To look backward for a while is to refresh the eye, to restore it, and to render it the more fit for its prime function of looking forward.
- Margaret Fairless Barber

The end of the year is here again, a time when we traditionally take a long look at the progress we've already made and then turn our eyes toward the future, attempting to forecast the year to come. With this in mind, Hitesh Seth sought input from industry leaders on their end-of-year predictions, and caught up with Michael Champion, research and development specialist at Software AG.

Michael has been active in the World Wide Consortium's Document Object Model (DOM) Working Group for more than three years and was an editor of the core XML portion of the DOM Level 1 Recommendation. He is now co-chair of the Web Services Architecture Working Group. Michael's current focus is on technical business development activities, writing articles on XML technology, and building example integrations between XML applications and Software AG's database and enterprise integration products.

Here, Michael shares his thoughts on the current state of XML and his expectations for the future.

Current Realities
XML is becoming pervasive, the "safe" choice for documentation, data interchange, application integration, business messages, and other areas. Everyone in the IT industry has an XML story, and to a very great extent these stories are based on XML standards that make interoperability much easier than it was in the past.

While some lament the complexity and diversity of the overall body of specifications from the W3C, OASIS, and other organizations, the real power of XML comes from its most basic and universal properties: the ability to represent any kind of data, with labels ("tags") associated with each value to make it easier to work with, in convenient packages that conveniently group together related information, leveraging the Unicode standard to support all computing platforms and human languages.

A Look Ahead
Consumers will vote with their feet against the complexity around the edges of XML just as they vote for the simplicity at its core. Products such as Microsoft Office 2003, which support XML only via the complex and widely criticized W3C XML Schema specification, will fail to meet their full potential until simpler approaches to document structure definition are supported by "aftermarket" products or a subsequent edition of the core product.

XML itself will become increasingly taken for granted as part of the "plumbing" rather than something visible to end users. This implies that XML infrastructure products and vendors will be chosen on the basis of their support for real standards, reliability, performance, and so on, and not on the basis of marketing and mind share.

Because XML is such an important part of the IT infrastructure, increasing attention will be paid to the obstacles that the current standards offer to those who deploy XML in environments where data communication is slow and XML's verbosity is a limitation, and in environments where processing speed is critical and XML's human-oriented grammar and features just slow things down. It's not clear whether these challenges will be overcome by improving the processors or improving the specifications, and there will be much experimentation in both directions.

The XML standards will be "refactored" (as the term is used in the Extreme Programming community) to support high-performance processing and to put what is really universally useful in the core, and relegate the complex and specialized features to an optional periphery.

All this will cause a certain amount of fragmentation: the data-oriented specifications such as W3C XML Schema will not be widely supported by those focusing on producing human-oriented documents, and the features of XML that appeal to human authors but pose great processing burdens (such as general entities and references) will not be widely supported in data-oriented applications. Alternative serializations of the XML data model that are more compact and/or faster to process will find support in specialized areas where these considerations are critical.

In spite of the challenges and fragmentation we'll see, however, the core XML vision and standards will provide a point of unity across all the diverse communities that develop specialized tools and formats to meet their own needs.