You get the defiant flavor of the conference from its slogan: "There is nothing so practical as a good theory."  The attendees readily voice disagreement with speakers or even confess
incomprehension (with rare candor) because they care about the result.  They want it to work right. Some of the attendees were in on the cultivation of markup as a concept, so they get markup as only those who saw the bare dirt can.

The conference is a good spot to catch glimpses of new critters in the markup ecosystem, some of which may become important for DITA adopters down the road:

• The trend away from single, static, monolithic document schemas and toward processable type declarations and generated validations.
  
  For instance, DSRL (pronounced Disrule) supports aliasing for XML elements and attributes.  While DSRL provides aliasing for arbitrary vocabulary definitions, I would think DITA might benefit from structured aliasing for the DITA types.
  
• XML Processing Language (XProc)
  
  Norm Walsh's presentation was a case study in defining a solvable problem.  XProc sets out to define a processing pipeline. The potential for the DITA Open Toolkit is to integrate a set of DITA processing modules into both a platform-independent processing pipeline and a higher-value, Ant-based toolkit for managing dependencies.
  
  
• The maturation of formal semantics
  
  Steve Pepper presented what appeared to be a solid solution to RDF and Topic Map integration, conserving the insights from TopicMaps in an idiomatic RDF implementation through annotations on the classes and instances.  In HistoryNet, Sturla Flem Rinvik and Eszter Harvati are doing pragmatic work on temporal semantics. In RxPath, Adam Souzis is working on enabling the same script-like processing for RDF that XSLT offers to XML.  Good stuff.
  

If you have a taste for the exploratory, you can see for yourself through the high-quality, publicly available set of proceedings made available by the Extreme folk.  Fortunately for me.  Through the mixed blessing of VPN, the work is too much with us sometimes, and I had to miss a couple of circled timeslots.

I noticed on the final night that I had yet to

• Spend any of the Canadian currency I had converted.
• Leave a two-block radius of the conference hotel.
  

So, I took a stroll to the top of Mont Royale ("he that will reach her, about must and about must go") and found the plaza by dusk. Some photographers with tripods were set up at the balustrade for the view of Montreal as the lights came on.  The towers looked like  stacked-up boxes, and the flat arc of the St. Lawrence carried the accumulated thought downstream.