H3. describe a linkage between a region and a publication

[congruili]

Relate a term to describe a linkage between a region and a publication. Maybe make it as a subclass of dcterms:Relation. Currently, we are using http://schema.org/isRelatedTo; please inform whether is a correct use of this term.

[zednis]

@justgo129 can you provide a few examples?

One option that comes to mind is the use of dcterms:subject and/or dcterms:spatial.

[zednis]

schema:isRelatedTo is defined as a relationship between 2 products, so I do not think it is the correct property to use in this scenario.

from http://schema.rdfs.org/all.ttl

schema:isRelatedTo a rdf:Property;
    rdfs:label "Is Related to"@en;
    rdfs:comment "A pointer to another, somehow related product (or multiple products)."@en;
    rdfs:domain schema:Product;
    rdfs:range schema:Product;
    rdfs:isDefinedBy <http://schema.org/Product>;

[justgo129]

@zednis in response to the need for examples, I am finding that this may be an outdated request. I think we're using dcterms:Relation now. As it's a class in-lieu of a property, the need for clarification about a substitute term still stands. e.g. http://data.globalchange.gov/report/ucollaw-ti-2011.thtml

[congruili]

I think that dcterms:spatial makes more sense according to the definition: spatial characteristics of the resource.

[justgo129]

I'd be fine with that since it has range dcterms:Location which itself has a broad definition.

[justgo129]

"dcterms:Spatial" it is. @lic10 or @zednis please make the appropriate change, merge the change, and close this issue (#62).

[zednis]

@justgo129 are you sure we are not currently using dcterms:Relation for regions?

I found this in the RDF for http://data.globalchange.gov/report/nca1.thtml:

## Region(s):
   dcterms:Relation <http://data.globalchange.gov/region/northeast-us>;

[zednis]

I created two pull requests in gcis repository.

USGCRP/gcis/pull/189 updates dcterms:Relation to dcterms:spatial for publication->region reference

USGCRP/gcis/pull/190 updates dcterms:Relation to dcterms:relation for region->publication reference.

I see now why we requested an inverse property to dcterms:spatial. No inverse property is defined by dublin core. We can use dcterms:relation as per the open pull request or define a new property in the GCIS namespace.

[justgo129]

I would prefer if we define a new property in the gcis namespace to be related to dcterms:relation in some way. dcterms:relation seems rather vague to me.

[zednis]

Ok, how about this?

gcis:spatialOf
  a owl:ObjectProperty ;
  subPropertyOf dcterms:relation ;
  rdfs:comment "spatial characteristics for the resource" ;
  owl:inverseOf dcterms:spatial .

If there is consensus to add this add it to the ontology and update USGCRP/gcis#190 to use this.

[justgo129]

I really like it. Do you mean "of the resource?" +1 on making that update to USGCRP/gcis#190

[zednis]

@justgo129 No, the definition of dcterms:spatial is "Spatial characteristics of the resource." In that case "of the resource" implies this (the subject in the RDF statement).

I was hoping to have a similarly worded definition for the reverse association. My first pass may not have been clear enough though.

[justgo129]

I think I get it now. Yep, could we clarify the proposed definition?

[zednis]

"resource with these spatial characteristics"?

[congruili]

looks good enough

[justgo129]

"resource with contents having these spatial characteristics?" "spatial characteristics of resource content?"

[zednis]

To me "resource with" and "resource with contents having" mean basically the same thing...

How about "a resource related to these spatial characteristics"?

[justgo129]

Interesting, I perceive it differently. I see the resource as something physical or electronic and which contains content. The content, as opposed to the document containing it, would have the spatial characteristics under discussion (i.e. a physical resource itself doesn't have the characteristics; it's the information contained in the resource with the characteristics). Basically, I see it as the "this is not a pen" issue.

@rewolfe and @CurtTilmes I'll defer to you on this matter. This is for relating a region (subject) with spatial characteristics such as states contained in that region (object).

[zednis]

With dublin core terms the resource and the resource's content seem to be synonymous.

The definition of dct:spatial is "Spatial characteristics of the resource" not "Spatial characteristics of the resource's contents". This usage of resource as an informational resource is consistent through DCT and terminology I tried to copy in my comment on the inverse property.

[justgo129]

Good catch, @zednis. I wonder why the resource and its content are treated in dublin core as synonymous as it seems rather counter-intuitive to me given various distinctions in the Semantic Web (e.g. a URL versus the page to which it directs are separate objects in various ontologies). It would be a good discussion topic for another venue.

Back to this particular issue, @zednis are you proposing the following?: A resource: subject something="a resource related to these spatial characteristics" Predicate the spatial characteristics: object?

This would seem more fitting to https://github.com/USGCRP/gcis/pull/189 than https://github.com/USGCRP/gcis/pull/190.

Maybe rephrase to "a resource is related to these spatial characteristics?" That makes a bit more sense to me and follows some of the examples of definition formatting that already exist in the GCIS Ontology. Otherwise, +1 for going with "a resource related to these spatial characteristics" as a definition but I'd really like to hear other opinions before proceeding. @CurtTilmes @bduggan @aulenbac @rewolfe