Fix sameAs issues in ARCRC

amoeba commented 2 years ago

While looking at something unrelated I noticed twoowl:sameAs triples pointed at RDF literals instead of URIs. See the below turtle-formatted triples and the owl:sameAs triples.

###  http://purl.dataone.org/odo/ARCRC_00000006
<http://purl.dataone.org/odo/ARCRC_00000006> rdf:type owl:Class ;
                                             rdfs:subClassOf <http://purl.dataone.org/odo/ARCRC_00000013> ;
                                             obo:IAO_0000115 "A water temperature which inheres in water close to the surface of an ocean or sea. The exact meaning of surface varies according to the measurement method used, but it is between 1 millimetre (0.04 in) and 20 metres (70 ft) below the sea surface." ;
                                             dc:creator <http://orcid.org/0000-0002-5300-3075> ;
                                             dc:date "2020-11-04T19:43:19Z"^^xsd:dateTime ;
                                             rdfs:label "Sea Surface Temperature Indicator"@en ;
                                             owl:sameAs "http://purl.dataone.org/odo/ECSO_00001523" ;
                                             skos:altLabel """SST
temperature of sea surface
ocean surface temperature
temperature of the ocean surface""" .

###  http://purl.dataone.org/odo/ARCRC_00000007
<http://purl.dataone.org/odo/ARCRC_00000007> rdf:type owl:Class ;
                                             rdfs:subClassOf <http://purl.dataone.org/odo/ARCRC_00000013> ;
                                             obo:IAO_0000115 "Ice which has formed by the freezing of sea (saline) water." ;
                                             dc:creator <http://orcid.org/0000-0002-5300-3075> ;
                                             dc:date "2020-11-04T19:43:23Z"^^xsd:dateTime ;
                                             rdfs:label "Sea Ice Indicator"@en ;
                                             owl:sameAs "http://purl.obolibrary.org/obo/ENVO_00002200" .

mbjones commented 2 years ago

Yeah, seems like they should be URIs.

Another question is why owl:sameAs was used rather than owl:equivalentClass. I have always thought sameAs was used for resources/instances. However, I also read that when sameAs refers to classes or properties, it also means that an equivalentClass or equivalentProperty relation exists, respectively. This is from: http://www.ksl.stanford.edu/software/jtp/doc/owl-reasoning.html

When an owl:sameAs relationship is asserted or inferred between two entities that are known to be classes, an owl:equivalentClass relationship is inferred between the classes. Similarly, when an owl:sameAs relationship is asserted or inferred between two entities that are known to be properties, an owl:equivalentProperty relationship is inferred between the classes

If that is correct, it probably doesn't matter which we use and they should work the same from a reasoning perspective wrt alignment. Thougths on that?

amoeba commented 2 years ago

The sameAs<->equivalentClass thing is really good to know.

Thougths on that?

I think we should test it with the tools we use, specifically Jena where we use RDFS and OWL-Full reasoning (it can be tweaked).

mbjones commented 1 month ago

@justinkadi when you are fixing the sameAs issues in ARCRC, can you check on this assertion? It is probably safer to use owl:equivalentClass in any case when that is what we mean, and reserve owl:sameAs for instances. Let's try to be consistent.

DataONEorg / sem-prov-ontologies

Fix sameAs issues in ARCRC #112