryable.
We also have a hall of mirrors that keeps us stupid. This time it's not performance that dazes us—it's popularity.
The modern twist that we face is that browsers are nearly commoditized already. They're free, and there's little money in them—certainly not the megabucks of the early relational era. In fact, lots of web technology is free. Apache is free; HTTP is free; PHP, JavaScript, and Perl are all free. The web therefore represents a no-value zone for tools vendors such as Microsoft. If standards are broadly applied to that zone, Microsoft is effectively prevented from extracting any monetary value.
Standards have historically not been broadly applied to the web, and Internet Explorer is the reason. Deep in its little mind, open, manageable, standardized data is not that welcome. Microsoft is silent on standards compliance, and in our hall of mirrors mentality, all of the discourse on browser popularity does nothing but support Microsoft's position. It distracts us from the true issue of the value of standard data. Microsoft gets the required smokescreen rhetoric for free. (Oracle must be livid that they had to do the work themselves to delay full SQL92 support.)
And So, to the New Breach
Fortunately, some players care little about money or popularity. The Mozilla Foundation has systematically been building open, manageable, standardized data support into its offerings. Profit is not a primary motive. Mozilla technology has a bloody-minded attitude: standards-compliant or bust.
Perceptive customers and consumers see this standards support emerging. The web development community is starting to grasp the use of standards: XHTML is now an issue, and CSS and HTML 4.01 are in popular use. Many market segments have an interest in a standards-oriented environment: education, government, the military, large corporations, and enthusiasts. All these factors serve to strengthen the web as a global commons and as a marketplace—in which there is no value for tools vendors such as Microsoft.
Make no mistake: Microsoft really hates the web. The new browser war may appear to be about the emergence of Mozilla and friends with their polished eye-candy interfaces, but it's really about Microsoft versus the W3C. Internet Explorer is Microsoft's blocking tactic—never to be properly web-compliant, never to give the W3C a day in the sun—and Longhorn technology is the big-stick alternative being built. One of the purposes of Longhorn is to destroy the web as we know it.
The web is used to provide a variety of services and communities. Part of the Longhorn strategy is to extract from the web all of the services with any profit model at all: web magazines, auction sites, news, online retailers, and so on. When Microsoft tempts these organizations and communities to Longhorn, the web suffers the death of a thousand cuts. Over here will be the standards-based web, with a gradually shrinking set of web sites. Over there will be the future Longhorn-based proprietary global infrastructure—a global version of the early Novell NetWare, a sort of stock market/CNN fusion for content delivery. For Microsoft, the best possible outcome is for the standards-based web to be reduced to the profitless: a few idealistic hippies, some idle perverts, and the disaffected. Few others will want to go there; so every day there will be fewer traditional websites, every day less relevance.
This is a fundamentally different attack from that of the browser wars, round 1. Instead of fighting for control, the new browser war is a fight for the survival of the web itself. In this new war, the eye candy offered by new and polished browsers is a necessary but insufficient response to the stonewalling of Internet Explorer as a precursor to Longhorn. It's the presence of standardized data in web content—whether current standards such as XHTML or some yet-unknown future standards, perhaps based on XUL—guaranteeing that the web will remain a global commons, an information highway, and a free marketplace. The alternative is a corporate Diaspora and a tollway.
The decision to generate such content is not the province of consumers or hobbyists; it's the province of organizations. Organizations must wake up to the value of open and manageable standards-based web data, and cease being stupefied by irrelevant popularity arguments.
Historically, to standardize your web data was something of a commitment to an ideal. Ideals alone are not very attractive to pragmatic organizations. Standardizing data should not be an act of penance; it should be about sustaining communities and markets—ones from which service agendas and profits derive. If organizations don't see the web as a useful global commons into which they can deliver their services, that global commons will vanish as a community and as a marketplace. That is the new browser war. Either tweak your web data to approximate standards now, or risk the cost of a massive Longhorn infrastructure upgrade in a few years' time. If you don't standardize now, you'll be forced to buy or build that Longhorn infrastructure in order to access the communities that Microsoft has managed to attract and bottle up inside Longhorn. It's very easy to do the math and conclude that it's cheaper to support your existing public rather than relocate and rebuild when that public goes west.
Fortunately, some organizations are indeed modernizing and standardizing their web data. We need to see that movement reported more often. No doubt it's harder work than reacting to press releases about the latest browser share, or playing with a new lifestyle tool, but readership is served just as well by investigative journalism as it is by idle review.
In this new war, individual action is still important, so choose a standard compliant browser if you value the web, or if your job earns value from it. This time around, however, it's just as important to discuss with your peers the state of your web data—the very data that will protect the web from self-implosion in the face of competition. Standard data guarantees that you won't have to migrate to Longhorn in order to stay where you are.