A few years ago at an early XML conference, an attendee made the point that XML was such a useful technology for data portability that it would eventually become ubiquitous - part of every tool, server, and application. He went on to predict that XML would become so commonplace that the idea of attending an XML conference would eventually seem silly. "In a few years, a conference on XML will seem as ridiculous as a conference on ASCII would today," he quipped.

He was on the right track, but by focusing on XML only as a way to move data between proprietary applications even this bullish prognosticator was quite likely understating XML's long-term potential to improve the capabilities of information systems.

XML has been a smashing success. It's widely used in publishing as a lightweight alternative to SGML for separating form and content. It plays a key role in messaging and business-to-business data exchange, offering a low-cost replacement for technologies such as EDI. Its on-the-fly transformation capabilities make it ideal for personalized user interfaces on non traditional computing devices. It makes software more interoperable by providing a language-independent foundation for defining vendor-neutral interfaces, such as OASIS's WSRP portlet initiative and JSR 172 for wireless devices. XML is also the key technology that makes Web services stand out from previous generations of distributed computing models. And XML is well on its way toward ubiquity. No self-respecting developer today even thinks of defining a new interface or config file without turning to XML, and some level of XML support has crept into just about every product in every category of software.

As significant as these contributions are, however, their impact pales in comparison to the longer-term change that XML will help bring about, which involves a fundamental shift in how we manage, access, and share information. The schism between the document-centric and data-centric camps actually highlights the source of what will be XML's most lasting contribution: its ability to bridge the worlds of structured and unstructured data.

Today less than 10 percent of the world's information is managed as the valuable resource that it is, and most of that is traditional "rows and columns" structured data. Being able to manage - to capture, store, index, search, analyze, and share - more of the remaining data will be a Very Big Deal. XML will make this possible by enabling the conversion of unstructured data into semi-structured data that can then be actively managed, both on its own and in hybrid applications that require structured and unstructured information. The canonical but prosaic example of catalog information, which combines structured data such as part numbers and prices with unstructured content such as images and text, only hints at the potential for knowledge and efficiency when we have information systems that can truly manage all our data.

When unstructured information becomes a managed resource, it can be integrated into more organizational processes. It can be analyzed with data mining and business intelligence tools. Users can search across information that was previously stored in silos, such as file systems, document repositories, Web sites, and e-mail. Automating collaborative processes becomes easier. And capturing information as it is created becomes the norm. Consider the enormous amount of business information that resides in spreadsheets today. As these documents and the tools that create them become XML-enabled, it will be easier to capture that data in repositories where it can be made immediately available to applications for processing.

XML's flexibility and applicability to a wide range of data management problems allow the industry to progress toward convergence on several fronts, without the need for a "big bang" paradigm shift. Document-intensive industries are benefiting from standardizing their document formats with XML, and Schema continues to gain momentum as the preferred way of defining the structure of complex documents. Tool vendors are XML-enabling content-creation tools to make it easier to capture information in content repositories. Server vendors are XML-enabling other infrastructure software to make it easier to share and repurpose XML-based information. These incremental steps toward convergence represent an organic process that is more likely to happen in the real world, where investments in new technology must deliver immediate tangible business benefits.

XML standards and XML-enabled tools will not, by themselves, turn data into information and eliminate the problems caused by silos of information stored in different formats. To realize the full benefits of convergence, organizations need to rethink how they store and manage information. An architecture based on fewer integrated information repositories that provides cost-effective scalability, security, and manageability will be key.

Such a repository will need to provide support for key XML standards such as Schema and DOM, and support for navigation-based access via XPath, as well as highly optimized querying through SQL, XQuery, and SQL XML (SQL/X). To support as many tools and other clients as possible, it should provide direct access from a variety of protocols including HTTP, FTP, and WebDAV.

Because progress toward convergence is happening in an incremental way, it's possible to misinterpret these advances as less significant than they really are. But at a certain point a difference of degree becomes a difference of kind, and we are left with a very different world than the one we started with. At Oracle, we are continuing to leverage XML's capabilities in a wide range of ongoing product releases and R&D efforts, on the firm conviction that today's information systems represent only a glimpse of what will be possible tomorrow.