XML 2005

Volume 8, Issue 149; 20 Nov 2005; last modified 08 Oct 2010

A week in Atlanta with the XML tribe.

Lauren and the rest of the program committee pulled together another great conference. Kudos to everyone involved!

For my part, I mostly stuck to the schedule I proposed. A few highlights and random thoughts…

Things began for me on Monday when I gave my half-day DocBook tutorial. I think I'm trying to cram too much into a half-day, but I'm not sure what parts I should cut. I aim to spend about one-third of the time introducing DocBook markup (because you have to know that), one-third talking about stylesheet customizations (because you know you're going to want to tweak them) and one-third talking about schema customizations (because lots of folks want to make DocBook simpler in some ways or extend it in some ways).

On Tuesday, Bob Stayton did a great job introducing OLink; Bob Glushko spoke enthusiastically about combining document analysis and data modeling to close the “data/document divide”; Microsoft's Erik Meijer presented their integration of XML and query into the CLR (John Cowan sat in the front row and asked all the hard questions, mostly he got good answers, though some, like support for mixed content, were incomplete); finally, Henry Thompson talked about names, naming, and XML versioning, summarizing a lot of the current TAG ruminations.

Wednesday's technical sessions started with Bob DuCharme and Tim Bray in a double header about schema languages. Bob encouraged us to use (or extend) industry standard schemas instead of inventing our own and Tim exhorted us not to invent new languages. Good stuff. After lunch, Michael Sperberg-McQueen told us how schema validation isn't like being married (read “pregnant”). Michael argued persuasively that there's value in being able to get more information from your schema validation process than a simple “yes or no” answer. Dan Connolly’s late-breaking talk about semantic web calendaring was excellent. Jim Melton overcame a hard disk catastrophe to give an unexpected presentation about SPARQL; and Tony Coates described his experiences designing a system of several hundred (!) different, but related schemas. I hope we hear more about that at some future conference.

Thursday began with Marc Hadley’s overview of his Web Application Description Language (WADL). WADL is exceptionally cool. It's simple, it's clean, it makes good sense, it supports REST web services (which also make good sense) and RELAX NG. And Marc's actually implemented all of the infrastructure to build applications with it. I can't say enough about how cool I think it is. I took the stage again before lunch, talking about unit testing in XSLT 2.0, followed by Bob describing how to build stylesheets from spreadsheets (which is cooler than it sounds). After lunch, Sam Ruby cataloged some of the pitfalls that can arise in XML processing. His list included some obvious user errors, some obvious programmer errors, some examples of variation specifically allowed by the relevant specifications, and some problems that are only marginally related to XML. Overall, I enjoyed it, though I'm not sure that his point came through very clearly.

David Megginson’s closing keynote, Everyone's using XML, but does anyone care? was excellent. His conclusion, yes, they'll always care, is good news for XML hacks like myself. I hope he's right.

This conference marked the end of an era, at the end of the banquet, Lauren announced that she's stepping down as program chair. Lauren has done an outstanding job, raising the conference to new heights. David wrote an elegant tribute to Lauren, what he neglected to mention is that he's agreed to chair the conference next year.

Good luck, David! If there's anything I can do to help, don't hesitate to ask. And see you all in Seattle next year!