pennyl67 / lexmeta Goto Github PK

1. the relation "skos:exactMatch" between frbr:Manifestation and bibo:Document must be changed because:

• The skos mapping relations (exactMatch, closeMatch) imply that Manifiestation and Document are skos concepts.
• we can use owl:equivalentClass or rdfs:subClassOf, taking into account the implications, as outlined at see https://stackoverflow.com/questions/4192435/owls-equivalentclass-vs-subclassof (thanks @gkirtzou for spotting this). To be honest, I find rdfs:subClassOf "less dangerous"

2. Have we found the mapping relation between frbr:Manifestation and bibo:