RE: versioning...

> If this is the case, then it should be defined as such. 
> However I wonder 
> how useful automated mechanisms really find this? (I expect 
> not at all.) 
> Plus, if you change the syntax or processing the data will 
> have a different 
> namespace anyway.

I will copy in the info from the SAML spec.

> If someone can present a scenario and definitions of what 
> these things mean 
> I might be more comfortable; otherwise it's seems to be 
> optional/advisable 
> cruft that's best left out.

This is a pretty standard approach for protocol compatibility
testing. It is already defined in HTTP.

		Phill