Shibboleth Developers

[Chad La Joie <>]

Well, the real question is what data is needed to make the decision and 
where is it kept?  Should the decoders/encoders, for example, know 
whether the binding is back-channel or front-channel?  If so, are there 
are other possibilities in other protocols that we should take in to 
account?

On 6/23/11 7:21 PM, Cantor, Scott E. wrote:
> On 6/23/11 4:29 PM, "Chad La 
> Joie"<>
>   wrote:
>
> > Yeah, I'm iterating through a lot of ideas on this right now.  So far,
> > I've considered three possibilities:
> >   - leave it as is
> >   - Double the properties we have today and make one set for
> > front-channel and the other set for back-channel.  However, this means
> > the IdP will need to directly understand those concept (it doesn't today).
> >   - Have a pluggable strategy.
>
> I definitely favor the last one, even if it's rarely used, it just avoids
> arguing over it. Assuming our new profile context objects are sufficiently
> easy to use as a generic input to the decision process anyway.
>
> -- Scott

--
Chad La Joie
http://itumi.biz
trusted identities, delivered