Re: Interpretation of URIs

On Monday 21 October 2002 02:52 pm, Hallam-Baker, Phillip wrote:
> 	Question, who defines the semantics of a URI?

The confluence of the parts of the URI. The scheme tells you something, if a 
URN the registrar tells you something, if URL the use of DNS tells you 
something, the host tells you something, and the path might tell you 
something, etc. At the W3C, we like "self-describing" URIs -- by 
derefencing it you get a description of the meaning.

My goal with comments on XKMS is to *clarify and standardize* what was 
present in XKMS 1.1 and to avoid new features/complexity where possible -- 
unless its really compelling in our requirements and supported by 
prototypes. One of the clarifications I've sought is the meaning of 
"validate" which invites discussion of the (previously) implicit policy. 
The policy could be bound to the service URI, or to another (policy) URI 
(so one can use the same service under different policies) and those are 
good, clarifying things. But I don't want to get into the meanings of the 
policies, URL versus URNs, nor how to combine multiple policies them under 
some logic <smile/>.

> 	I.E. if we have an XML schema bound to a URI who gets to say what it
> means?

I'm not sure what you mean with respect to the schema as the namespace URI 
might be different from the schemaLocation URI...?

Received on Wednesday, 30 October 2002 15:30:28 UTC