- From: Michael Kay <mhk@mhk.me.uk>
- Date: Wed, 24 Mar 2004 16:00:17 -0000
- To: "'David Carlisle'" <davidc@nag.co.uk>, <dnovatchev@yahoo.com>
- Cc: <public-qt-comments@w3.org>
# I agree with Dimitre that xsl:key and xsl:id should have the
# same argument forms but I do question why this is needed (for
# id() or key()) It is true that the XSLT1 for-each hack was
# horrible but given that the
# XPath2 versions allow id() to be used within a path
# expression and index the document current at that point, what
# is gained by the new syntax. Isn't id("a",$somedocumentnode)
# the same as $somedocumentnode/id("a") in the case when
# $somedocumentnode is a single document node (and perhaps also
# when it is not?)
It's almost the same. There are probably two reasonable justifications for
the extra argument
(a) the $doc/id(@x) approach isn't obvious to the average beginner
(b) the above changes the context node for "@x" which means you need to
introduce a variable to follow cross-document references.
I'm not personally overwhelmed by the weight of these arguments but I can
live with them.
Michael Kay
Received on Wednesday, 24 March 2004 16:51:34 UTC