Shibboleth Developers

["RL 'Bob' Morgan" <>]

On Thu, 12 Jun 2003, Keith Hazelton wrote:

> The reason I think it's ok to put things like cn and mail under the
> eduperson arc is because the eduPerson spec points to the relevant
> defining rfcs or to x.520 if there's no rfc, for each of these person,
> orgPerson and inetOrgPerson attributes. The eduPerson spec even adds
> some notes on usage that amount to profiling their use in higher ed
> within the bounds of the x.520/rfc definitions.

Thinking about all this makes me think, contrary to where I started out,
that basing any of this naming on documents is just a bad idea.
Documents are revised.  They only include some stuff of interest, not
other stuff.  Some stuff you want isn't in any particular document (yet).
In this case, if eduPerson doesn't cover some stuff but eduOrg does, do we
have to put some defs under urn:mace:dir:eduorg#foo?  This is the sort of
hand-wringing it would great to avoid.  We're going to have to live with
this stuff for a long time, which is why I'm pressing for a scheme that
requires the least justification and the least process-churn going
forward.  Which is why I (now) recommend just having a bucket where these
named definitions go, without adding other dependencies to it.

 - RL "Bob"


------------------------------------------------------mace-shib-design-+
For list utilities, archives, subscribe, unsubscribe, etc. please visit the
ListProc web interface at 

    http://archives.internet2.edu/

------------------------------------------------------mace-shib-design--