What is the rationale for supporting functionality beyond the Infoset: e.g. documents with empty content, with multiple element children, etc ? As Michael Kay has said, backward compatibility concerns should mean that removing this functionality is not even under consideration, however this feature is not only there for backward compatibility. It is a useful (and much used) feature. There is a requirement on XSLT (and any reasonable XML transformation language) to be able to construct external parsed entities as well as complete documents. Xpath models these using the same model of a root node (Xpath 1) (Document node in Xpath 2) but without the constraint that there need be exactly one element child, and it similarly merges concepts of an xml declaration on a document and a text declaration on an external parsed entity, and models them both in the same way ie, doesn't model them at all in the data model and generates them based on the same parameters to the serialisation. 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 Tuesday, 17 February 2004 09:43:09 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:14:05 GMT