RE: poe-ACTION-30: Can we only have a json-ld serialisation? will it impact righstml?

Let me explain how I wound up with both “assignee” and “assignee_scope” in JSON. And how both the XML and JSON representations differ from the Ontology model, as I see it.

Assignee (and assigner) are kinds of Party. They are represented in XML like this:

        <o:party uid="http://example.com/class:IT01" function="http://w3.org/ns/odrl/vocab#assignee" scope="http://w3.org/ns/odrl/vocab#group"/>

But because assignee is a particularly important kind of party, I made it a “first class” properties in JSON:

        "assignee": "http://example.com/class:IT01"

As I understand it, this is also how it is modeled in the Ontology (assignee inherits from Party):

                                odrl:assignee <http://example.com/class:IT01>

In XML, the o:party/@scope indicates what kind of party is this – in the above example, it is a “group”. So, in JSON there’s an assignee_scope property, which is equivalent to XML’s party/@scope :

        "assignee_scope": "http://www.w3.org/ns/odrl/2/group"

But in the ODRL Ontology, there is no concept of “scope”. Instead, the assignee instance is a Group:

<http://example.com/class:IT01> a odrl:Group .

In fact, for the Ontology, the possible values of “scope” in XML and JSON are instead subclasses of Party. This may well be a better way to model it, but that’s how JSON and XML have the scope property and the Ontology does not. There is no concept of inheritance in JSON. (Or namespaces or …). Instead, applications which handle JSON are meant to just “know” what the properties mean.

Regards,

Stuart


From: Ivan Herman [mailto:ivan@w3.org]
Sent: Sunday, October 09, 2016 7:17 AM
To: Dr. Renato Iannella
Cc: Myles, Stuart; W3C POE WG
Subject: Re: poe-ACTION-30: Can we only have a json-ld serialisation? will it impact righstml?


On 9 Oct 2016, at 11:56, Renato Iannella <renato.iannella@monegraph.com<mailto:renato.iannella@monegraph.com>> wrote:


On 8 Oct. 2016, at 07:35, Myles, Stuart <SMyles@ap.org<mailto:SMyles@ap.org>> wrote:

So far as I can see, this approach wouldn't require us to alter the properties that we've identified for ODRL in JSON, so long as there is a direct mapping to the ODRL ontology for each one.

I think the answer is “no” ;-)

Because the point (I think) of JSON-LD is that it uses the same property URIs as defined in the OWL Ontology.

So, for example, in the current JSON (only) encoding, we have defined the “permissions” property (where as the ontology only has “permission”)

That can be mapped in the context file. any name can be mapped onto the ontology term (ie, the URI)


Also (for example), the current JSON defines “assignee_scope” which does not exist in the ontology.


O.k. That is indeed a problem. The JSON-LD should be a faithful serialization (modulo naming of terms) of the ontology.

Ivan



Renato Iannella, Monegraph
Co-Chair, W3C Permissions & Obligations Expression (POE) Working Group



----
Ivan Herman, W3C
Digital Publishing Technical Lead
Home: http://www.w3.org/People/Ivan/

mobile: +31-641044153
ORCID ID: http://orcid.org/0000-0003-0782-2704