Re: issues with skos:related and Semantic Tags

This will cease to be an issue when we revise the model to enable roles
associated with individual bodies.
Semantic tags will just be a URI with role of oa:tagging. No need for
skos:related or a special class :)

Rob

On Mon, Sep 14, 2015 at 2:50 AM, Stian Soiland-Reyes <
soiland-reyes@cs.manchester.ac.uk> wrote:

> Any named linked data resource can be (or in many way is) a
> skos:Concept, inferring the type skos:Concept does not add any further
> semantic implications (e.g. does not add any constraints or further
> interpretations). I consider SKOS a non-intrusive way to relate to
> "any kind of identified concept".
>
> SKOS can be used in conjunction with any other existing schemes (e.g.
> OWL) - it is meant exactly as a bridging technology.
>
> See http://www.w3.org/TR/skos-primer/
>
>
> I would still agree on questioning the need for the blank node and
> oa:SemanticTag - in many ways just identifying dbr:Paris *as* a
> skos:Concept should be sufficient to make its usage a 'semantic tag'.
> (In some ways the 'tag' is the annotation relating the concept with
> the annotated target resource, not the concept itself which is used
> for many things.
>
>
>
> What is the challenge with oa:SemanticTag is if it should be a
> localized resource per annotation (e.g. blank node as you say), or if
>
> In the original Open Annotation Data Model  ontolology, oa:SemanticTag
> was meant to be used directly on the concept, similar to skos:Concept:
>
> > A class assigned to the Body when it is a semantic tagging resource; a
> URI that identifies a concept, rather than an embedded string, frequently a
> term from a controlled vocabulary. It is NOT RECOMMENDED to use the URI of
> a document as a Semantic Tag, as it might also be used as a regular Body in
> other Annotations which would inherit the oa:SemanticTag class assignment.
> Instead it is more appropriate to create a new URI and link it to the
> document, using the foaf:page predicate.
>
> http://www.openannotation.org/spec/core/core.html#Tagging
>
>
> But this would mean those also become oa:Tag instances - so dbr:Paris
> would become a oa:Tag - which I think is more intrusive than saying it
> is a skos:Concept (as any dbpedia resource) that is related to a tag.
> If you then add a "value" (or cnt:chars as it was in OA) to the
> oa:Tag, you are trying to add new (possibly annotation-centric) labels
> to an existing third-party concept, which could be wrong.
>
>
> I would wonder a bit about skos:related rather than skos:closeMatch or
> even skos:exactMatch, as it is a quite a loose relation.. which leaves
> the blank oa:SemanticTag a bit meaningless.. a target resource has
> been tagged with something that is related with Paris, but we don't
> know what that something is (except hopefully it would also have a
> "value": "Paris" from oa:Tag, and as such is a ad-hoc skos:Concept).
>
>


-- 
Rob Sanderson
Information Standards Advocate
Digital Library Systems and Services
Stanford, CA 94305