What the Next-Generation Web Means for Managers

Imagine a company's computer system that could read and make sense of any other company's paper documents—including purchase orders, invoices, loan applications, medical forms, and contracts. Or a customer service center whose database automatically updates, across separate documents, all treatments of a specific problem with the latest technical fix or workaround.

Imagine a medical publisher able to take content related to a particular illness from a variety of sources (books, magazines, Web sites) and "aggregate" it into a new document—all without having to recode the information. Or a robust search engine that retrieves only the precise information you're looking for—without the reams of irrelevant data that typically accompany most Web searches today.

	Today, you need a human being to access information and services on the Web. But a set of standards known as XML will soon enable computers to do most of the work.

It's all very heady stuff, and it's also very close at hand. Such a future is being made possible by the ability to manipulate information in incredibly rich ways. That capability, in turn, rests on a set of standards known as XML. To appreciate XML's potential, you need to know a little bit about its technological lineage.

The genealogy of markup languages
SGML (Standard Generalized Markup Language) has been the international standard for the formatting of text and documents since 1986. Although it's very popular among organizations that manage large volumes of data, SGML is not supported by the most common Web browsers. As a result, HTML (HyperText Markup Language) has become the predominant markup language for the Web because of the ease with which it displays simple documents. Basically, all HTML does is describe how the information appears—say, in a 12-point Helvetica Bold font. By contrast, SGML is much more flexible: it enables you to organize and manipulate information according to its content.

XML (Extensible Markup Language) makes the richness of SGML available on the Web. Technically, XML isn't a markup language at all. It's "a syntax that allows you to separate data—information assets—from the processes that act on them," explains Brian Travis, chief technology officer for Architag International, a firm in Englewood, Colo., that provides consulting and training on issues of information asset management. XML makes it possible for you to mark up or "tag" information however you choose. So instead of just being able to use formatting tags like "12-point Helvetica" or "bold," you can also use tags like "article," "summary," "innovation," or "leadership" to describe the content.

A new age of B2B activity
"HTML allows you to describe how information looks," Travis continues. "It describes the information to human eyeballs. But XML lets you describe the information in terms of what it is, which allows you to expose your data to computers." In other words, XML ushers in an age in which computers, not humans, will do most of the accessing of information on the Web.

Just what would this look like in practice? Arbortext of Ann Arbor, Mich., one of the members of the W3C, provides an extended example in a white paper located on its Web site (See HBS Working Knowledge Close up below): a faucet manufacturer with an online parts catalog. Using XML, the manufacturer could create tags for the various rooms for which its faucets are designed, the types of controls on each faucet, the finishes available, and the price of each model. Such application-specific tags make the following activities possible:

• data interchange. Building contractors could take faucet information off the manufacturer's Web site and incorporate it into their estimating software and design systems.

• automation. Clicking on a faucet's part number could bring up an order-entry form.

• precise searches. If other manufacturers use the same tags, a customer researching faucets could create a query that would retrieve only the appropriate data.

These possibilities will become realities with the next-generation Web, says Howard Smith, director of strategy for e-business at CSC Europe, an IT solutions provider based in Farnborough, U.K. "Today we have the publishing Web, which was invented in 1992-93. By 1994, it had early adopters—large companies that had created home pages. The publishing Web is now ubiquitous—without any enabling infrastructure of your own, you can have a service provider create a Web site for you.

"The transactional Web represents the next horizon. It was invented in 1994-95, when the first companies started using the Internet to sell products and services. By 1996-97, companies were using the Web for back-office transactions; it will probably be another year before the transactional Web is ubiquitous.

"And on the horizon beyond that is the process Web, which will enable businesses to publish and share business processes in all the ways that they can share data today." Not just transactions, which are the end results of processes, but the processes themselves in their entirety. Ordering, payment, marketing, new-product development—everything will be transparent. "Companies will be able to share processes without intermediaries," Smith continues, "and the sharing will be as pervasive as e-mail is today. The process Web is being created now, but it may be 2004 before it is ubiquitous."

The problem of agreement
XML enables you to create an infinite number of tags for your information. But before you can achieve interoperability—that is, before you can share information and intent with your trading partners—you need agreed-upon conventions. All told, some 200 groups are working to create XML conventions and vocabularies for a host of document types, processes, and vertical industries. For instance, Ontology.Org, a CSC-sponsored think tank, focuses on interoperability for Internet commerce. The Universal Description, Discovery, and Integration project—a consortium of 35 companies including IBM, Ariba, and Microsoft—is working to create a Web-based business registry containing details about what a company does, the software standards it uses, and how other companies can connect to it to complete transactions. Vendor-sponsored solutions also seek to solve the agreement problem for specific functions or processes. For example, Extricity, Inc., of Redwood Shores, Calif., makes software that helps companies coordinate cross-enterprise processes with private processes.

Why all this activity? Because, says Smith, "interoperability, not security, poses the chief obstacle" to the development of the next-generation Web. Stay tuned.

Reprinted with permission from the Harvard Management Update, November 2000.

See the latest issue of Harvard Management Update.