Shibboleth Developers

[Velpi <>]

> I know there are a number of technical considerations to this and 
> haven't completely thought through how this would actually be 
> implemented, but I wanted to bring it up anyway.  Has anyone else ever 
> had similar concerns?  Does this seem like a reasonable concern?  Does 
> this sound like something that may be more easily addressed in Shib2 
> with the planned changes to the ARP structure?

We have exactly the same concerns at K.U.Leuven. The artifact protocol requires 
an SP to have an entry in the metadata for any statement to be released, so 
maybe there can be found an easy solution in that area: if it's possible to 
force all/unauthenticated providers to use the artifact protocol then that issue 
could be solved easily in a Shib1.3 environment.


-- Velpi