RE: Namespace of ODRL

Hi Renato

Could you tell us where we are in this discussion and how we will come to a
conclusion? To "look at how the single namespace looks in the first draft of
the ODRL Ontology Spec" may not be sufficient - and do we have timeline for
looking at it?

And to raise the confusion I add some comments below ;-)

Michael

> -----Original Message-----
> From: Renato Iannella [mailto:ri@semanticidentity.com]
> Sent: Thursday, August 08, 2013 6:13 AM
> To: Benjamin Hück
> Cc: public-odrl@w3.org
> Subject: Re: Namespace of ODRL
> 
> On 6 Aug 2013, at 05:28, Benjamin Hück <bhueck@uni-koblenz.de> wrote:
> 
> > we are surprised about the decision to use a single namespace for them
> ODRL Ontology.
> 
> Benjamin, I think a single namespace can still meet your requirements:
> 
> > First, using different namespaces for the basic structure and for each
> > set of vocabulary terms eases the process of managing these terms.
> 
> Yes and No ;-) Given the nature of rdf/owl ontologies, loading a single
> namespace or
> multiple namespaces into common editors (like Protege, TopBraid
> Composer, etc) makes no
> difference to the editing/managing the terms etc

A note from having a deeper look at the schema.org vocabulary -
http://schema.org/docs/full.html
- it has a hierarchical structure (yes, in know: classes and sub-classes)
which provides wrappers, e.g. Action has AchieveAction ... UpdateAction as
children - this is essential for finding out which terms in the schema.org
namespace represent an action
- and in addition for some of the wrappers all its child terms inherit the
wrapper term, see Action above, see also MedicalEntity.
- to my feeling this clearly indicates that a flat list of terms as the
current ODRL vocabulary would not work properly for the not-expert users.

> 
> > Different namespaces would help to get a very clear distinction between
> > the different vocabularies. It is not surprising that other ontologies
> > such as KAoS [1] also use separate namespaces.
> 
> There are pros and cons for both, and many similar example of ontologies
> using one namespace
> (eg schema.org, SKOS, PROV...)
> 
> > Second, the use of different namespaces reduces the possibility of name
> > clashes. For example, a term X in the action vocabulary can be
> > distinguished from a different term X in the constraint vocabulary.
> 
> There should not be cases of name clashes in our case, as for example,
> Actions will always be verbs
> and constraints Nouns.
> 
> > Third, each namespace can be used as a URL providing exactly those terms
> > associated with the namespace. For example, a namespace unifying all
> > actions can also be used as an URL to access all action terms.
> 
> This still holds true with a single namespace, as all terms would be
> subclasses/subproperties of a parent term.
> (or in our case, individuals of the same class)
> Hence, an editor can see all the relevant terms based on the expression in
> the ontology.

This implicitly requires that one MUST use an ontology editor/viewer to
understand the structure of classes in a namespace, accessing only the URL
would not help - as the URL of an IPTC vocabulary, e.g.
http://cv.iptc.org/newscodes/scene/, delivers all members.

> 
> > Finally, the use of ODRL should not get more complicated when using
> > different namespaces. Instead, different namespaces can even help the
> > process of creating ODRL policies. A policy editing tool could easily
> > distinguish between actions, constraints, and other terms based on their
> > namespaces.
> 
> See above comment.

See my above comment, to filter out all Actions from the generic ODRL
namespace needs to use an ontology viewer.
(Btw: currently I try to build a ODRL policy builder using JavaScript - a
vocabulary-specific namespace would help there.)

Michael

> 
> >  Thus, we suggest discussing the namespace modifications for a
> > second time and not just decide on it.
> 
> 
> I suggest we look at how the single namespace looks in the first draft of
the
> ODRL Ontology Spec, then we can have further discussion if required.
> 
> Cheers...
> Renato Iannella
> Semantic Identity
> http://semanticidentity.com
> Mobile: +61 4 1313 2206
>

Received on Tuesday, 13 August 2013 17:33:55 UTC