RE: URNs, Namespaces and Registries

Those are good points.  The use of metadata in uris and constraints by
the authority to approximate myri schemes isn't discussed in perhaps
enough detail in the finding.

Cheers,
Dave 

> -----Original Message-----
> From: www-tag-request@w3.org [mailto:www-tag-request@w3.org] 
> On Behalf Of Schleiff, Marty
> Sent: Friday, August 11, 2006 11:22 AM
> To: www-tag@w3.org
> Subject: RE: URNs, Namespaces and Registries
> 
> 
> Comments on section 2.6 (Uniform access to metadata) of URNs, 
> Namespaces and Registries [1].
> 
> Perhaps the XRI notion of metadata differs from the other 
> myRIs. It certainly differs from what I've read in "The use 
> of Metadata in URIs"
> [2].
> 
> It's probably important to understand that XRI is not a 
> single type of identifier; rather, it's more of a framework 
> within which other kinds of identifiers can be expressed. 
> OIDs, IP address, distinguishedName, UUID, HIT, identifiers 
> that are case sensitive, identifiers that are case 
> insensitive, numeric identifiers, and others can all be 
> expressed within the XRI framework. Now let's move on to 
> XRI's notion of metadata. 
> 
> XRI metadata consists of tags/indicators/data about the 
> _identifier_ instead of data about the named resource. 
> Identifier metadata informs XRI-aware applications about 
> characteristics of the identifier.
> Following are examples of why identifier metadata might be usefull:
> 
> To inform the application about normalization and matching 
> rules for an identifier expressed in an XRI. Simple string 
> matching would not recognize that the following two DNs are 
> equivalent:
> 
>      cn=smith\, joe,ou=Marketing;   O=Acme; c=us
>      CN="Smith, Joe";   OU=marketing,o=acme, c=US
> 
> To inform the application of inherent features of an 
> identifier such as an embedded check digit or some crypto 
> properties like the identifier is a hash of the subject's 
> public key. For the application to derive any value from such 
> features, the presence of such features mst be conveyed to 
> the application.
> 
> To inform the application of non-http resolution capabilities 
> that might be native to the identifier (e.g., DNS, or Open 
> Group's notion of UUID pair where one UUID represents the 
> issuing authority that assigned the other UUID to a subject). 
> 
> To inform an application of how to treat an identifier like 
> "1.2.3.4" -- such an identifier in an XRI will let the 
> application know if it can ping the value as an IP address, 
> or treat it like an OID, or treat it in some other fashion.
> 
> I think the statement in URNsAndRegistries [1] that "Naming 
> authorities can impose such constraints on the http: URIs 
> under their control"  also covers XRI metadata requirements 
> if we use a naming authority like "http://xri.net" instead of 
> the "xri:" scheme.
> 
> [1] http://www.w3.org/2001/tag/doc/URNsAndRegistries-50.xml
> [2] http://www.w3.org/2001/tag/doc/metaDataInURI-31
> 
> Marty.Schleiff
> 
> 
>

Received on Monday, 14 August 2006 18:55:16 UTC