Lists Home |
Date Index |
On Tue, Nov 09, 2004 at 05:19:01PM -0500, Jonathan Robie wrote:
> I also think most programmers prefer a keyword syntax to an XML-based
> syntax for everyday programming.
<nounphrase>types of discourse</nounphrase>
is not the most comfortable
Stephen Pemberton has repeatedly raised the idea that sometimes it's
be more useful to define a mapping (e.g. as we're doing for
XQuery <-> XQueryX) so that people can work in a comfortable idiom
but can then join the happily family of XML users in a group cuddle,
using XML tools to manipulate their data / progams, and then
rewrite them if and as needed.
I think a good example might be the idea of an XML representation of
CSS -- one could use client-side XSLT in a Web browser to manipulate
a style sheet.
But in the comparison between XQuery and XSLT people are comparing the
non-XML textual representation of XQuery with XSLT.
I think it's better to emphasise that XQuery is designed for efficient
optimization to be possible for a different set of use cases than XSLT.
This applies both to human users -- some tasks are easier in one than
another -- and for implementations.
I don't want to use XSLT to identify all REPAIR-SEQUENCE elements with
a STEP element having a PREREQUISITE of part-no 1976 in a dozen
150,000-page aircraft manuals. But this operation might be almost
instant with an XML Query implementation, and would be very easy.
(there are already multi-document implementations that use indexes
as Jonathan of course is aware, but many other xml-dev readers might
not know this)
Liam Quin, W3C XML Activity Lead, http://www.w3.org/People/Quin/