XTech 2005

Volume 8, Issue 81; 27 May 2005; last modified 08 Oct 2010

Some thoughts on the XTech conference, '05 edition.

First off, my heartiest congratulations to Edd and the rest of the program committee for organizing a great conference. It'll definitely be on my calendar again next year.

I didn't try to blog the conference while I was there because there are a sea of smart, interesting people and so many interesting conversations that it always takes me a few hours to organize my thoughts. These things tend to be 7am to midnight affairs by the time you've had a couple of drinks and dinner with some subset of aforementioned smart, interesting people. The hours only come later.

If there was a pervasive meme at the conference, I think it was the XML processing model. A half-dozen or more sessions were about, or relied on, some sort of pipeline language. I didn't see them all, but Henry Thompson talked about MT PipelineAs I write this, I'm travelling northbound from London, bound for Norwich. Without connectivity, I'm not going to be able to provide links to the topics or products I mention. Check out the XTech website, you'll find everything there. Leave me a comment if you can't find something and I'll dig up the URI., now a freely available product, that implements a full suite of pipeline technologies. Thomas Pitner discussed Improving Document Reusability with Adaptive XML Inclusions. Erik Bruchez described the Orbeon PresentationServer. And Jeni Tennison capped it all off with a thorough and balanced presentation about Managing Complex Document Generation through Pipelining.

My tour through the twisty maze of talks, all interesting, went something like this:

First, both opening keynotes get high marks. The average opening keynote doesn't move me, but Paula Le Dieu from Creative Commons International discussed the BBC's plans to publish an open archive of their content and the licensing issues that are involved. If I had three arms, the BBC would get three thumbs up for their outstanding and forward-thinking decision to open up their archives. We can only wish that the massive media conglomerates in the United States could be so couragous. You give knighthood's in England for public service, right? Someone find out who did all the hard work that I'm sure was required to reach that decision and nominate him or her for a knighthood. And keep doing so until they receive it.

Mike Shaver from the Mozilla Foundation suffered some AV difficulties but carried on with aplomb, giving a great keynote on the future of web technologies. His plea for “little bangs” really resonates with me.

After the keynotes, Michael Kay presented a comparison of XSLT 2.0 and XML Query, giving good advice about how to think about them and how they solve different kinds of problems. Patrick Chanezon presented ROME, an attempt to provide a uniform Java API to the various flavors of syndication feeds. Leigh Dodds talked about Connecting Social Content Services Using FOAF, RDF and REST, highlighting along the way some interesting issues about API keys and the amount of ambiguity over licensing of the results of querying services. The presentation dieties did not smile on Uche Ogbuji. He struggled mightily to get his presentation of Matching Python Idioms to XML Idioms off the ground. In vain, as it turned out (AV problems were all too common for a modern conference center), but he was generous enough try again the next day and it was worth it (thanks Uche!). I'm slowly becoming a Python convert and some of the excellent work that he's done at Fourthought to provide Python access to standard XML in ways that feel natural in Python is part of the appeal.

And that was all on the first day. I started my second day by listening to Michael Day present his case for why CSS is Better than XSL for Printing. Michael has already presented his case in public, and I've replied. I don't think anything new was presented, and I wasn't much pleased by the tone of his presentation, but we shook hands and had a couple of productive conversations I think. There’s lots of room for disagreement, but lots of room for agreement too, I hope. I predict that he and I will square off again some day on this topic. Paul Prescod talked about structured authoring and wikis. He works for Blast Radius and they've already demonstrated a pretty cool bit of technology in their collaborative authoring platform (if only it ran on a platform I was willing to use). If they bring the same innovation and engineering prowess to this topic, I don't know what they'll come up with, but I bet it'll be cool stuff. Tony Graham introduced OMAR, the open source registry and repository that's used by ebXML. I keep thinking I could use one of those repository things, but I never quite manage to figure out what for, exactly. But lots of folks know why they need them and OMAR looks pretty solid. Alistair Miles introduced SKOS ,and I'll be converting my weblog taxonomy to SKOS (Update: did that on the plane ride home). Tomas Pitner talked about Improving Document Reusability with Adaptive XML Inclusions, a flexible pipeline-like mechanism for automatically including transformed resources. I regret missing Jon Trowbridge's talk about Beagle, but I rounded out the day on a slightly different note with with Kal Ahmed's presentation about his experiences Topic Mapping The Restoration—17th Century London on the Semantic Web, building a topic map of the events in Samuel Pepys diary. I need to find a reason (and time and energy) to play with topic maps some more.

