RE: [saml-dev] Encoding of URI in "Alternative SAML Artifact Form at"

["Mishra, Prateek" <pmishra@netegrity.com>]

>>> 
>>> UTF-8 is my encoding of choice for a possible implementation.
>>> The RFC2396 Yuri pointed out uses US-ASCII as default for 
>>URI's, so that 
>>> UTF-8 is correct for the US-ASCII subset and the best 
>>choice if other 
>>> characters might appear in URI's in the future.
>>
>>Yes, I agree that UTF-8 SHOULD be the default charset for  '
>>SourceLocation'  URI. 
>>However, if 'SourceLocation' isn't used by a destination site 
>>to specify 
>>the  source site's identity , it is not REQUIRED because the 
>>destination 
>>site just uses it  to retrieve assertion and the source site 
>>must have  
>>already known what charset is used for 'SourceLocation'.

The source and destination site must have out-of-band agreed upon a
relationship before the artifact arrives. In this sense, there is no
"specification of the source-sites identity" through the artifact. But the
specification should clearly explain the standard way characters are mapped
to bytes and vice-versa. This is where UTF-8 should be recommended.

- prateek mishra