Eight Greats of XML.com 1999

December 29, 1999

Edd Dumbill

As I wrote last week, 1999 has been a terrific year for XML. It has also been—may I be so bold as to suggest—a great year for XML.com. In this article, I've gathered together some of the most interesting, controversial, and useful articles published by XML.com this year.

There can be few W3C recommendations with a higher per-word controversy ratio than the XML Namespaces. Issued in January, the recommendation has spawned some of the most involved developer discussions this year, including the infamous "XHTML 3 Namespaces" debate. While initially conceived as a mechanism for enabling names belonging to different applications to be distinguished within a document, the Namespace recommendation opened up larger issues. Those issues included the question of whether a URI associated with a namespace resolved directly to a schema defining that vocabulary associated with the namespace, or whether the namespace URI was free of any semantics. This was one of the core questions at the heart of the XHTML debate. Even the W3C's Director, Tim Berners-Lee, weighed into the discussion on the XML-DEV mailing list.

In January, XML.com ran two articles on XML Namespaces. The first, from Mark Walter, reported on the adoption of the recommendation; the second, from Tim Bray, gave a simple explanation of what the recommendation meant.

• Namespaces in XML Adopted by W3C (Mark Walter)
• XML Namespaces by Example (Tim Bray)

This year, Microsoft's Internet Explorer 5 became the first (and, at the time of writing, only) released browser with native support for displaying XML. IE5 enabled display of XML using either of the two W3C-supported style technologies, CSS or XSL (to be more accurate, an XSL similar to the Working Draft at the time the browser was released). Whether you love or hate Microsoft, there can be no doubt that the inclusion of XML support in IE5 enabled many developers to find immediate real-life uses for XML, and encouraged much experimentation with the technology.

The fact that the XSL support was not updated as the Working Draft evolved was a bugbear to many developers who feared that an "MS-XSL" might become a de facto standard before the final XSL recommendation was issued. Microsoft has recently confirmed its intention to support the XSLT recommendation (now that it is stable) and one hopes that the danger of a dual XSL standard is now minimized.

In March, Tim Bray took his shiny new IE5 browser for a spin, and applied the real-world test of trying to write an XML document for production on XML.com.

• XML Support in IE5 (Tim Bray)

War broke loose on XML.com in May as Michael Leventhal—launching an angry reaction to Ken Holman's exploration of XSL, published earlier in the month—lambasted XSL as the "most hideous and unwieldy language imaginable." Among other points, Leventhal suggested that XSL provided no functionality that couldn't already be achieved by the DOM or CSS.

Leventhal's article provoked lengthy debate both in mailing lists and in the XML.com forums. Although his contention that XSL "stands absolutely no chance of acceptance by the Web community" has not been borne out in reality (although one might suggest it is still too early to tell), it is true that XSL(T) has not always proved an easy fit for web developers. I suspect that from time to time the old debate will raise its head once more.

• XSL Considered Harmful (Michael Leventhal)

With vendors spitting out XML tools left, right, and center, which one should you use? One measure of a tool's quality is its conformance to the XML 1.0 specification. With interoperability high on the XML agenda, your XML processor ought to be able to read and write standard XML.

Armed with the XML conformance tests from OASIS, David Brownell embarked upon the appreciable task of assessing the XML 1.0 conformance of XML parsers. The results he found were moderately disappointing: "Few processors passed all the XMLTEST cases, much less the whole OASIS suite. The class median on this open book test was about eighty percent, which suggests that many implementors just haven't applied any real effort to getting conformance right."

Since the publication of these results, many vendors (reporting Brownell's article as being helpful) have applied more effort to enhancing the conformance of their parsers. Recently, David applied these tests to Microsoft's MSXML.DLL parser, which wasn't tested in the first article.

• Conformance Testing for XML Processors (David Brownell)
• Conformance of Microsoft's MSXML.DLL (David Brownell)

In October, XML.com took a slight diversion from its usual course and published Tim O'Reilly's keynote address to the Tokyo Linux World (given in September). O'Reilly's address covered the major issues impacting today's Web: Linux, Open Source, XML. As Dale Dougherty, XML.com's publisher, said: "I'd argue that XML itself would not make sense to anyone without this greater context in which open standards and open source have emerged."

The article itself proved very popular, drawing attention from Slashdot. O'Reilly concluded his article by saying "Open source has been one of the engines of enormous change in the computer industry, and it will be the engine of even greater change as we go forward." This statement has repeatedly been proved true in the world of XML: open source implementations have brought XML technology to a wider audience and lead the way for many new innovations.

• Where the Web Leads Us (Tim O'Reilly)

You cannot stay long in the XML world without hearing the name of James Clark. Technical lead on the original XML 1.0 specification activity, and prolific author of open source XML and SGML tools, James Clark inspires respect among the XML community. He has been called simultaneously the worst and best thing to happen to XML. Best, for his energy and provision of reliable software, and worst, because a manager or client will tend to say "Well, if James Clark can do it, why can't you?" Why not indeed? Ask anyone who's posted to XML-DEV or the XSL-List claiming to have found a bug in any of Clark's software....

Arguably, expat has been the most successful so far of Clark's tools. Its name a subtle pun on Clark's own expatriate status (he is a Briton living in Thailand), the C-based XML parser has been used to provide the XML support in many projects, including the Mozilla web browser and the PHP web scripting language.

In October, Clark Cooper presented a tutorial on using expat in applications.

• Using Expat (Clark Cooper)

Is XML, itself a great simplification of SGML, too complicated? Don Park certainly thought so. In November he proposed a simplified form of XML, dubbed SML for "Simplified (or Stupid) Markup Language," and posted it on the XML-DEV mailing list. His suggestion found great sympathy with a certain section of the community, predominantly those using XML for data-exchange (as opposed to document markup) applications.

As you might imagine, however, Park's suggestion was not without its opponents, and another long-running debate was spawned. One of the major arguments against an SML was that every application writer wants different things simplified, so there is no agreement on what such a simplified language might contain. Another concern was that a new variant on XML would impede the acceptance of XML itself.

Robert La Quey, a proponent of SML, wrote a feature for XML.com on the SML initiative. In turn, Rick Jelliffe provided a critical assessment of the SML proposals. The SML group has now created its own mailing list, SML-DEV, to pursue its aims.

• Simpletons vs. DocHeads: Simplifying XML (Robert La Quey)
• Simplified Markup Language: Your Responses (Feedback from the XML.com forums)
• Goldilocks and SML (Rick Jelliffe)

Providing a fitting end to the year, XML'99 was the best-attended XML show yet. Over 2,200 attendees traveled to Philadelphia for the four day conference, of which XML.com was a proud co-host. During the week we provided daily coverage, bringing readers a taste of the tutorials, conference sessions, and the expo.

Simon St. Laurent, Lisa Rein, and myself reported from the floor, and Digitome's Sean McGrath contributed a version of his tutorial on Python and XML. The show itself was remarkable for the earnestness and thoughtfulness of the attendees, many of whom were developers applying XML in real-world situations and sharing their experiences.

• Reports from XML'99

That's All For This Year

We hope that you've found XML.com an enjoyable and helpful resource this year. As ever, we welcome comments about the site, article proposals, criticism, and even praise! Please send them to me at edd@xml.com. Thanks for reading.

Edd Dumbill
Managing Editor, XML.com