- From: Sean M. Burke <sburke@cpan.org>
- Date: Tue, 01 Nov 2005 22:17:50 -0900
- To: xml-names-editor@w3.org
I am confused about a particular aspect of XML namespaces as defined in the "Namespaces in XML" document and its errata. Namely: I can't find anywhere in the spec (or errata) that explicitly explains how to treat unqualified attributes -- i.e., explaining how they get their namespace. That is, given this: <Stuff xmlns ="urn:Pies" xmlns:cakes ="urn:Cakes" > <a y="123/> <cakes:b z="123/> </Stuff> I am looking for some part of the spec that explains what namespaces apply to "y" and "z". I'm guessing that each of the two attributes gets its namespace from its element, so that y's namespace is a's namespace, which is "urn:Pies", and z's namespace is cakes:b's namespace, which is "urn:Cakes". But this is just my guess. The second example in section "A.3 Expanded Element Types and Attribute Names" seems, sort of, to hint that my guess is right, but it's unclear. Section 5, "Applying Namespaces to Elements and Attributes" does clearly explain namespace application to elements of all kinds, but I see nothing comparable for attributes, specifically unqualified ones. There is a sentence in section 5.2 saying "Note that default namespaces do not apply directly to attributes." -- but I see no statement as to what /does/ apply to attributes, and/or with what degree of indirection. It's very apophatic. Section "A.2 XML Namespace Partitions", insofar as I can even understand it, seems compatible with my guess, without actually confirming it in any terms I find intelligible. Can you point me in the right direction?
Received on Wednesday, 2 November 2005 07:18:10 UTC