Re: naming attributes

On Tue, Jul 11, 2006 at 07:05:22PM +0200, Leif Johansson wrote:
> Note that URNs include OIDs by reference so using URNs don't preclude
> using OIDs.

There's no real good reason for using OIDs other than that they're
already used in the GSS-API so they are an easy fit.

> I guess what I am arguing for is both to include a larger namespace and
> thereby to avoid having to define attribute semantics in many cases.

That's unconvicing to me, but Martin's point about attribute identifier
self-description may be.  Think of an attribute identifier that tells
you, essentially, how to extract some element of a larger structure
delivered/authenticated by a mechanism.

Nico
--

--