December XML Conferences and Meetings

by Rick Jelliffe

I was pleased to see that Debbie Lapeyre is presenting a tutorial session Introductory Schematron at XML 2007 in Boston next week (Dec 3-5). I've looked over the online course notes, and it looks a really good introduction. If you want a good introduction to Schematron, this is good opportunity for you. Now that XML Schemas 1.1 also has adopted a stripped down version of Schematron's assertions, it is even more worthwhile getting to know.

Other sessions that looking interesting to me include the XML/XSLT hardware sessions and Miguel de Icaza's session.

I won't be attending, because I am going to the ISO/IEC SC34 meeting in Kyoto that weekend, sponsored kindly by Allette Systems and Topologi, for WG1 issues. It seems that our recent jumping up and down made some national bodies vote, and the last set of ballots had a quorum (just). I hope we can slough off the document format single-issue people to an new working group, which UK has proposed. That would free up the grid lock.

The voting problem has meant that I have had no desire to push ahead Schematron's evolution, but there are quite a few issues that I want to raise with WG1 members (they probably cannot be discussed formally, because the WG has agenda rules to make sure that delegates get adequate backgrounding time): I want to propose that the next version of ISO Schematron has an official XSLT2 query language binding and a better library import mechanism (there are some editorial fixes too); and users have requested that ISO Schematron Validation Report Language (SVRL, which has become quite popular) should allow more information from the schema. SC34 tasked me to look at an ISO standard for ZIP: my approaches to PKWARE have not been answered, so it looks like we will have to go ahead without their involvement which would be a shame.

I also have a discussion draft for a set of abstract application profiles for office applications I want to run up the Kyoto flagpole. There is clearly a strong need for application profiles for office applications, if we want interoperability: interoperability requires both that an application provides a minimum set of features and that it warns when exporting documents with more than those features (where no graceful degradation is possible.) The need is so strong that people expected that somehow document format standards would provide it, which has lead to a lot of heated criticism that that oranges should be apples. An example of an abstract application profile might be "Unstructured Word Processor with Western and Asian text support with medium quality (styles, hyphenation, kerning, autonumbering, nested tables, hyperlinks, standard images, running headers) but no complex script support. Software reports when features more than this profile allows are used." By defining a set of features, it helps procurement specify minimum requirements and it gives software developers a series of bars which they can demonstrate their software reach when trying to win contracts. And by making the checks (or validation) ability part of the profile, it promotes interoperability.

Unfortunately, while I am out of Sydney Standards Australia has scheduled the first meeting to discuss OOXML and our response to it. ECMA has released (to national bodies) the first hundreds of responses to the national body comments, and it would be interesting to discuss these. Actually, Standards Australia has taken a very strong line that only issues we raised in our comments would be discussed (and our comments in turn had a requirement that only issues with an Australian angle made it to our comments lists, which precluded us parroting the form-letter material from an American company.)

This SC34 meeting has another angle: it marks a changing of the guard for several key players: ANSI's Dr Jim Mason has been the chair for is it 20 years now, and has brought a cool head and enormous procedural experience (he has seen a lot of hijinks over the years I imagine); BSA's Martin Bryan has been the chair of WG1 since it came out and has been particularly involved recently in several DSDL standards such as ISO DSR; Canadian Standard's Ken Holman has provided the Secretariat over the last few years, an outstanding effort in view of his multiple cancer onslaughts (successfully repelled, I am happy to say) and especially in view of the recent fast-track palava. I have i the past urged that companies that champion strategic technologies especially in fast-tracking should contribute to secretariat expenses: Microsoft and IBM are doing the wrong thing by not doing this, perhaps out of some spurious daintiness that there would be some conflict of interest (but it is not buying a standard, because NBs vote not the Secretariat!).

These three come from SGML days and have the strong user-oriented bias that makes ISO-created standards uniquely different from standards from boutique bodies where commercial companies are first-class citizens.