RE: odrl-ISSUE-1 (renato): Namespace for Vocab terms [ODRL Version 2.0 XML Encoding (Contributors) ]

ODRL group:

After a short private exchange with Renato I have re-opened this issue. My reasons for that are:

* We have to be aware of two slightly different interpretations of namespaces, local names, QNames and web resource identifiers:

** in the pure XML context [1] namespace and local name is a pair of values, if a namespace prefix is concatenated with a colon and a local name the results is a QName. There is not any definition that concatenating the namespace URI and the local name should make another valid URI.

** But in the context of semantic technology this is a requirement: appending the local name to the namespace URI must result in a valid URI which identifies a resource.
See one of Jeni Tennison's musings about that issue [2]

In other words: if an XML serialization is understood to be a serialization of an ontology like Dublin Core this applies:
Namespace URI = http://purl.org/dc/elements/1.1/
Namespace prefix = dc
Local name of a metadata property: creator
QName of the creator property: dc:creator
URI of the creator property: http://purl.org/dc/elements/1.1/creator
(Note: the same applies beyond XML serialization, e.g. to N3 or Turtle)

* Therefore I cannot support Renato's statement in Issue-1:
“I have been thinking that it might be better to not define an XML Namespace for the Common Vocabulary, but instead, simply define them as URIs, as they are not ever XML elements - only the values of attributes.”

** ODRL wants to define this thing: “The Asset upon which the Action is performed” and ODRL assigns this URI to it http://w3.org/ns/odrl/2/target  
When I reverse engineer what was done there: take the basic URI http://w3.org/ns/odrl/2/  and appended “target” to it.
To my feeling this is exactly what is done by defining and using a namespace.

In other words: currently it is denied that a namespace is defined for the Common Vocabulary but anybody familiar with semantic technologies would conclude from seeing URIs like http://w3.org/ns/odrl/2/target, http://w3.org/ns/odrl/2/output, http://w3.org/ns/odrl/2/payeeParty , … etc: “ah, there is an obvious namespace of http://w3.org/ns/odrl/2/ “
And I guess many adopters of ODRL will start doing this:
reuse the already defined namespace http://w3.org/ns/odrl/2 , they will used the prefix o and then you’ll see <o:action name="o:publish"/>

And they are completely right, they are not doing anything wrong, the ODRL specs allow applying a URI OR a QName OR a QCode to attributes like @name above. Further I think it is not of help to make a distinction between resources which may be a subject or a predicate or an object in an RDF triple.

* Therefore I propose that the ODRL standard defines the namespace and not the people using ODRL

* And I propose to think about this: the URIs for the common vocabularies should not be defined in the XML Encoding document but in the Controlled Vocabulary document. Only this URI makes the identifier of terms like target, output … globally unique, and this should be definitely a goal.
The same applies to the elements of the Data Model.

[1] http://www.w3.org/TR/REC-xml-names/ 
[2] http://www.jenitennison.com/blog/node/49

Thank you for your considerations

Michael


Michael Steidl
Managing Director of the IPTC [mdirector@iptc.org]
International Press Telecommunications Council 
Web: www.iptc.org - on Twitter @IPTC
Business office address: 
20 Garrick Street, London WC2E 9BT, United Kingdom
Registered in England, company no 101096




> -----Original Message-----
> From: ODRL Community Group Issue Tracker
> [mailto:sysbot+tracker@w3.org]
> Sent: Sunday, December 04, 2011 1:09 PM
> To: public-odrl-contrib@w3.org
> Subject: odrl-ISSUE-1 (renato): Namespace for Vocab terms [ODRL Version
> 2.0 XML Encoding (Contributors) ]
> 
> 
> odrl-ISSUE-1 (renato): Namespace for Vocab terms [ODRL Version 2.0 XML
> Encoding (Contributors) ]
> 
> http://www.w3.org/community/odrl/track/issues/1
> 
> Raised by: Renato Iannella
> On product: ODRL Version 2.0 XML Encoding (Contributors)
> 
> At the Barcelona meeting we agreed to have two XML Namespaces, one for
> the V2 Core Model and one for the V2 Common Vocabulary.
> 
> I have been thinking that it might be better to not define an XML Namespace
> for the Common Vocabulary, but instead, simply define them as URIs, as
> they are not ever XML elements - only the values of attributes.
> 
> It also seems that the general approach (e.g. microdata, RDF, and Linked
> Data) is to simply use URIs for everything.
> 
> For example, instead  of this:
> 
>  <o:permission>
>         <o:asset uid="http://example.com/asset:9898" relation="o:target"/>
>         <o:action name="o:publish"/>
>   </o:permission>
> 
> We would have this:
> 
>  <o:permission>
>         <o:asset uid="http://example.com/asset:9898"
> relation="http://w3.org/ns/odrl/2/target"/>
>         <o:action name="http://w3.org/ns/odrl/2/publish"/>
>   </o:permission>
> 
> (of course, the "o" would be an XML Namespace)
> 
> Any views?
> 
> 
>

Received on Thursday, 16 February 2012 18:16:18 UTC