[This local archive copy is from the official and canonical URL, http://www.businessweek.com/1999/99_24/b3633183.htm; please refer to the canonical source document if possible.]


SCIENCE & TECHNOLOGY

A 'Rosetta Stone' for the Web?
The XML lingo could make it easier to find and use data

Miracle that it is, the World Wide Web is not without shortcomings. Mind-numbing delays and irrelevant search results are enough to try the patience of the most saintly Web surfers. And E-business companies have a long litany of woes, ranging from security problems to difficulties in swapping data across the Net.

A powerful technology has emerged that could cure many of these ills--if conflicting corporate agendas will allow it. Called XML, for extensible markup language, it's a new and clever way to tag electronic information with identifying codes. The coding scheme is based on the same principles as HTML, the current lingua franca of the Web. But next to XML, HTML looks like baby talk. XML employs far more precise tags to define and format electronic information (table, page 100). A number on an HTML Web page, for example, could be a price, a measurement, or a heart rate, as far as computers are concerned. An XML page, in contrast, will specify what that number represents.

BIG HITCH. If XML were widely adopted, searches would be more accurate and almost instantaneous, and business partners could link operations without all the crashes and fat consulting fees that E-business now entails. ''XML could turn the Web into one giant database,'' says Anthony J. Blake, a vice-president at AT&T Labs in Menlo Park, Calif. For translating information wending through disparate computer systems, he adds, XML ''could be the new Rosetta stone.''

There's just one hitch: For XML to perform as promised, there must be agreement on the tags used within various markets and interest areas. While XML's basic language was standardized last year by the World Wide Web Consortium, the tags that will be used to define specific data for banking, telecommunications, retailing, and other areas are still up for grabs. But software rivals aren't waiting. IBM, Microsoft, Oracle, Sun Microsystems--plus a host of E-business startups such as Ariba Technologies and Commerce One--are all trumpeting their unique expertise in XML. So the technology is already starting to fracture. Without better coordination, warns Timothy S. Sloane, an analyst with Boston's Aberdeen Group, the whole idea of universal standards ''could drift into never-never land.''

On May 25, a Boston-based standards group called OASIS (Organization for the Advancement of Structured Information Standards) tried to draw a line against fragmentation. It created a central clearinghouse on the Web, called XML.org, to coordinate XML news and proposals for tags. Unfortunately, Microsoft Corp. had the same idea. One day before XML.org went live, Microsoft unveiled BizTalk.org as the main repository not only for its own XML tags and products but also for those of its partners in E-business, such as Ariba and Commerce One.

XML enthusiasts studiously avoid bashing Microsoft, whose huge reach gives it unmatched ability to promote the language around the globe. ''Microsoft is not trying to co-opt XML,'' insists Forrester Research Inc. analyst Joshua Walker. Last week, the Redmond (Wash.) giant took steps to defuse the tension by joining OASIS. Still, Walker thinks Microsoft should go further and ''let OASIS host BizTalk.org.'' OASIS Executive Director Laura Walker says her organization would welcome Microsoft'sw participation--and Microsoft may not be averse. ''We have no interest in fracturing XML,'' says James Utzschneider, head of BizTalk.

Coordination would bring many benefits. Suppose, for example, you wanted to compare prices on digital TVs. Today, you either slog through online catalogs at each manufacturer's site or you punch in a generic search query. Either way, good luck. With keywords like ''digital'' and ''price'' saturating Web-site descriptions, you'd be better off browsing a magazine rack. With XML tags, a single search sentence would elicit all published prices. What's more, search engines would know to knock only at doors relevant to digital TV. Such efficient targeting, multiplied by billions of searches a year, would greatly ease the current strain on Web servers.

EYES ON THE ROAD. Smarter searches are just the beginning. Unlike HTML, which is a fairly rigid set of standards, the XML language is designed for flexibility. Groups with shared interests--from footwear to ice-cream manufacturing--can create their own dialects. Wireless equipment makers, for example, are working on a variant called WXML whose goal is a speedier Internet experience for road warriors. Web sites coded with WXML will recognize the memory, modem, and display limitations of the user's phone or other handheld device. Instead of sending pages laden with graphics, they'll dispatch text synopses.

Cell-phone companies Motorola, Lucent Technologies, and AT&T have another variant called VXML. It will convert XML text into speech so car phone users can keep their eyes on the road. This versatility matters to companies such as Charles Schwab Corp. ''We intend to give customers access to our services over wireless,'' says Dawn G. Lepore, Schwab's chief information officer. ''XML is part of the future.''

The technology's greatest impact, however, will be in E-business. Today, manufacturing giants such as General Motors Corp. connect to their channel partners and parts suppliers through an expensive and cumbersome protocol called Electronic Data Interchange, or EDI. This software painstakingly codes business information such as financial statements and purchase orders so disparate computer systems can accurately swap data. While EDI works, it usually requires that partners link up over the same private network. So a supplier serving two or more masters may need more than one EDI system.

XML changes all that. GM is working with AT&T and DataChannel Inc., a small company in Bellevue, Wash., to tailor XML tags so that any computer on GM's ''extranet'' will be able to deal with GM's data. That doesn't mean GM's system will be an open book, though. Its XML will be customized so that each business partner can see only the information it needs. Parts suppliers can come into the system, but they can't examine executives' salaries.

INSIDE EDGE? If XML's benefits are crystal clear, the business models are less obvious. Nobody can make money selling basic XML technology. It's free, as are many of the tools and tags already developed. For Microsoft, Sun, and IBM, the key is XML's efficiency for integrating networks. It means these systems giants can serve more corporate customers--and go after companies now using rival software.

Today, many companies are struggling with multiple computer systems that communicate only with difficulty. Sun's Java programming language was the first step toward a remedy, since a Java application can run on any computer. Now, explains Marie Wieck, director of XML technology at IBM's software group, ''XML will do for data what Java does for applications.''

Mention Java, however, and conflicting strategies come into sharp relief. Java has already fragmented, with Microsoft promoting its own approach. Nobody will be surprised if XML tags developed under Microsoft's BizTalk framework perform most effectively with Microsoft's own Web browser, Java dialect, and software development technologies.

There's no way to avoid such conflicts. But it would help if the giants could hold their fire long enough for XML to launch as a cohesive technology. Microsoft's Utzschneider promises that when the last pieces of the XML language are released this fall, ''we will adopt them and make any changes needed to BizTalk.'' For that, all consumers and companies will be grateful.

By Neil Gross in New York, with Michael Moeller in San Mateo, Calif.