OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

xacml message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [xacml] examples in specification


Now I understood your point and agree with you.
The description of rule 3 for line 98-114 looks a little misleading.
Since Section 5.35 describes that "the values of the obligation
arguments SHALL be interpreted by the PEP", the sentence
should have been described in more unambiguous way.

It would be great help if you could post such misleading portions
you have already found in the current specification.


                      Seth Proctor                                                                                                      
                      <Seth.Proctor@Sun        To:       Michiharu Kudoh/Japan/IBM@IBMJP                                                
                      .COM>                    cc:       xacml@lists.oasis-open.org                                                     
                                               Subject:  Re: [xacml] examples in specification                                          
                      2003/10/24 00:13                                                                                                  

Hi Michiharu. I don't think I got my point across. :) Let me try again.

> I don't agree that the example in section isn't true. The
> obligation described in that rule is "email" with three arguments, an
> address in the medical record referred by a specific XPath, a text
> and subject id in the request context. These three arguments are not for
> PDP but for PEP. PDP does not have to interpret those arguments and the
> whole text string below the obligation element is sent back to PEP as a
> part of the decision. No interpretation by PDP is not required. Instead,
> PEP must understand those parameters but this kind of agreement between
> and PEP is already assumed, as described in section 5.35.

I actually didn't say that the example isn't true. What I said is:

  While this isn't illegal, the example implies something about the
  specification that isn't true

What I think the example implies (and what I've had others tell me they
see in the example) is that the PDP is supposed to recognize that there
is a Selector or Designator, do the attribute retrieval, and then fill
in the AttributeValue for the AttributeAssignment. Then what the PEP
gets back is the email address as the first parameter, not a Selector
that points to an email address. This is not behavior that is defined in
the specification, but most people I have talked with think that this is
what the example is showing. [1]

This is why I was concerend about this, and other examples. I don't
think that this example is invalid, I just think that it's misleading
people trying to use Obligations. Is that clearer?


[1] I fully understand why this funcationality would be useful, but
that's not the issue I'm trying to raise here.

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]