Comments at end:
On Tue, 5 Nov 2002, Anne Anderson wrote:
> Polar,
> I am having trouble parsing the second paragraph Section 5.3 Complex type
> CategorizedSubjectAttributeDesignatorType (ref:
> http://lists.oasis-open.org/archives/xacml/200211/msg00066.html),
> even apart from the typos :-)
>
> I have reworded it according to my understanding. Is my
> rewording true to the intended semantics? (not necessarily
> better, just correct)
>
> ORIGINAL:
>
> A "subject" is represented by a <Subject> element of the
> <Subjects> element the <xacml-context:Request> element. A
> "categorized subject" a "subject" that contains a particular
> "subject category attribute". A "subject attribute" is an
> attribute located in a particular "subject". A "named subject
> attribute" is a "named attribute" for a "subject". A "subject
> category attribute" is the "subject attribute" that matches
> "named subject attribute" with the AttributeId of
> urn:...:subject-category and the DataType of urn:...#string. A
> "named categorized subject attribute" is a "named subject
> attribute" for a particular "categorized subject".
>
> REWORDING:
>
> A principal, or "subject", involved in making an XACML Request is
> represented by a <Subject> element of the <Subjects> element of
> the <xacml-context:Request> element. Each <Subject> element must
> contain one and only one attribute with AttributeId of
> urn:...:subject-category. This is called its "subject category
> attribute". The DataType of this attribute MUST be
> urn:...#string. No two <Subject> elements may have "subject
> category attributes" that have the same value.
I don't like referring to "subject" as a principal at this point. The
subject can be a group etc. It is also not the thingy making the XACML
reqeuest. But I do see your point about the awkward definitions in
paragraph form.
I've attached a reword of this section in both Doc & PDF files.
-Polar
