- From: David Carlisle <davidc@nag.co.uk>
- Date: Thu, 15 Jan 2004 09:59:44 GMT
- To: jim.melton@acm.org
- Cc: public-qt-comments@w3.org
I agree that making the section explictly non normative would be helpful but I cannot agree with this: Section 5 was intended, not to specify a normative way to do a trivial embedding of XQuery in an XML document, but to demonstrate that the subject has been considered and to suggest some (certainly not all!) approaches that a user might follow if a trivial embedding is required for his or her purposes. There is effectively only one way that an Xquery can be placed in XML: treat it as a string, quote all XML markup and place it in element or attribute content. Of course the actual quoting may use any one of several equivalent XML quoting syntaxes, CDTATA sections, predefined entities, or numeric character references, there is no need to elaborate those choices here. However the current section seems to indicate that unless it has a mismatched < and so directly embedding the unquoted query would produce non-well formed XML, the Query can be simply inlined without quoting into an XML document. As I tried to indicate this is a very dangerous practice that (because of different meaning for the same syntax of character references in XML and Xquery) may cause misinterpretation of the Query: what is, apparantly, a single Xquery string can become arbitrary Xquery code. The section should explictly warn against that, unless the treatment of Ӓ in XQuery can be changed to match the behaviour that would result from first parsing the Xquery as a fragment of XML. David ________________________________________________________________________ This e-mail has been scanned for all viruses by Star Internet. The service is powered by MessageLabs. For more information on a proactive anti-virus service working around the clock, around the globe, visit: http://www.star.net.uk ________________________________________________________________________
Received on Thursday, 15 January 2004 05:00:20 UTC