How are publisher and contact point modelled?

4 years ago

Issue

The modelling difference between publisher and contact point made by DCAT and DCAT-AP may be hard to understand for implementers.

The previous discussions regarding this issue are available here.

Current situation

The base recommendation DCAT and DCAT-AP make a distinction on the modeliing level between a publisher and a contact point. The publisher is modelled as an Agent, i.e. an organisation as a real-world entity identified by its URI, while the contact point is modelled as a vCard, i.e. information about the entity.

DCAT-AP v1.1 defines:

Publisher: the real-world entity of class foaf:Agent. An entity (organisation) responsible for making the Dataset available.

Contact point: contact information that can be used for sending comments about the Dataset. The contact information of class vCard:Kind associated to an organisation, not the organisation itself.

Recommendation

The way that DCAT and DCAT-AP distinguish between the publisher (the organisation that makes the catalogue or dataset available) and contact information (address where more information can be requested or feedback can be given) is a continuing source of confusion. It is important to differentiate and provide the two types of information: Publisher is necessary to identify the entity and Contact point allows any person/organisation to communicate and provide feedback.

Rationale

The reason for this distinction is that, on the conceptual level, there is a difference between the semantics of the class foaf:Organization (which represents a real-world entity like the European Commission) and the semantics of the class vcard:Organization (which represents the information that could be printed on a business card). On one hand, it is sufficient to refer to the organisation to be able to state “The publisher of this dataset is the European Commission” without knowing exactly where this entity is located or what its telephone number is. On the other hand, if the objective is to communicate, it is necessary to know the exact details, e.g. e-mail address and/or telephone number, without necessarily knowing whose address or telephone number it is. Even if in practice the e-mail address on the vCard business card would be often an e-mail address associated with the organisation, it is important to distinguish between the two. This also allows for situations where the contact address is not the address of the organisation that publishes the dataset but the address of a subcontracted entity that takes care of customer support, or when the publisher has different contact information for different types of dataset it publishes.

Example

The example is based on the Nobel Prize catalogue, which is available via http://www.nobelprize.org/datasets/dcat. Some modifications were made in order to clarify the guideline.

<rdf:Description rdf:about="https://dcat-editor.com/store/17/resource/1">

    <foaf:mbox rdf:resource="mailto:technical@nobelprize.org"/>

    <rdf:type rdf:resource="http://xmlns.com/foaf/0.1/Agent"/>

    <foaf:homepage>http://www.nobelprize.org/</foaf:homepage&gt;

    <foaf:name>Nobel Media</foaf:name>

    <dcterms:type rdf:resource="http://purl.org/adms/publishertype/Company"/>

</rdf:Description>

<rdf:Description rdf:about="https://dcat-editor.com/store/17/resource/2">

    <ns3:fn>Hans Mehlin</ns3:fn>

    <rdf:type rdf:resource="http://www.w3.org/2006/vcard/ns#Individual"/>

    <ns3:hasEmail rdf:resource="mailto:hans.mehlin@nobelmedia.se"/>

    <ns3:hasTelephone rdf:nodeID="_node19gmorse5x1"/>

    <ns3:hasAddress rdf:nodeID="_node19gmorse5x2"/>

</rdf:Description>