[29 May 2009; typo corrected 13 August 2009]
I spent Memorial Day weekend in a bruising, exhausting, and exhilarating three-day program committee meeting putting this program together, and I have high hopes for this year’s conference.
On Monday 10 August the week will start with a one-day International Symposium on Processing XML Efficiently: Overcoming Limits on Space, Time, or Bandwidth, chaired by Michael Kay (who knows a thing or two about efficiency). The format of a one-day symposium on a single topic of current interest has worked well in previous years for symposia on markup for overlapping structures and on versioning XML vocabularies; this one, too, promises to provide an extremely useful compressed overview of issues ranging from hardware-assisted parsing to questions of efficiency for high-level application design.
The conference proper runs from Tuesday morning through Friday midday. Papers get 45-minute slots, so there is time for a serious paper and substantial discussion. Some of the sessions are divided into tracks (it turns out not everyone in the world is fascinated by problems of overlap; those who aren’t need some other topic to listen to, sometimes), but substantial parts of the conference are plenary.
The program starts with a talk by Alex Milowski looking back over ten years of XML support in standard browsers, and looking forward to what is needed next and what can be built now with that support.
Michael Kay and Norm Walsh each have papers on pipelines in XML processing: Norm discusses issues of streamability as they relate to the new XProc spec and Michael considers the relative merits of push and pull as ways to organize pipelines (and how to smooth things out if some stages want to push and others want to pull; he suggests reviving the venerable concept of automatic program inversion to turn the control structure of a stage inside out).
Several talks deal with various aspects of markup for overlapping structures and the encoding of multi-versioned texts, with contributions describing work done in Bielefeld, Lyon, Mannheim, Bologna, and Brisbane. The group in Bielefeld are applying TEI feature-structure markup, and Fabio Vitali’s group in Bologna is pursuing the possibility of applying RDF and OWL to problems of overlap. Last December at a meeting, Fabio suggested in passing that SPARQL might be a convenient notation for querying and manipulating Goddag structures; this appears to be further work along the same lines and looks very promising to me.
Papers on test assertions and Schematron, on the XRM vocabulary for managing and checking inter-document links, a practical introduction to EXPath (a set of extension functions for XPath 2.0), a queryable compressor for XML (XSAQCT, which I assume is pronounced ‘exact’, from the redoubtable Thomas Müldner of Acadia University), case studies from the (U.S.) National Archives and Records Administration on the management of digital archival objects and from HL7 on the management of changes to a complex spec, and other topics round the program out. Allen Renear and Karen Wickett of the University of Illinois will provoke listeners with a paper arguing that documents cannot be edited; Wendell Piez of Mulberry Technologies will meditate on the nature of markup as a nomic game (you can look it up, and probably should). Peter Flynn of University College Cork will report on his investigations into XML editors and how they might be modified to make them more accessible to the kind of people who currently wouldn’t use them on a bet.
And Liam Quin, the World Wide Web Consortium’s XML Activity Lead, always ready to risk life and limb for the common good, has summoned up the courage to propose a change in the way XML deals with namespaces, despite the very real probability that anyone suggesting such a thing in public risks being torn limb from limb by a snarling mob of markup theorists shouting “Death to namespaces and their lackeys!”
Mark your calendars (10-14 August 2009, Montreal) and plan to attend; it promises to be an interesting event.