With Web’s Growth Comes the Standard Question

The growth of the Internet’s World Wide Web has been so rapid as to be frightening. Like the movie monsters that swallowed New York, its relentless advance is seemingly impervious to obstacles and outside of human control.

But, just as every monster turned out to have its Achilles heel, the Web seems to have a serious flaw. The problem is with the language used by the Web to describe the appearance and content of a document. Known as HTML (HyperText Markup Language), it acts as the lingua franca for the Web, ensuring that anything posted on the Web can be viewed by any of the many different Web browser programs.

At least, that’s how it used to work. But in the last few months, companies such as Netscape, Prodigy and Microsoft have announced Web browsers that support non-standard, “enhanced” versions of HTML. This means a Web page that uses enhancements designed for the popular Netscape browser may appear jumbled when read by a browser from Microsoft. Jeff Veen, a designer for the Hotwired Web site, likens the situation to having TV channels that can only be viewed on certain TVs. Suddenly, the Web has become a lot more complicated.

This is a critical issue for the Web, and it’s also part of a larger problem that seems endemic to digital technology. Whether it’s video discs or personal computers, there always seems to be some sort of battle over standards. And yet, despite the issue’s importance, it’s a problem we’re not very good at solving.

Looking at the current battle over HTML illustrates problems with the ways we usually choose standards. But it also provides the glimmer of a solution.

Let’s start at the beginning: What’s wrong with the existing standard? HTML was touted as well-designed, even visionary, just a few years ago. And we know it works. So why change?

The obvious reason is user demand. Web users are starting to ask for better security, prettier graphics, and support for on-line shopping--features that HTML simply wasn’t designed for. After all, the Web was originally used just by physicists for exchanging information. It’s no wonder HTML looks a little dowdy.

But there is a second force that is pushing the standards process far faster than user demand is pulling it: the dozens of ambitious companies that develop Web browsers. Think about how they compete. If every browser supports the same standard for describing documents, all the browsers end up pretty much alike. But if a browser supports a unique variant of HTML that includes extra bells and whistles, well, that browser is something special and worth buying. The companies are really selling standards as much as browsers.

Given this situation, how can the Web community ever decide on a new standard? It’s as if Coca-Cola and Pepsi had to agree on the best soft drink. Forming a standards committee to decide--the usual solution--doesn’t work when there is no incentive for companies to abide by the decision. Just look at Netscape, which currently controls the largest share of the browser market. Despite belonging to the appropriate Web standards committees, Netscape was one of the first to introduce an “enhanced” HTML.

According to Netscape’s Marc Andreessen, the right place to decide Web standards is the marketplace. Let the invisible hand choose the winners and losers. If Netscape extends HTML in a way that people find useful, other browsers will be forced to support the extension and it will become a de facto standard. Similarly, if some other company introduces a popular extension, Netscape will add support for that. In this way, the Web will rapidly evolve as browser companies fight to distinguish their product from the rest.

That’s the theory, anyway. If you want to know how it works in practice, talk to today’s Web designer, who has to create four versions of every page. Or talk to the users who find themselves upgrading their Web browsers every month. Right now, the process is just too anarchic to produce stable standards.

But that may change now that Microsoft has entered the arena. By including a Web browser with every copy of Windows 95, Microsoft will probably come to dominate the market. And, based on Microsoft’s history, you can bet it’ll exploit its market share to dictate the future of HTML.

True, the result would be a single, stable standard. But it’s all too likely to be the wrong standard. From Qwerty vs. Dvorac keyboards to Beta vs. VHS cassettes, history shows that market share and technical superiority are rarely related.

It might seem as if this leaves us at an impasse. If neither standards committees nor the invisible hand of the marketplace can be trusted to choose HTML’s successor, what’s left? Simple: Avoid choosing. Instead, let a million standards flourish, but carefully design a meta-standard--a method for describing standards. By sending information about the standard along with the document, a browser can translate. The Web becomes like a library that has books in many languages, but where each book comes with its own Rosetta Stone.

This is possible with specialized programming languages like Sun Microsystems’ Java. It allows small programs to be sent over the Web and then run on the receiving computer. These programs could act as the Rosetta Stones. For example, a Web site that uses an esoteric language for document description would first transmit a short Java program. That program would then run on the user’s machine, acting as a translator for the subsequent data sent by the Web site.

This approach lets us return to the time when any Web page could be read by any browser. Also, by allowing multiple standards to flourish, it helps speed up innovation. True, you still need to pick the meta-standard. But this should be less contentious, since browser vendors aren’t already tied to their own private systems.

It’s really an old-fashioned solution: When a problem seems insolvable, pass the buck up the ranks. But it answers a critical need by preventing the Web from fragmenting into a Babel of dialects, or from being stuck with an awkward tongue. This way, the Web can continue growing unchecked.