Danny Ayers wrote:
>On Thu, 16 Dec 2004 15:05:13 -0500, Sam Ruby <[EMAIL PROTECTED]> wrote:
>
>>Danny Ayers wrote:
>>
>>>The attributes aren't namespace-qualified, yet aren't defined in the
>>>RSS 2.0 spec.
>>
>>They don't need to be defined in the RSS 2.0 spec. Semantics of
>>attributes on an element are defined by the element itself, not by the
>>containing document.
>
>
>Do you have a reference, please?
>
>What about:
>
><xx:thing yy:another="elsewhere" />
>
>Where would you expect to the semantics of yy:another to be defined?
Maybe explaining it in terms of the infoset[0] will help. We are
interested in two properties:
[namespace name]
[owner element]
<xx:thing yy:another="elsewhere" yetanother="..." />
yy:another: [namespace name]: some namespace
[owner element]: xx:thing instance
yetanother: [namespace name]:
[owner element]: xx:thing instance
No "null" namespace, just no value. It is defined only by its owner
element and local name. It's not sharing a "null" namespace with
unqualified element names.
Robert Sayre
[0] http://www.w3.org/TR/xml-infoset/#infoitem.attribute
Robert,
The spec writers of the Infoset Recommendation cleverly/dumbly omitted to define how a value was or was not determined for the [namespace name] property.
The definition of the [namespace name] property , "The namespace name, if any, of the attribute. Otherwise, this property has no value.", is essentially without any meaning as far as determining how to decide what the value of the [namespace name] property ought to be.
This tends to throw the reader back on Namespaces in XML which, as I indicated earlier, is less than totally lucid at points.
Andrew Watt
