RE: [xacml] proposal for action-id, subject-id,and required Attr ibutes

[Daniel Engovatov <dengovatov@crosslogix.com>]

I would second that.  There is no need to standartize any attributes - XACML
system may be
used in cases where something we mandate will make no sense at all.  Having
an ID attribute
for "subject" is not critical to anything - especially since we do not
guarantee, or even
suggest that it is unique.  
It is not harder to agree on the name of an attribute used for several
interoperating systems,
then to try to specify some scheme, of what this attribute are supposed to
mean, for all systems
using XACML.
Daniel.


-----Original Message-----
From: Anne Anderson [mailto:Anne.Anderson@Sun.com]
Sent: Tuesday, September 03, 2002 10:41 AM
To: XACML TC
Subject: Re: [xacml] proposal for action-id, subject-id, and required
Attributes


On 3 September, Polar Humenn writes: Re: [xacml] proposal for action-id,
subject-id, and required Attributes
 > Anne, again, just requiring the Subject to have "one other" attribute,
 > which is "typically" the "subject-id", doesn't really buy you much.
 > 
 > I suggest that we should take a stance of being consistent across all
 > subjects, actions, and resources. For each, either that no attributes are
 > guarranteed to be there, or they all have at least one attribute
available
 > and we know explicitly what that attribute is.

I vote for "no attributes are guaranteed to be there".  I do not
care whether we make minOccurs=0 or 1, although I think we can
eliminate some meaningless cases by requiring at least 1, even if
we do not specify which one that is.

Anne
-- 
Anne H. Anderson             Email: Anne.Anderson@Sun.COM
Sun Microsystems Laboratories
1 Network Drive,UBUR02-311     Tel: 781/442-0928
Burlington, MA 01803-0902 USA  Fax: 781/442-1692


----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>