XML as salvage yard (was RE: James Clark: XML versus the Web)

["Simon St.Laurent" <simonstl@simonstl.com>]

I know a lot of us here are surprised that XML 2.0 hasn't happened. 
Most of the people I know who've looked closely at the XML family of 
standards are aware that there are serious problems in the specs, though 
which pieces are the problem is one of those never-ending conversations.

This past year has been a return to markup for me. Not that I ever 
really left, but I haven't worked in the code mines for a while.  I'm 
using web technology - HTML, CSS, and JavaScript - to build mobile 
applications.  You can get a sense at:

http://htmlref.labs.oreilly.com/

We started with DocBook - everything at O'Reilly is DocBook.  I needed 
to get to HTML5, so out came the XSLT (1.0).  Yes, of course there's 
been some JSON in there, actually generated with XQuery I was happy 
someone else was willing to write.  As I move forward on related 
projects, I'm pulling out XLink and XPointer again, applying them to 
HTML5 through the wonders of JavaScript.

There's been virtually no angst over which pieces of XML mattered.  XML 
syntax has been a constant in the markup, but apart from the original 
DocBook DTDs schemas haven't mattered, and namespaces barely figured in. 
  The core of the project has been ever-evolving markup constructs that 
we'll reuse in a largely unconstrained way.

I know the XML world emerged from a standards-centric universe, but that 
approach has provided us with mixed results.  We had a burst of 
creativity through committees that worked in some ways and failed in 
others.  "SGML on the Web" was completely a failure, at least as 
originally articulated, and while XML syntax for HTML remains a common 
best practice, the revolt against that syntax has many adherents driving 
current standards.

Instead of wondering about which direction the standards could, should, 
or will go next, let's take this pile of parts and do more interesting 
things with them.  In the Web context, JavaScript has become powerful 
enough that we can begin to extend the browsers ourselves and implement 
functionality we've craved for years.

There are lots of good parts here, but I don't see the critical mass 
needed to do to XML what XML itself did to SGML.  In the meantime, 
there's plenty of good work we can without endless committees.

Thanks,
-- 
Simon St.Laurent
http://simonstl.com/