Making chunking and linking work to full effect requires a whole set of enabling technologies: servers than can process XPointers efficiently to return appropriate chunks, browsers that can cope with the XML-Linking specification to allow embedding of linked content, and authoring tools that will help site managers keep track of the data stored in these new document structures. The revolution this can make possible is probably a few years off, if the speed at which HTML tools have been developed is any indication.

Web Structures as Application Architecture

Even though Web browsers have rapidly grown into a popular interface, they haven’t yet been able to offer the kinds of services that users expect of their computers. The latest rounds of improvements to Web browsers and servers have beefed up the Web’s claim to being a universal interface for all kinds of data. Although Java arrived several years ago, Java development is finally reaching the point where applets and Java applications are capable of competing with full-blown operating system-specific applications. Dynamic HTML and the DOM have made it possible to create polished interfaces inside of a Web browser that offer considerably more functionality than form fields, drop-down boxes, and clickable buttons and images. The scripting languages, particularly JavaScript, have grown up, acquiring object-oriented extensions and other improvements along their path from form validators to interface managers. In the midst of this explosive growth, XML has appeared to clean up the messes created by six years of rapid development and provide a firm foundation for all these technologies.

XML is not likely to make the browser a better interface for creating graphics editors or video games. XML is not the solution for every problem by any means. Still, XML is a better solution for projects from CAD/CAM (where it lets users exchange information) to document management to presentations to possibly even the worlds of word processing, spreadsheets, and databases. XML’s structures are flexible enough to store relational database tables, spreadsheet information, and complex documents. Developing interfaces that will make writing XML documents as easy as reading them will take years of development and improvements in other key standards. Moving the Web from presentation engine to workhorse will require many more steps, and probably years of development. XML may find a “killer application” that moves it quickly to the fore, or it may need years of quiet infiltration.

Despite the challenges, however, XML seems likely to succeed. The SGML community has provided an initial base of applications and support, and the interest of key players like Microsoft, Netscape, and Sun promises it a bright future. So far the W3C has provided XML with a strong center, a place where these competitors can participate in discussions leading to common standards and prepare for their implementation. If XML continues to develop as it has, it should quickly find favor as the architecture that allows the Web to finally deliver on its promises of convenient, friendly, cheap, and interactive information access .