[UC] Re: RDF-star in ARCO and CIDOC CRM

Peter,

I wasn't following the Use Cases work too closely and I'm having trouble finding stuff. 

I remember you making some proposals w.r.t. templates and prototypical use cases and I also seem to remember use cases mentioned in mails to this list. However I can't find a page that links to all of that. The best start seems to be https://github.com/w3c/rdf-ucr/issues, and then dig for everything else in the mail archives.  

lt also seems like use cases were often on the agenda but never actually discussed in the WG calls, at least according to the meeting minutes. 

Is that indeed the status quo or am I missing something?

Thomas


Am 22. Mai 2023 23:32:49 MESZ schrieb "Peter F. Patel-Schneider" <pfpschneider@gmail.com>:
>I would say instead that CIDOC-CRM is based on events (https://cidoc-crm.org/Entity/e5-event/version-7.1) that are situated in space-time.  An E13 Attribute Assignment (https://cidoc-crm.org/Entity/e13-attribute-assignment/version-7.1) is then an event (via E7 Activity - those events intentionally carried out by actors) that is an intentional assignment or assessment of some some value on some property on some entity.  For example an instance of E13 would be "the assessment of the current ownership of Martin Doerr’s silver cup in February 1997".
>
>So I would model the state of being married in CIDOC-CRM as an event (E5) and a ceremony of marriage as an attribute assignment (E13) that assigns married status to two people and is connected to the married state of those two people via a temporal relationship.
>
>In my opinion RDF-star would have problems using quoted triples to model CIDOC-CRM events.  Consider the marriages of Richard Burton and Elizabeth Taylor.  If one has a married-to relationship and wants to use it as a quoted triple one can end up having difficulties like two start and end dates on the quoted triple.  CIDOC-CRM gets around these problems by not requiring events to be uniquely determined by the values of P140, P141, and P177.
>
>There is a contributed use case to the WG from a project that uses CIDOC-CRM. I am still trying to determine exactly what is needed for the use case but you can see the current status of my attempt at creating a complete use case at https://github.com/w3c/rdf-ucr/wiki/RDF-Star-for-contextualizing-historical-assertions and discussion at https://github.com/w3c/rdf-ucr/issues/12.
>
>peter
>
>
>On 5/22/23 06:49, Vladimir Alexiev wrote:
>> Thomas Lörtsch wrote:
>>  > You may know ARCO [0] and its decision to provide "shortcut" relations that help users navigate the very involved and OWL-enabled data structures in the background. Such shortcut relations combined with shapes that define complex objects may well be the future of modelling on the Semantic Web. A sound facility to connect the shortcut with the underlying "full" data would be very welcome to enable such a modelling style.
>>  > [0] Valentina Anita Carriero, Aldo Gangemi, Maria Letizia Mancinelli, Ludovica Marinucci, Andrea Giovanni Nuzzolese, Valentina Presutti, Chiara Veninata:
>> ArCo: The Italian Cultural Heritage Knowledge Graph. ISWC (2) 2019: 36-52
>> 
>> Thomas, I'd like to add some details to the potential Use Case that you describe.
>> Since both reification and RDF-star are about adding extra data to relations, I think that tackling complex reification scenarios is relevant.
>> If you like, we can try to write out a explicit Use Case together?
>> 
>> 
>> I believe ArCo uses CIDOC CRM.
>> CIDOC CRM has a lot more involved reification than most ontologies.
>> The prototypical reification class is crm:E13_Attribute_Assignment:
>> https://cidoc-crm.org/Entity/e13-attribute-assignment/version-7.1 <https://cidoc-crm.org/Entity/e13-attribute-assignment/version-7.1>.
>> E13 is similar to rdf:Statement and has these props:
>> 
>>   * P140 assigned attribute to : rdf:subject
>>   * P141 assigned : rdf:object
>>   * P177 assigned property of type: rdf:predicate
>> 
>> P177 was recently added.
>> It's punned as individual of E55 Type, which is like skos:Concept:
>> "the properties defined by the CIDOC CRM also constitute instances of E55 Type themselves."
>> 
>> Furthermore, E13 has these subclasses:
>> 
>>   * E14 Condition Assessment,
>>   * E15 Identifier Assignment,
>>   * E16 Measurement,
>>   * E17 Type Assignment
>> 
>> They add specifics, for example E15 has:
>> 
>>   * P37 assigned: a sub-prop of P141
>>   * but also P38 deassigned, which is also a sub-prop of P141,
>>     but inverts the logic from positive to negative!
>> 
>> These subclasses should fix the value of P177.
>> Eg E15 should fix that P177 = P1 is identified by
>> (because P37 and P38 operate on E42 Identifier, and P1 is the incoming property of E42).
>> Although CRM has a first-logic axiomatization, it fails to fix the value of P177.
>> 
>