RE: [Serialization] IBM-SE-015: Serializing QNames

I don't think the serialization process contains any mention of a
dependency on casting: this is one of the advantages of specifying it as
a set of constraints, rather than as an algorithm. It's true that an
implementation is going to have to perform some kind of operation
analogous to casting. But the environment is very different: the
serializer always has a namespace context available for the operation,
because the QName always exists as the value of a node, not as a
free-standing atomic value. Conversion of lexical QNames to expanded
QNames and vice versa is no problem so long as the QName exists in the
context of a node on a tree, it only becomes a problem when the QName is
detached from its namespace context. So I don't think there is any
inconsistency that needs to be fixed.
 
Michael Kay

-----Original Message-----
From: public-qt-comments-request@w3.org
[mailto:public-qt-comments-request@w3.org] On Behalf Of XML Query
Sent: 03 February 2004 00:16
To: public-qt-comments@w3.org
Subject: [Serialization] IBM-SE-015: Serializing QNames



It is worth noting that the serialization process depends on the ability
to cast a QName into a string (for example, when it is found in the
value of an attribute such as xsi:type="xs:decimal"). Casting a QName
into a string is currently not defined in the Functions and Operators
document. This inconsistency between documents should be fixed. 

--Don Chamberlin