Re: [xml-dev] RE: James Clark: XML versus the Web
Doug,
Actually, one of the facets I've always found fascinating is that most of the @type attributes tend to be overwhelmingly dominated by a single language. In theory, there's nothing wrong with saying
<?xml-stylesheet type="text/javascript" href="myTransform.js"?>
<script type="text/xquery"/>
<style type="application/xslt+xml">
etc., but in practice, of course, such extensibility is so foreign to most web browsers that I suspect most browsers don't actually even have stub code for any alternatives (and please don't ding me on wrong mimetypes here, just trying to get the idea across).
XQuery actually works surprisingly well in a script context (take a look at XQIB as at least partial verification of this - http://www.xqib.org) but because there is no modularization of the scripting engine in a browser, it will always remain a very secondary technology except for the few XQuery geeks out there. There is a move to create XQIB using Javascript as the foundational language, and I suspect that this very well may be the only real route for such support moving forward.
Perhaps this is one of those areas that the W3C should push back on to the HTML5 group (which is largely synonymous with the vendors) - providing a way to extend the HTML5 model to better support modularization of functional capabilities, even if only through a consistent Javascript bridge (in the same way that you can use Java or C# as the VM for Python, Ruby or Perl.