Posted at 5:16 AM PT, Jul 19, 1997
Although it's still making its way through the standards committees, the Extensible Markup Language (XML), scheduled to appear in products in 1998, is already shaping the future of the Web.

XML provides the underpinnings for both Microsoft's proposed Channel Definition Format (CDF) for pushing content and Netscape's Meta Content Framework (MCF) for browser presentation. And if Sun Microsystems has its way, XML will be the content standard that gives Java something useful to do.

Although HTML is emerging as a dominant file format, XML promises enough flexibility as a programming language to fundamentally alter the development of client software and potentially upset the balance of power in the industry by eliminating the need for large, monolithic clients such as Windows, Microsoft Office, and Lotus Notes.

"XML does for data in general what Java does for programming in general: Java gives you portable code, XML gives you portable data. They are the two shoes of an entire platform-independent programming picture that is going to move us into the future," said Jon Bosak, an online information-technology architect at SunSoft and World Wide Web Consortium (W3C) XML working group chairman, in Mountain View, Calif.

Miko Matsumura, a JavaSoft evangelist, in Cupertino, Calif., believes that Java, in concert with technologies such as XML, will significantly shift the programming paradigm from developing "containers" -- traditional applications such as Microsoft Office -- to wiring together a series of JavaBeans to create easily deployable yet highly customized applications.

XML, which is currently a W3C working draft, is a subset of the ANSI's Standard Generalized Markup Language (SGML) designed for use on the Web.

XML provides a meta language for creating Web markup languages and augments HTML by letting industry groups create tags that are specific to their needs yet understandable to any XML-enabled client. For example, a health-care consortium could agree on tags such as or to embed in patient records applications.

Although XML clearly will extend HTML, the technology is not intended to replace it.

"I wouldn't say XML is the second coming of HTML because HTML is not going away," said Lauren Wood, a vice president at Soft Quad, in Ottawa, and a member of the XML standards committee. "HTML will be useful for a certain class of application, and XML will be useful for another."

SunSoft's Bosak said that XML will ease application development in five main categories: data interchange; greater distribution of processing from servers to client; different views into data; intelligent agents; and meta data, or data about data.

IBM, for example, sees XML as a useful carrier of information between applications.

"We could use XML as a way to make two other programs talk,'' said David Singer, a program manager working on advanced Internet technologies, in Somers, N.Y. "I could have a database query generate XML and have a program somewhere else deal with that XML and not have to worry about things like the number of columns in a field changing.

"Right now the best you can do is mark something up as a table and hope that something at the other end knows it is a parts catalog," Singer explained. "But with XML you can mark it up as a parts catalog, creating that markup dynamically, and have an applet at the other end that knows how to deal with that markup."

Singer and others believe that XML will also take the pressure off HTML, which is getting bigger and bigger as developers continue to add tags in order to get the technology to do things it wasn't designed to do. Unlike HTML, XML could allow an online catalog, for example, to describe the different properties of its contents, such as color, size, and shape.

With Internet heavyweights Netscape and Microsoft putting their muscle into XML development, the technology's future looks bright. Microsoft's CDF uses XML to describe tags that will enable push-technology vendors to more easily create Web channels.

Also, a proposed Open Financial Exchange language derived from XML aims to integrate Microsoft, Intuit, and CheckFree electronic banking and payment protocols.

Microsoft is girding up for an XML products push in the area, said Tom Johnston, product manager of platforms marketing at Microsoft, in Redmond, Wash. The company already is building an XML parser into Internet Explorer 4.0.

Last month Microsoft released a stand-alone XML Parser program written in Java. Netscape's MCF, which will be supported in the company's forthcoming Web client, Mercury, will let Netscape and third parties create visual display controls and define pushed content.

Because vendor tags, such as Microsoft's CDF and Netscape's MCF, overlap, the W3C is expected to announce by the end of next week a Resource Description Framework working group that will serve to coordinate vendor efforts and iron out XML meta data-format issues, said Dan Connolly, the W3C Architecture Domain lead, in Cambridge, Mass.

Tools vendor SoftQuad plans to release a version of Hot Metal Pro that, though not based on XML, will introduce some concepts and capabilities associated with the language.

The W3C also has under way a method for making the Document Object Model -- a proposed means of formatting document objects for use in Dynamic HTML -- work with XML, according to Microsoft's Johnston.

"Content developers and providers will only need to learn one object model, not two," according to Johnston.

Because this technology is in its very early stages and only a small circle of developers are now keyed into the technical aspects and programming ramifications of XML, most third-party and corporate developers appear a bit leery of the complexity it may introduce.

"We are in the business of generating HTML through C and C++, and the thing we like about HTML is it is nice and simple," said Richard Smith, president of Phar Lap Software, in Cambridge. "So we worry a little [about] what complexity XML may introduce."

Writing to XML will be more complex and may demand more programming talent.

"The difficulty will be that people who do HTML development may not necessarily be able to write things like XML parsers that can manipulate meta data," according to Phil Costa, an analyst with the Giga Information Group, in Cambridge. "You are talking about writing in a layout language compared to writing data manipulation and C++ code."

But some vendors disagree, pointing to the parsers already available and contending that writing in XML will eventually become easier than writing in SGML and even HTML.

"One of the aims of the XML group is to write a spec that could be done by your average computer science graduate in two weeks. It is not rocket science," Soft Quad's Wood said. "It could be as easy as grabbing one of the public domain parsers and folding your own code in."

Because XML continues to be a relatively closed conversation among the programming elite, most IS professionals remain far removed from its implications.

"I am for anything that makes development easier and results in more flexible applications," said John Henderson, a technical consultant with a large Maryland-based utilities company. "An important issue is, how much will it cost me in training? But this all sounds like it will be committee-ware for a while."