There is no such assumption. The idea is that when resource is an XML
document, it can contain (maybe partially) the context. Context is
notional. Even if it is a representation of a resource, it may include
any attributes.
Personally, I think AttribtueDesignators should not have a kind either,
as they have no other meaning other then to group attributes in the
context to some traditional categories. Maybe I should suggest for 3.0
to reconsider that and make attribute kind an extensible identifier for
association with arbitrary data sources.
D;
-----Original Message-----
From: Prakash Yamuna [mailto:]
Sent: Wednesday, March 30, 2005 6:33 PM
To: Daniel Engovatov
Cc:
Subject: Re: [xacml-users] Reg. <ResourceContent>
Hmm - interesting...
Section 6.4 states:
"The <ResourceContent> element is a notional placeholder for the
content of the resource. If an XACML policy references the contents
of the resource by means of an <AttributeSelector> element, then the
<ResourceContent> element MUST be included in the RequestContextPath
string."
The key phrase that struck me was - "notional placeholder for the
content of the resource" - hence my assumption that <ResourceContent>
can be used only for xml content representation of resources.
prakash
On Wed, 30 Mar 2005 18:16:22 -0800, Daniel Engovatov
<> wrote:
> It is explained partially in section 2.5, line 460 : you can have
> subject attribute as part of this context.
> One can include all that information within a single XML document if
so
> needed. No need for separate contents.
>
> Division of attributes into several kinds is purely for convenience.
> They do not have any fundamental differences.
>
> There is really only one context, that contains all kind of
attributes.
> See section 3.2.
>
> Daniel;
>
> -----Original Message-----
> From: Prakash Yamuna [mailto:]
> Sent: Wednesday, March 30, 2005 5:48 PM
> To:
> Subject: [xacml-users] Reg. <ResourceContent>
>
> I was wondering if anyone could provide insight on why the XACML spec
> supports <ResourceContent> but not <SubjectContent>, <ActionContent>,
> etc...
>
> thanks,
> prakash
>
>
