* Daisuke Miyakawa <
d.miy...@gmail.com> [2012-11-30 23:41]:
> I understand this is the right uri:
> - urn:oasis:names:tc:SAML:1.1:nameid-format:emailAddress
If that is what the SAML specs say (which you clearly have found and
consulted yourself), then it is. You don't need other people to tell
you what's in the spec or or.
> ..while technically SimpleSAML can handle these:
> - urn:oasis:names:tc:SAML:1.1:nameid-format:email
> - urn:oasis:names:tc:SAML:2.0:nameid-format:email
> - urn:oasis:names:tc:SAML:2.0:nameid-format:emailAddress
As I already said this might just be an error in the documentation and
metadata-template file. Some developer would need to confirm what, if
anything, SSP would do with invalid NameID formats.
> In the email thread "Help Passing Authenticated Information to Google",
> "urn:oasis:names:tc:SAML:1.1:nameid-format:email" is mentioned while
> [SAMLCore] just has
> "urn:oasis:names:tc:SAML:1.1:nameid-format:emailAddress".
> I'm not sure if the shorter one is valid in a strict manner (from the spec
> perspective).
If software behaves differently at all based on looking at this, it
will be a simple string comparison, so yes, a single byte or char
might make all the difference.
Needless to say, whatever is in the spec is "correct", as there is no
other measure for correctness. And my reference to that thread was
made only regarding the use of a SAML1.1-defined nameid format within
SAML2 protocol messages.
So I don't really understand your questions. You've found the spec and
the nameid formats it defines. You've seen mails that state that there
should be no reason you couldn't use a SAML1.1-defined nameid format
within SAML2.0 (as it makes no sense to redefine all existing nameid
formats only because a new version might be available).
So you know what you should be doing, I'd say.
-peter