Dxwg: dcat:Resource usage note

[agbeltran]

In the usage note for dcat:Resource, indicates "It is strongly recommended to use a more specific sub-class when available."

This is clear for the subclasses that are defined in the spec but, should we clarify that it is recommended to extend dcat:Resource when no specific sub-class is available?

If changes are made, they also need to be reflected in the dcat.ttl file.

[riccardoAlbertoni]

Should we force the user to extend dcat:Resource whenever the type of resource considered in not represented explicitly as a class?
I wonder if there is any case in which a cataloger does not want to define explicitly the subclass, and use Resource as a generic resource...

[agbeltran]

I don't think we should force the user, but I was suggesting recommending to extend dcat:Resource. I would imagine that it is always useful to know what kind of entity you are cataloguing.

When we said "It is strongly recommended to use a more specific sub-class when available.", the point is that the only available subclasses are dcat:Dataset and dcat:DataService, so I think we should add more clarifications on what to do when a sub-class is not available.

[riccardoAlbertoni]

what about the following sentence?

"It is strongly recommended to use a more specific sub-class when available. dcat:Resource SHOULD be extended with more specific sub-class when the cataloged resource is not a dcat:Dataset or dcat:DataService"

[smrgeoinfo]

Since dcat:Resource has a dct:Type property, it might be better for interoperability if the recommendation is to use dcat:Resource for other resource types if no additional properties are required to describe them, and define new sub classes if additional properties are required. Soft typing vs. hard typing....

[dr-shorthair]

Indeed - rdf:type gives you the 'hard type' with entailments, while dct:type is a classifier without consequences (... soft) ;-)

Soft-typing can be quite useful, perhaps as a first step towards developing a fully axiomatized hard-type, but maybe just as a flag for use in a particular context.

[smrgeoinfo]

You could also set up SWRL rules based on soft types...?

[agbeltran]

Then, what about

It is strongly recommended to use a more specific sub-class when available. If cataloging other resources different from dcat:Dataset or dcat:DataService, it is recommended to rely on the dcat:Resource or extend it, while including the dct:type to indicate the specific type/genre.

?

ping @riccardoAlbertoni @dr-shorthair @smrgeoinfo

[dr-shorthair]

It is strongly recommended to use a more specific sub-class. When describing a 
resource which is not a dcat:Dataset or dcat:DataService, it is recommended to create a 
suitable sub-class of dcat:Resource, or use dcat:Resource with the dct:type property to 
indicate the specific type.

[riccardoAlbertoni]

I agree that we should let the users decide between soft- and hard- typing or even both.
IMO the following @dr-shorthair's proposal is a more direct formulation than the previous

It is strongly recommended to use a more specific sub-class. When describing a 
resource which is not a dcat:Dataset or dcat:DataService, it is recommended to create a 
suitable sub-class of dcat:Resource, or use dcat:Resource with the dct:type property to 
indicate the specific type.