RE: [saml-dev] strongly matches

> This perhaps should be errata.

The weasel language was all that was agreed to during the process, if people
can agree to more now, that's fine.

I think I probably misspoke on the attributes though. I think the IdP is
free to insert or modify the conditional attributes to some degree.
Certainly so in the case of an AuthnRequest.

> > along with the other simple content,
> 
> Such as a nested <NameID> element?  If so, this, too, should 
> be errata.

NameID in this context is somewhat like KeyInfo, identifying something for
implementations, but not really in a normative way. It would be legal for
the NameID to identify the same principal in different ways, I think. At
least, that would be my opinion.

That said, it's probably not all that useful to do that, not to the same
degree as with KeyInfo.

All in all, I don't think anything but Method is very safe to assume
anything about.

-- Scott