[OpenID] Canonical OpenID url form

[Martin Atkins]

Drummond Reed wrote:
> Martin's right, Peter -- XRI is one option for Unicode. But you can also use
> an internationalized domain name
> (http://en.wikipedia.org/wiki/Internationalized_domain_name) in a regular
> URL. It uses Punycode (http://en.wikipedia.org/wiki/Punycode).
> 

I hadn't thought of punycode. Certainly I think many of the existing 
implementations would struggle with unicode characters in the domain 
part of the URL. The spec doesn't really seem to say anything about this.

Should libraries be applying the mapping set out in RFC3987[1] section 
3.1 to incoming URLs? What about legacy servers that rely on their URLs 
not being encoded in UTF-8? The spec should probably say something about 
this, so that different implementations treat non-ASCII characters in an 
interoperable fashion. If it does already and I've missed it, then 
please point me to it!

Cheers,
Martin

[1] http://www.ietf.org/rfc/rfc3987