Yadis conformance

[Dan Lyke]

On Wed, 31 May 2006 15:58:27 -0700, Jonathan Daugherty wrote:
> What I mean is: URL comparison rules for interoperability have nothing
> to do with how your conformance tester connects to an input URL.  Do
> you see the difference?

I don't see the relationship, I think mostly because I'm taking a  
blinders-on read of the spec, which says nothing about comparison.

6.2.1 says:

    Note: This Specification does not require a Relying Party Agent
    to resolve an XRI or other identifier. Some Relying Party Agents
    may work only when the Yadis ID is a URL.

I think that should be a "MAY" rather than "may", but either way, the fact  
that the conformance tester fails danlyke.pip.verisignlabs.com is an  
indication that other conforming Relying Party Agents MAY fail with that  
same Yadis ID and still be conforming.

If we do add a test to conditionally prepend http:// to the Yadis ID, then  
at the very least we need a warning which says "6.2.1 says some Relying  
Party Agents may fail this or interpret it differently" (ie: they could  
prepend "https://" and still be conforming, but this could give different  
results from prepending "http://"), or we need to put a note somewhere in  
sections 6.1 or 6.2 that explicitly states any conditionals and  
transformations necessary to convert a Yadis ID to a URL.

(And, thanks to searching for "equivalent", I think I've discovered  
another bug in the tester relative to its treatment of URI elements in the  
Yadis document. Thank you for keeping me on my toes...)

Dan