xacml — archive

RE: [xacml] XACML Core Spec

From: Polar Humenn <[email protected]>
To: [email protected]
Date: Fri, 23 Apr 2004 15:46:00 -0000

 MHonArc v2.5.0b2 -->

xacml message

[Date Prev]
 | [Thread Prev]
 | [Thread Next]
 | [Date Next]

--

[Date Index]
 | [Thread Index]
 | [List Home]

Subject: RE: [xacml] XACML Core Spec

From: Polar Humenn <[email protected]>
To: Tim Moses <[email protected]>
Date: Fri, 23 Apr 2004 11:59:02 -0400 (EDT)

On Fri, 23 Apr 2004, Tim Moses wrote:

> Polar - We have RuleIdRef defined as a string and RuleId defined as a URI.
> Is this intentional?  All the best.  Tim.

You mean in the RuleCombinerParameter?

This must be historical. RuleId's were URI's right? I think this approach
just comes from the PolicyId or PolicySetIds being URIs. But they are URIS
because they can be externally referenced in a PolicySet (god only knows
how).

We don't have RuleReferences, (and we shouldn't), so a rule really didn't
even need a RuleId other than having a descriptive name.

Now a RuleCombinerParameter can reference it from only within the Policy,
much like how VariableDefintions are referenced.. Therefore, the RuleID
really shouldn't be more than a string.

I would like to see the RuleId be a string, so that less complicated
string-equality can be used to deal with these RuleCombinerParameters.

We wouldn't be loosing anything, since there was nothing you would do with
a RuleId as a URI in the first place.

Cheers,
-Polar

>