Agree on details how to measure supported metadata standards
More specific definition of The service provides metadata according to a metadata standard which is established (and specific) in the ESS
is needed.
Currently, we are planning to answer the following questions:
- "Does the service support metadata standards"?
- "Does the service support cross-domain metadata standards"?
- "Does the service support metadata standards which are specific for the domain of the service"?
The 1st question is easy to answer, questions 2 and 3 are a bit more difficult:
The list of metadata standards in the KH is currently being harvested from the RDA MSC. Each metadata standard should now have the property dcat:theme
, assigning it to a certain thematic area (=research subject area). They are mapped to the DFGF ontology where possible, but if not we use the controlled vocabulary from the RDA MSC (which contains some custom terms and terms from the UNESCO Thesaurus.
Based on this, we could currently answer question 2 by simply checking if one of the n4e:supportsMetadataStandard
has dcat:theme <http://rdamsc.bath.ac.uk/thesaurus/domain0>
(= multidisciplinary).
Note: ISO 19115
is currently not classified as a multidisciplinary standard in the MSC. However I think we should consider it as multidisciplinary within the KH, meaning that it can be used across all subdomain of the ESS.
The second check will be harder. Based on the current data, we could query the SPARQL endpoint regarding a certain service/repository at least one standard it supports is assigned to a research subject area, that is at the same time assigned to the service itself:
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
prefix dcat: <http://www.w3.org/ns/dcat#>
prefix dct: <http://purl.org/dc/terms/>
prefix n4e: <http://nfdi4earth.de/ontology#>
prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
SELECT ?standardTitle ?subjectLabel
WHERE {
?repository dct:title "PANGAEA"@en .
?repository dcat:theme ?theme .
?repository n4e:supportsMetadataStandard ?standard .
?standard dcat:theme ?theme .
?standard dct:title ?standardTitle .
SERVICE <https://nfdi4earth-knowledgehub.geo.tu-dresden.de/fuseki/dfgfo/sparql> {
?theme rdfs:label ?subjectLabel
}
}
However, there are some issues with this: Some services as well as metadata standards have dcat:theme dfgfo:34
, meaning that they belong to the subject "Geosciences". This is too general for our scope (should be true for all resources in the KH), therefore we need to hardcore the specific (sub)domains we are interested in for the evaluation of question 3.
According to the NFDI4Earth proposal we are interested in (meta)data from all Earth System Sciences (ESS) disciplines, which affect the 4 systems solid earth, hydrosphere, atmosphere and biosphere. This roughly maps to all subject areas which are a subclass of dfgfo:34
(Geosciences):
- dfgfo:313 "Atmospheric Science, Oceanography and Climate Research"
- dfgfo:317 "Geography"
- dfgfo:314 "Geology and Paleontology"
- dfgfo:315 "Geophysics and Geodesy"
- dfgfo:316 "Mineralogy, Petrology and Geochemistry"
- dfgfo:318 "Water Research"
Only the biosphere is not very well represented in this. We could consider adding additionally (even though they are from the biology domain):
- dfgfo:203-03 Ecology and Biodiversity of Animals and Ecosystems, Organismic Interactions
- dfgfo:202-02 Ecology and Biodiversity of Plants and Ecosystems
Additionally we have to consider, that the research subjects are organized in a hierarchical system. However, currently in the KH there is now semantic reasoing in place which makes sure that a service with e.g. dcat:theme dfgfo:317-01
(Physical Geography) also has dcat:theme dfgfo:317
, the parent concept. However, we can solve this using SPARQL property path syntax:
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
prefix dcat: <http://www.w3.org/ns/dcat#>
prefix dct: <http://purl.org/dc/terms/>
prefix n4e: <http://nfdi4earth.de/ontology#>
prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
SELECT ?standardTitle ?theme ?subjectLabel
WHERE {
?repository dct:title "PANGAEA"@en .
?repository dcat:theme ?theme .
?repository n4e:supportsMetadataStandard ?standard .
?standard dcat:theme ?theme .
?standard dct:title ?standardTitle .
SERVICE <https://nfdi4earth-knowledgehub.geo.tu-dresden.de/fuseki/dfgfo/sparql> {
{
?theme rdfs:subClassOf+ <https://github.com/tibonto/dfgfo/34> .
?theme rdfs:label ?subjectLabel
} UNION {
?theme rdfs:subClassOf+ <https://github.com/tibonto/dfgfo/202-02>
} UNION {
?theme rdfs:subClassOf+ <https://github.com/tibonto/dfgfo/203-03>
} UNION {
?theme a owl:Class .
FILTER ( ?theme = <https://github.com/tibonto/dfgfo/202-02>||<https://github.com/tibonto/dfgfo/203-03>)
}
}
}
We should try to implement this and then look how it works out