OASIS eXtensible Access Control Markup Language (XACML) TC

 View Only

RE: [xacml] [glossary] Second Comments

  • 1.  RE: [xacml] [glossary] Second Comments

    Posted 10-26-2001 13:59
    Title: RE: [xacml] [glossary] Second Comments Michiharu - Thanks for the comments.  Some explanation ... I think the ISO work is less influential than it might be, because it is not as easily accessible as the OASIS and IETF works. The terms that currently appear in the XACML glossary are gleaned from the XACML charter, Ernesto's contribution on models and other important messages to the list.  So, it seems to be the set of terms that this group of experts uses.  Of course, I consulted the popular glossaries, and used their definitions where appropriate.  But, I found them overly-general, poorly-tailored to our domain, and collectively, they did not form a closely-bound, self-consistent, set. You propose a couple of new terms (SDA and clearance).  Neither of these showed up in the various contributions to XACML.  So, I suggest we add them only when (and if) we find we need them.  On the topic of clearance and label, I know that Hal has objected to their use because, conventionally, they are bound to just one access-control model (that described by Bell and Lapadula) and it is our hope to be more general. The decision made by SAML is to use "Subject" to mean the "name" of the entity.  Whereas, the term Principal refers to the entity itself. I wrestled with the terms "Requestor" and "Initiator".  I am certainly willing to be persuaded otherwise.  But, I wonder whether we will, in fact, find it important to have a term for this.  We will have to be clear that we don't mean the PEP.  Also, it is not clear in my mind that the Requestor or Initiator has to be represented in the policy language.  If it is represented, then the definition of Principal should serve. On the use of "Target", Hal has argued for its inclusion.  However, I believe his definition conflicts with the use of the term in the XACML charter.  We'll have to resolve that conflict. On the topic of "Authorization", I believe we prefer this term over "Access control", because "Access control" is closely associated in people's minds with a fixed set of actions: read, write, execute, etc..  And, we want to allow a more general set of actions.  Choosing a different term signals to the reader that something different may be intended. All the best.  Tim. ----------------------------------------- Tim Moses Tel: 613.270.3183