An Association News Commentary

The Future of XML
By Simon St. Laurent
XML Editor

1998 was a good year for XML, though hardly the triumph that some of us had hoped for. February saw the release of the XML 1.0 W3C Recommendation, and the W3C Recommendation for the Document Object Model (DOM) arrived in October. The foundation specifications are out - but applications have been very slow to appear. A number of key supporting specs, like XLink, XPointer, and Extensible Style Language (XSL), remain deep within the bowels of the W3C, developing slowly and not very publicly. XML has achieved tremendous levels of hype, with almost no use on the Web, and fairly little use elsewhere.

1999 should see the release of key tools that ordinary Web users can apply to XML documents. Both Microsoft and Netscape have promised support for XML and Cascading Style Sheets (CSS) in their browsers, and Microsoft is actively developing XSL support for Internet Explorer as well. How completely the browsers will support style sheets appears to be a larger question at present than how well they will support XML.

On the editor front, a number of small companies are building XML editors (a few have released, though mostly as betas), and perhaps the SGML vendor community will release their much-rumored-but-not-yet-for-sale XML editors. Macromedia is incorporating XML support into DreamWeaver 2, and hopefully other HTML editor vendors will begin providing XML support in their products. With any luck, XML+style sheets will start to move in on the Web turf that so far has belonged exclusively to HTML.

Web developers also got a major XML boost from the recent addition of XML support to Perl, the most popular language for CGI development. The continuing spread of Java servlets provides another easy way for developers to handle XML in Web contexts, applying the large number of Java XML tools available to Web-specific situations.

Another key project that should bring XML closer to the Web is the W3C's proposed rebuilding of HTML as a set of XML modules - effectively redefining HTML using XML syntax. The implications of this project are many and vast. First, XML is much less forgiving of syntax than HTML has been traditionally. End tags are mandatory; elements can't overlap; script and style tags may need some extra help. A lot of HTML people are going to be irritated when they go to work with HTML 5 (or whatever they call it) only to find that they have a lot more work to do than they did before. Second, modularizing HTML is a great way to make important parts available (like forms support), while letting developers discard other parts in favor of straight XML+style sheets. Finally, describing HTML in XML terms means that a lot of developers will encounter XML directly. Hopefully, the experience will be much more pleasant than the prior encounter with SGML.

A big question mark for 1999 is the approach the W3C takes in its standards development process. At present, the W3C is a vendor consortium, taking only as much account of user sentiments as its committee members feel like bothering with. The XML community has proven somewhat boisterous, and created at least two standards on its own through more open processes - the Simple API for XML, or SAX, and XSchema, a set of tools for describing XML documents using XML syntax. SAX has achieved wide implementation in XML parsers and software, in an area the W3C ignored, while XSchema may have more effect as an input to the new W3C schema working group.

The W3C and standards in general have gotten a big boost from the Web Standards Project, an organized group of Web developers and consumers, but how long that boost will last remains to be seen.

A number of other XML projects and areas of development look promising. XML has proven to be an important field for open source software as well as open process, and large numbers of Java, Python, Perl, and C projects are available. Document management and XML storage facilities are under development, offering new levels of control and efficiency in document processing. Standards based on XML, like MathML for mathematical notations and SMIL for multimedia, are starting to proliferate, though only in a limited number of areas so far. An enormous number of XML-based standards announcements in financial, database, scientific, document management, and other fields have been made, though most of the work is, as always, in progress.

1999 should see some of those projects to completion, and hopefully XML will get the chance it deserves to become ubiquitous. The Web lowered transaction costs significantly for a large number of tasks, making possible electronic publishing, commerce, and the use of the Web as a general-purpose interface. XML can lower those costs still further, bringing the Web into far more fields than Web browsers, creating an interconnected world of organized information

[Back to Commentary]