xmlhack: XML 2001: open and balanced

XML 2001, opened by its new chair Lauren Wood, marks the arrival of XML schema languages and a new competition between the W3C and ISO for XML standardization.

XML 2001 has been a successful start in the attempt to give new momentum to the IDEAlliance (ex GCA) conferences, by appointing new chairs Lauren Wood and Edd Dumbill (XML Europe 2002). XML 2001 was also a personal success for Lauren Wood, who able to keep the conference buoyant -- unlike a number of conferences held in the US since September 11 -- by bringing back key contributors such as James Clark and Rick Jelliffe, who were absent from previous conferences.

Among the memorable themes from the conference was the surprising persistence of the quarrel between RDF (W3C) and Topic Maps (ISO) efforts, whose convergence seems to be taking longer than expected. XML schema languages also had a significant presence in many of the presentations.

Despite the recognition of the importance of XML schema languages in XML 2000, the view this year was much more open: only six months old, RELAX NG is already being considered as a credible alternative to W3C XML Schema. One presentation even demonstrated that RELAX NG may be more generic, and can be used as a basis for a layered implementation of W3C XML Schemas.

Progress at the W3C towards specifying a processing model for XML that includes multiple operations such as validation, inclusion, transformation and querying seems to be slowing under the weight of existing specifications, allowing ISO to take the initiative. James Mason announced the start of an effort to rationalize the range of existing schema languages. This will unite a basic processing model with grammar based (RELAX NG), rule based (Schematron), and object-oriented (W3C XML Schema) languages. James Clark, Murata MAKOTO, Rick Jelliffe and Ken Holman will all be participating in this effort.

The W3C "startup," propelled by the initial success of XML, seems to have reached the point where its size and the weight of its legacy places it in the same category as the major standard organizations, and makes it therefore less competitive.

Other stories:

Why this insistence that there is a "quarrel" between RDF (W3C) and Topic Maps (ISO)? I know of no quarrel, except possibly among a small minority of bigots, who don't count anyway.

There *is* healthy debate between people who genuinely want to understand how the two relate to each other and to what degree (if any) they overlap. Lars Marius Garshol, who gave a paper at XML 2001, is one such person. My own opinion is that RDF and Topic Maps complement each other rather than compete, and that they can - and often should - be used in tandem.

The whole thing is reminiscent of the old debate among librarians about whether "document languages" (read: RDF) or "subject languages" (read: topic maps) were best. The answer finally arrived at by that community was that neither was "best" and both were necessary.

So let's have less talk of "quarrels" and more constructive contributions to an understanding of what these two specifications are really about and how they can work together! And let's not assume that it necessarily makes sense to have them "converge" in the shortest possible time frame!

(P.S. I don't want to pick on the author, Eric van der Vlist, since he is one of those that have actually used RDF and topic maps together in practice, but I am getting tired of these sensationalist claims of a conflict.)

> I've been all over the XML web sites, and never heard of "RELAX NG" -- is this a little, journalistic bias?

On the part of the web sites you've been to, yes. Most so-called tech "news" sources cover only those things that large corporations deem interesting (ie. put out a press release on). XML got its start in independent grassroots thinking, and that's where we keep our coverage rooted, as practically all the best ideas came from there.

I've been all over the XML web sites, and never heard of "RELAX NG" -- is this a little journalistic bias?

I've waited long enough for schemas, don't ask me to wait for yet another standard to crawl out of beta.

Mark