On Friday, Dave Beckett described Bootstrapping RDF Applications with Redland, his RDF application framework. The next time I want to build something with an RDF angle, I'm going to try to use Redland. I've even considered porting JpegRDF to Python with Redland, just for the experience. Next, I checked out Christian Kravogel and Boris Hunter's presentation of DITA—Getting Started because I'm still trying to get my head around DITA. So far, I've concluded that they've got a notion of topics, which can be subclassed, and a framework for processing them that's based on using the (structured) value of the class attribute (rather than the element name) so that the “base” stylesheets “automatically” process the subclassed topics. And they've got this notion of building a document from a collection of topics with an external mapping file. I'm thinking I might pull together a set of stylesheets for building DocBook documents with an external mapping file. That might be useful in some contexts. I don't think there's anything you can do with DITA that you couldn't do with a DocBook customization, but they've got a different methodology so it'd be harder in some ways, just as I expect, there are some things that'd be harder in DITA. I'm an old Perl hack, there's more than one way to do it. Jeni gave her aforementioned excellent survey of, and call to action for, processing pipelines and then Liam Quin, Robin Berjon, and I participated in a Future of XML panel. Surprising discovery: there are a lot of people who don't even know XML 1.1 exists. “Lucky them,” I hear some of you thinking. Ya, well, let's just not go there again today, 'k? I regret that I missed the closing keynote. Sorry Jean. Can someone post a summary?

There was a bunch of activity on the exhibit floor, too. Justsystem are getting the XFY train into high gear. Antenna House showed off their new XSL FO features. Blast Radius were demonstrating their collaboration tools. And SyncRO Soft, the folks who make the oXygen XML editor were showing off their latest version (more about that later, it looks first class). And that's just the few that come to mind.

I'll say it again. Edd, Dan Brickley, Eric van der Vlist, Jeni, Matt Biddulph, and everyone else involved in setting up the conference: excellent job.

Comments

Norm Walsh succinctly encapsulated DITA's main concepts thusly:

**********

So far, I've concluded that they've got a notion of topics, which can be subclassed, and a framework for processing them that's based on using the (structured) value of the class attribute (rather than the element name) so that the "base" stylesheets "automatically" process the subclassed topics. And they've got this notion of building a document from a collection of topics with an external mapping file. I'm thinking I might pull together a set of stylesheets for building DocBook documents with an external mapping file. That might be useful in some contexts. I don't think there's anything you can do with DITA that you couldn't do with a DocBook customization, but they've got a different methodology so it'd be harder in some ways, just as I expect, there are some things that'd be harder in DITA.

**********

Thanks Norm. This is pretty much what I finally concluded, but only after I spending much too much time. I am reminded of some Eric Raymond statements on the bad baggage of unnecessary complexity that most commonly accompanies structured documentation frameworks. I suppose I could set out now with my red pencil and further pare down the above quoted block so that we end up with two sentences, maybe one.

Writing exposes thoughts, in all their glory, or in a new light in which we can more easily evaluate them and make corrections. I have urged Don Day to write a book detailing DITA, what it is, what it does, why someone would want to use DITA, how to use it, the details of its architecture, etc., etc.

I am still waiting for that book and hoping that Don Day takes responsibility for it. It is necessary.

—Posted by Steve Whitlatch on 22 Jul 2005 @ 07:57 UTC #