Shibboleth Developers

["Scott Cantor" <>]

> Right, so the E-Auth SAML 1.0 standard doesn't have much meaning.  For
> all practical purposes, the base standard is SAML 1.1.

No, it's 1.0.

1.1 is in no way wire compatible if you validate, thus the source of much
pain for us until Walter's filter idea solved the problem for unsigned
content anyway. The reengineering to support both at once was not
attractive.

> > and as far as I know there
> > is no installed base of fed e-auth sites to be interoperable with ...
> 
> But there will be...

Running 1.0? Maybe, but I doubt it.

> True, but as far as I can tell there is only one significant change
> moving from 1.0 to 1.1 (caching) and it's an addition to 1.1.  Most of
> the remaining differences are deprecated elements and such.  Really
> quite minor.

Nope, the ID attributes completely changed and a 1.0 document isn't valid
anymore. The incompatibility was a major problem, and in retrospect, we
shouldn't have done it, IMHO.

-- Scott