Take-aways from XML 2005

Well, another great XML conference has come to a close. It's always a pleasure to interact with the industry experts, and meet all the first-timers! My favorite sessions were David Megginson's closing keynote, Bob DuCharme's schema and XSLT sessions, Ken Holman's XSL case study, and Norm Walsh's XSL unit testing, and Bob Stayton's "Linking outside the box". Thursday had a heavy emphasis on XSLT, but all of the sessions were quite informative.

I learned a fair amount of new things this year. Here are just a few of my take-aways:

Cool things to follow up on:

• Piggy Bank (http://simile.mit.edu/piggy-bank/)
• Microformats (http://www.microformats.org)
• NVDL (http://www.asahi-net.or.jp/~eb2m-mrt/dsdl/)
• Ibex XSL-FO engine (http://xmlpdf.com)
• FO2HTML.xsl (http://www.renderx.com/fo2html.html)
• XSLStyle (http://www.cranesoftwrights.com/resources/xslstyle/)
• Automating XSL (http://www.snee.com/xml/xml2005/xls2xsl.zip)
• IBM DB2 Viper (http://www.ibm.com/db2/xml)
• Microformats (http://www.microformats.org )
• Unit Testing XSL: http://sourceforge.net/projects/docbook in CVS at /xsl2/tools/
• CMS ROI calculator (http://www.dakota-systems.net)
• eXtensibility Manifesto (http://extensibilitymanifesto.org)
• FlightGear (http://www.flightgear.org/) - this has nothing to do with XML, except David Megginson is working on it!

Did you know?

DocBook XSL stylesheets consist of:

• 2,500 match templates
• 800 named templates
• 250 modes
• 115 modules
• 62,000 lines
• with 485 parameters

Subversion will allow you to set up pre-commit tests, to make sure all tests pass before checking the code in.

End to end XML processing performance estimates:

• Best: XPath
• JAX-RPC
• SAX
• Worst: DOM

Security and Identity Standards in the land of XML are complementary, not overlapping!

• XML Signature – fine grained data origin authentication
• XML Encryption – fine-grained confidentiality
• XKMS – outsourced key management
• SPML – user provisioning services
• XACML – auth policy expression and evaluation
• WS-Security – end to end SOAP messaging headers
• SAML: the universal solvent for identity information

MS Linq is over my head...

Limitations of DITA: current stylesheets for FO are very limited, and not production quality. DITA is more complicated to produce output than DocBook.

Targeted search using XQuery in MarkLogic is lightning fast, and enables the creation of a variety of new outputs!

Not enough XML authoring tools provide validation against RelaxNG schemas! Good job, oXygen. Now, how 'bout the rest of you editors? Arbortext?! DocBook v5.0 is here!

I hope to see you all at XML 2006 in Seattle!

Categories: XML, xmlconf2005