I love me some CardSpace, but I sure wish there was better documentation on how to use it, especially from WCF. The current documentation is very conceptual so you really only have the samples to go on, which sucks not to put too fine a point on it.
One example shows how to use WCS with WCF’s wsHttpBinding. However, it turns out you can also use WCS with WCF’s wsFederationHttpBinding. I’ve been experimenting with that binding talking to an STS, and it’s never invoked the WCS UI before. Why does this example invoke WCS? I’m guessing because it doesn’t specify an issuer in the config file, but I’d need to see documentation to be sure.
Also, using the federation binding appears to be the only way to request/demand additional claims beyond the private personal identifier (aka the PPID). If you want the client’s email address, name, address, etc, you need to specify that via the claimTypeRequirements of the binding’s messagesecurity element. But that configuration isn’t valid for the wsHttp binding. Why?
Finally, the two bindings produce different results on the security token. Using wsHttp, you get three claims: RSA Identity, RDA Possess Property and PPID Possess Property. When using wsFederationHttp, you get Hash Possess Property and PPID Possess Property (plus the claims you request). What happened to the RSA claims? If you attempt to add RSA to the claimTypeRequirements, CardSpace throws an error as an invalid request. Again, why? Keith Brown recently wrote about how to use the RSA claim, so it sounds like a valuable piece of information to have. How come the federation binding doesn’t send it?