Re: [xml-dev] Global/Local attributes

[Benderjgxd@aol.com]

In a message dated 8/1/2002 8:36:12 PM Central Daylight Time, dareo@microsoft.com writes:


> What is broken is the existence of a default namespace. Further mangling
> Namespaces in XML to accomodate this mistake is unwise and borders on
> foolishness. 


I disagree. What is broken is the possibility of local naming, where any namespaces are used. Or rather, what XML Schema seems to call unqualified names.

This appears *all* the time in XML Schemas, where a locally named XML Schema attribute is intended to be part of the XML Schema namespace because the element containing is. This is just wrong. Where "xsd" has been declared as a prefix mapped to the XML Schema namespace uri, this is used for elements, but *not* for attributes, which are supposed to MAGICALLY be treated as being part of the XML Schema namespace. (Some "XML Schema" tools apparently break if "xsd" is used as a prefix for XML Schema attributes, which is even worse.)

Default namespaces are great, because, in contrast to local naming, the tags are declared *explicitly* to be part of a namespace. To be precise, what is broken is the possibility that there is any ambiguity regarding a locally named/unqualified named attribute. To me, no matter what the namespace of its parent element, if an attribute is given with no namespace prefix, it is part of the default namespace, if one has been declared, or it is *locally named*. There should be no possibility for an attribute to automatically take on the namespace of its element.

Jim