Profile is not permitted to define failure sub-property

[nitmws]

The section about Rule Class in the Information Model (IM) defines for the failure property:

A Rule MAY have none, one or many failure sub-property values of type Rule.

The IM currently defines two failure sub-properties: remedy, consequence.

But the ODRL Profile Mechanism does not permit to define a sub-property of failure (by not listing creating failure sub-properties this section).
This appears to be an overly strict constraint in particular as for the semantically quite similar function property of the Rule Class sub-properties may be defined by a profile.

[vroddon]

The only problem is that changing this would not be strictly an editorial change, I am afraid. El 05/03/2018 a las 14:46, Michael Steidl escribió:

…

The section aboutRule Class in the Information Model <https://www.w3.org/TR/odrl-model/#rule> (IM) defines for the failure property: A Rule MAY have none, one or many |failure|sub-property values of type Rule. The IM currently defines two failure sub-properties: remedy, consequence. But the ODRL Profile Mechanism <https://www.w3.org/TR/odrl-model/#profile-mechanism> does not permit to define a sub-property of failure (by not listing creating failure sub-properties this section). This appears to be an overly strict constraint in particular as for the semantically quite similar function property of the Rule Class sub-properties may be defined by a profile. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#300>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFLs5TEQvj2lmAzFcgE7-ApXi5gylOWIks5tbUHHgaJpZM4ScT52>.

-- Víctor Rodríguez-Doncel D3205 - Ontology Engineering Group (OEG) Departamento de Inteligencia Artificial ETS de Ingenieros Informáticos Universidad Politécnica de Madrid Campus de Montegancedo s/n Boadilla del Monte-28660 Madrid, Spain Tel. (+34) 910672914 Skype: vroddon3

[nitmws]

@vroddon at the call on 5 March it was agreed that I should write this down as an official issue.
Then it will be reviewed and either accepted or added to a list of "pending next ODRL version issues".

[simonstey]

The IM currently defines two failure sub-properties: remedy, consequence.

+ duty

[riannella]

The first para before the Table should have read:

"To create an ODRL Profile, direct extensions to the ODRL Core Vocabulary classes, properties, and instances are defined. The table below shows some of of the typical extensions:"

[nitmws]

@riannella what exactly is a "direct extension"? This must be explained in this para. Could it be:

"To create an ODRL Profile, the ODRL Core Vocabulary classes are extended by sub-classes, the properties are extended by sub-properties, and instances of both are defined. The table below shows some typical extensions:"

This leaves open: may new non-sub-property properties be added?

And another question: may a new (sub-)property be defined in the ODRL namespace? A quick check did not show any prohibition of that.

[riannella]

Given the IM is based on the open-world RDF/OWL model, anyone can define a property with domain and range from ODRL and use it in a Profile.

No, only W3C can now add terms to the https://www.w3.org/ns/odrl/2/ namespace

[nitmws]

Ok, then I suggest this definition of the Profile Mechanism using a more verbose but as I think also clearer wording:

To create an ODRL Profile, the ODRL Core Vocabulary classes can be extended by sub-classes with optional additional properties, the Core Vocabulary properties can be extended by sub-properties, further instances of classes and properties defined by the Core Vocabulary or this Profile can be defined. The table below shows some typical extensions, a new Profile is not limited to them:

What about these phrases in the Profile Definition examples:

• ... and define it as disjoint with all other Policy subclasses (except Set).
• ... and define it as disjoint with all other Rule subclasses.

Is that only a hint or a normative requirement?

[riannella]

In section 2.6 we do say:
" A Rule MAY have none, one or many failure sub-property values of type Rule"

So, it would be reasonably expected that you could define your own failure sub-property.

[riannella]

Summary: The first para of Section 3.3 should be more clearer as "To create an ODRL Profile, the ODRL Core Vocabulary classes can be extended by sub-classes with optional additional properties, the Core Vocabulary properties can be extended by sub-properties, further instances of classes and properties defined by the Core Vocabulary or this Profile can be defined. The table below shows some typical extensions, a new Profile is not limited to them:"