[This local archive copy mirrored from the canonical site: http://www.gcn.com/gcn/1998/August10/1B.htm; links may not have complete integrity, so use the canonical document at this URL if possible.]


GCN August 10, 1998


XML promises to transform government’s use of the Web

By Florence Olsen
GCN Staff

What’s on the horizon for XML

Intelligent agents
Structured records
Objects with methods and data
Metacontent for Web sites
Query results
Graphical user interfaces
Persistent storage formats
Online procedures documentation
Interactive parts catalogs
Electronic service manuals
Electronic data interchange

Government information managers and commercial application developers alike are buzzing about the Extensible Markup Language, which tags nontext Web content for easier searching and delivery.

“This is one of the rare technologies that all the big players, who tend to disagree on most everything else, are in favor of,” said Charles F. Goldfarb, the Standard Generalized Markup Language’s chief inventor. “XML is mass-market SGML.”

Goldfarb said he thinks XML will fundamentally change the ponderous nature of data processing and storage. “The vast majority of XML documents are going to be generated instantly from databases, read by other programs and then destroyed,” he said.

In XML, Microsoft Corp. and other industry leaders see the potential for transforming the way Web content is distributed. Government software developers like it for different reasons.

Engineers at NASA’s Goddard Space Flight Center in Greenbelt, Md., plan to use XML to develop an instrument control language for infrared devices on satellites and space telescopes.

“It’s a little different from what XML is normally used for,” said Troy Ames, a computer engineer in Goddard’s Advanced Architectures and Automation Branch.

NASA engineers will use XML syntax to describe classes of infrared instruments, control procedures, communications protocols and user documentation. Computers will parse the tagged data and generate instrument control code, most likely in Java, Ames said.

At the moment, the engineers are still trying to pick their XML tools. “There are so many new ones coming out every day that we haven’t committed to writing any tools ourselves yet,” Ames said.

Common denominator

As a format for data exchange, XML can package objects from multiple data sources, and XML-aware applications would be able to read the enclosed data.

“Millions of Excel programmers want to pull data into their spreadsheets” from mainframe and Web server sources, Goldfarb said. XML might become a universal translator, he said.

Unlike the Hypertext Markup Language, which has a built-in presentation style, XML is flexible.

It can pull data from databases, merge it with different style sheets and display it on everything from a large-screen computer to a small-screen pager.

SGML, which was developed for tagging and publishing large documents, “filled a need for the government that couldn’t be fulfilled anywhere else and still can’t be,” said Goldfarb, who has published books about XML and SGML. Later HTML, a subset of SGML, “came along and created this huge application of SGML for online short-document publishing,” he said.

But far more useful than HTML’s text description, Goldfarb said, is the way XML and SGML can describe nested object structures and properties.

“Object databases are great for rich hierarchies, which is exactly what XML is,” said Mike Hogan, vice president of business development for Poet Software Corp. of San Mateo, Calif.

Hogan said he sees XML as the killer app that could create a mass market for object-oriented databases, which so far have succeeded only in niches such as network management and process control.

About two years ago, Poet Software discovered another niche market for its object database—SGML publishing. A year later, as XML started to create a buzz, Poet Software began tuning its Poet Object Server database to store XML objects.

“XML will do for object databases what tabular data did for relational databases, only in Web time,” predicted Dirk Bartels, president and chief executive of Poet Software.

Bartels said he wants to make Poet Object Server the object repository for all kinds of XML, HTML and SGML applications. The object database server supports the Object Data Management Group’s application programming interface.

Nimble app

The company’s Poet Content Management Suite, built on top of Poet Object Server 5.0, has revision control, check-in and check-out capabilities, full-text indexing and workgroup collaboration support.

But Hogan said XML cannot reach its potential in electronic commerce, electronic data interchange, data reuse, improved search engines, push technology and interactive electronic publishing until there is an adequate base of XML-tagged data on the Web for such applications to tap.

The first step, he said, will be building a critical mass of XML-tagged Web data. Leading database vendors such as Oracle Corp., IBM Corp. and Microsoft are designing ways to return XML-tagged data in response to Structured Query Language queries.

Microsoft’s Internet Explorer browser supports XML minimally, and Internet Explorer 5.0 will do much better XML parsing, Hogan said. The next release of Netscape Communicator will support XML, too.

For XML information, visit http://wdvl.internet.com/Authoring/Languages/XML.