RE: [saml-dev] the <md:ContactPerson> element

["Scott Cantor" <cantor.2@osu.edu>]

> - If the contact is a real person, then <md:ContactPerson> works just
> fine. But what if the contact is a non-person (such as an e-mail
> list)? Was there ever any discussion of non-person contacts while the
> metadata spec was being developed by the TC?

No, but I don't see there's a major problem with that. One of the name
elements is probably unused, but it's not really a big deal. The best use of
contact information is to put it in front of technical humans anyway.

> Various federation metadata files I've looked at shoehorn non-persons
> into <md:ContactPerson> in their own unique way. Thus
> <md:ContactPerson> turns out to be fairly useless as a standard way to
> represent contact information.

That depends what you're doing with it, but how many unique ways are there?
There's SurName and GivenName. What else could you do?

> - Why is the type of <md:EmailAddress> specified to be anyURI (instead
> of an ordinary string)? I'll note that some federations use mailto:
> URIs to represent e-mail addresses and some don't, but both seem to
> validate okay.

Just a bad idea that didn't get flagged I think.

> - Why is the type of <md:ContactPerson> called md:ContactType (instead
> of md:ContactPersonType)? This inconsistent naming suggests the intent
> to reuse the type in some way.

Not every inconsistency gets caught. Virtually nobody reviewed the metadata
spec at the time that I recall.

I'm sure if we had been doing it from scratch, somebody would have suggested
using vcard or whatever.
 
-- Scott