[xml-dev] Has XML run its course?

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

[No, I'm not suggesting that all of us working with XML abandon our 
projects and tools and retreat to binary formats or some such.]

A number of messages and encounters over the last year have left me 
wondering ever more deeply about XML as a project.  Pretty much every new 
XML-related draft from the W3C has made me reconsider whether XML itself 
still has any life in it, and whether the XML project of today looks 
anything like the XML project that first began five years ago.

I wasn't there at the beginning - I was a relative latecomer, getting here 
only in 1997. The initial goals laid out in the XML 1.0 specification were 
downright inspiring, though, even as the draft was, well, forbidding to 
those of us without extensive SGML experience. Those goals were motivating 
enough to drive me to read and publish on that document, and to spend a few 
years of personal professional time pretty much devoted to XML.

I can't say it's been time wasted, by any means, but I'm starting to move 
away from the three letters "XML" as they seem to have morphed into 
something with only a distant resemblance to the original XML 1.0.  While 
XML 1.0 was a brave attempt to do much more with less, and standardize on 
that less, the follow-up development has been an effort to do more with 
more, and only roughly standardize the results.  (There are lots of 
'standard' specs, it's just not clear which get used by whom when and why.)

Looking over XML, SGML, and (X)HTML, there are definitely common threads 
which provide a recipe for success.  Human-readable embedded markup seems 
to make a lot of people happy, as does the ability to edit such markup in a 
text editor if necessary.  The explicit structures of XML - those enforced 
end-tags - also seem to strike a chord with a lot of people, even some of 
those who complained initially.  Flexible hierarchical structures seem to 
make everyone except relational purists happy.

Beyond those basics, I'm not sure there's enough to keep a standards 
community together.  I don't think it's possible any longer for one person 
to keep track of everything happening in the XML space, and the disconnects 
between specifications seem to be growing wider rapidly.  Recent threads on 
SOAP suggest that SOAP is leaving behind much of XML 1.0 while embracing 
notions from XML Schema.  Interactions between XPath, XPointer, XInclude, 
and W3C XML Schema are difficult to plot on a chart, and even XLink is now 
used as a means of validating conditions across multiple documents.  (XSLT 
has already played a role in Schematron and Examplotron.)

As a result of this divergence, and despite the power of many of the tools 
described above, it seems that the original promise of a common syntax that 
could support both data and documents intelligently is fading.  The unity 
of information representation XML once suggested is proving ephemeral, 
broken substantially by the weight of the additional features labeled "XML 
whatever" piled on to it.  Obscurity is moving in rapidly, and the pace of 
specification development is slowing as the intricacies and junctures 
become more complex.

I suspect XML as 'XML' is dying.  Long live markup!

Simon St.Laurent
"Every day in every way I'm getting better and better." - Emile Coue