Re: Boeing XRI Use Cases

Hi Noah,

Dave Orchard and I went over this a bit last week off line.

I believe that you are correct that "anyURI"  are valid for name- 
spacing and that an an XRI in what we call IRI-Normal form is  
compatible with the definition of an "anyURI"

Dave Orchard still needs some convincing on our use of some of the  
characters in the authority segment, but we are now much closer in our  
understanding than a couple of weeks ago.

I think the point I was making holds true in that not all possible  
IRI's meet the definition of an "anyURI".

I believe there some issues around http: IRI where you would need some  
additional rule about PUNICODE in the Authority segment.   There may  
be issues with other schemes as well.

The XRI-TC did attempt to address this in the design of the proposed  
xri: scheme.

For the moment the position of the XRI-TC remains that Section 2.3.1  
of XRI 2.0 Syntax
covers the transformation of an XRI into a valid IRI form.for a  
registry based Authority(Sec 3.2.1 of RFC2396).

That  XLink Version 1.0 sec 5.4 defines the rule for converting the  
remaining UTF-8 characters in the IRI to %HH form.

This produces a URI that is compliant with Sec 3.2.1 of RFC2396,  and  
thus meeting the definition of an "anyURI"

I think we got the transforms correct,  and I thank you for perhaps  
putting that position forward in the past.

There was perhaps a misunderstanding at some point where some people  
thought that the XRI-TC believed that all IRI were "anyURI".

We only contend that a XRI in IRI-Normal form is a valid "anyURI".

There are lots of other political and "Information Space  
Fragmentation"  considerations why people would prefer XRI not to have  
a URI scheme.
However I and the XRI-TC remain unconvinced that there is some  
technical rule or ABNF that we violated.

If there is please let us know.

XRI 2.0 stands as an OASIS committee standard.   If there is a  
technical mistake we will issue an errata to that spec immediately.

Given the process at OASIS any of the major changes we are talking  
about will take almost a year to go through all the stages.

I will say that even if XRI become a registered scheme and accepted as  
valid "anyURI"  I don't know that I would ever encourage people to use  
them in xml name-spacing. (personal opinion)

I could however see being a valid "anyURI" as useful in X.509  
certificates, infocard and other places where the definition of  
"anyURI" is referenced.

Without that a lot of software that relies on comparing strings is  
going to have to be modified to recognize
the equivalence of HXRI forms.   most things at the moment will not  
recognize that and 
  are equivalent from a XRI syntax point of view.

I think there is still discussion to be had around the value of a  
registered scheme and what are the appropriate uses of the URI form vs  
a http: sub-scheme HXRI form.

So below I apologize for using URI as opposed to "anyURI" which was  
what I was trying to get at.

Thanks for the feedback.
I enjoy digging into this stuff:)

John Bradley

On 5-Aug-08, at 7:49 PM, wrote:

> John Bradely writes:
>> A IRI is NOT a URI,  it would be WRONG to use a IRI in an XML
>> document for name-spacing.
>> The XML specs are clear and unambiguous use a URI.
> Well, they do seem to me to be clear, but I read them differently  
> than you
> do I'm afraid ;-).  From the XML Namespaces 1.1 Recommendation [1]:
> "Abstract: XML namespaces provide a simple method for qualifying  
> element
> and attribute names used in Extensible Markup Language documents by
> associating them with namespaces identified by IRI references."
> and indeed the formal definition [2] says:
> "[Definition: An XML namespace is identified by an IRI reference
> [RFC3987]; element and attribute names may be placed in an XML  
> namespace
> using the mechanisms described in this specification. ]"
> Regarding schemas to validate such namespace names, note that XML  
> Schema
> 1.0 Datatypes anticipated IRIs by defining the xsd:anyURI type [3]  
> in a
> manner that effectively allows for IRIs (that's why there's the  
> somewhat
> confusing reference to XLink mapping rules in the 2nd paragraph);  XML
> Schema 1.1 formalizes the connection [4] now that IRIs are "official",
> stating:
> "[Definition:]   anyURI represents an Internationalized Resource
> Identifier Reference (IRI).  An anyURI value can be absolute or  
> relative,
> and may have an optional fragment identifier (i.e., it may be an IRI
> Reference).  This type should be used when the value fulfills the  
> role of
> an IRI, as defined in [RFC 3987] or its successor(s) in the IETF  
> Standards
> Track."
> When writing a schema to validate namespace-qualified content, so- 
> called
> "targetNamespace" attributes of type xsd:anyURI are used to specify  
> the
> namespaces to be matched. So, both the specifications for namespaces  
> in
> XML content, and the specifications for the schemas to validate that
> content quite explicitly allow for IRIs.
> Noah
> [1]
> [2]
> [3]
> [4]
> --------------------------------------
> Noah Mendelsohn
> IBM Corporation
> One Rogers Street
> Cambridge, MA 02142
> 1-617-693-4036
> --------------------------------------

Received on Wednesday, 6 August 2008 04:44:02 UTC