Re: Boeing XRI Use Cases

Let me respond to your points in reverse order, as the 2nd seems to be the 
more fundamental:

> Perhaps some mention should be made in
> that it is no longer authoritative.
> I guess that makes the answer to Julian's original question.
> No is not "the namespaces 
> specification for XML 1.1",
> should be used.

No, that's not the case.  The XML Namespaces Recommendations continue to 
be the normative source on what is and is not a legal namespace in an XML 
document.  XML Schema does define an anyURI type that is intended to be 
useful to signal that a given datum is intended to be a URIs and/or IRI. 
IWe realized early that schema datatypes could provide practical 
validation rules that would in fact reject all illegal IRIs while 
accepting all correct ones.  As a trivial example, the URI specifications 
delegate to the specifications for particular schemes for syntax details, 
and we knew there was no way we wanted to put in separate rules for http, 
mailto, tel, etc.  XSD 1.0 broadly indicates that the strings were to be 
validated as URIs, but realizing that this was a fuzzy and ultimately 
impractical burden to put on processors, the rules are loosened in XSD 
1.1, which now accepts any string of legal XML characters. 

I have tried to be careful that when responding to email questions in this 
thread about legal namespaces I've referred to the Namespaces 
Recommendation, and where questions have been raised about XSD anyURI I've 
given the rules for that.  I believe I did mention once that xsd:anyURI 
can be used to validate namespaces, but I meant that only to indicate that 
IRIs are not excluded.  The normative rules for namespaces continue to be 
in the Namespaces Recommendations.

> So I conclude that contrary to
> , the recommendation 
>  is not making it into XSD 1.1?

Again, we're crossing up Namespaces Recommendations and XSD 1.1.  The 
pertinent normative reference is, which says:

"The use of relative IRI references, including same-document references, 
in namespace declarations is deprecated. "

XSD 1.1 in no way supercedes that.  It merely defines a type that can be 
used to declare one's intention that a data value be considered as an 
IRI/URI.  The validation semantics are, for better or worse, essentially a 
no-op, but those affect only what will be accepted during schema 
validation.  They in no way alter what's a legal URI, IRI or namespace 
name, or remove the deprecation of relative namespace names.


Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142

Received on Wednesday, 6 August 2008 22:07:15 UTC