Charting the XML territory with XMLMap (2/2) - exploring XML
Charting the XML territory with XMLMap
Styling of documents
XSL was chartered with the transformation of all things XML into today's most prevalent display and print formats, HTML and PDF, among others. The effort was never completed, and subsequently split into XSL Transformations (XSLT) for the more display-oriented, low resolution and low precision typesetting, and XSL Formatting Objects (XSLFO) for high-resolution, high-precision print requirements. Again the on-line community was clashing with the off-line folks, trying to mold conflicting requirements into one specification.
XSLT was quickly completed after the split, specifying how to transform XML into other forms of XML, text, or (X)HTML. An XSLT processor consists of a rules engine for finding and executing applicable formatting rules, including an XPath component for identifying nodes in the source XML document, plus a template processor for generating the new output.
XSLFO dragged on for a while but now is also available as a set of standard formatting objects, associated with various style sheets for producing PDF and the like.
XML Schemas and Validation
XML Schema was the W3C's effort to pull together a set of competing specifications for replacing DTDs with an XML-based way of defining content models. While many sources of input were honored, the result is arguably convoluted and inconsistent, with many weak points.
RELAX NG was devised to improve on most of XML Schema's weaknesses. It is now a competing specification. TREX, another schema development, was subsequently merged with RELAX into RELAX NG.
Older efforts such as Schematron and Schema for object-oriented XML (SOX) are also still alive and kicking. We have now a handful of competing definitions, fighting a battle that will be decided by tool support and user adoption, with no clear winner to date.
We can see that balkanization has already started in the fundamentals of XML, because simplified SGML DTDs were initally picked for specifying content models, and subsequently many parties felt their efforts in superceding them needed to survive. To make matters worse, the XML schema recommendation has serious issues. While a similar constellation arose in the area of vector graphics, the resulting Scalable Vector Graphics (SVG) was technically sound with broad industry support. Furthermore, it is reasonably compatible with the only widespread predecessor implementation, Vector Markup Language (VML).
We walked through the fundamentals of XML, and critically assessed the W3C's mixed success in laying a firm groundwork for all further work based on XML. We also pointed out earlier WebReference articles:
- XML:  XML - the better HTML?
- XLink:  The Foundation: XML, XSL, X-Link
- XSLT:  Inside XSL-T
- Schemas:  XML Schemas
- Weak points:  Schema Wars: XML Schema vs. RELAX NG
- RELAX NG:  Introduction to RELAX NG
In our next installment on the topic, we will continue building our XMLMap with the most important document-based XML standards used for presentation.
Produced by Michael Claßen
Created: Dec 09, 2002
Revised: Dec 09, 2002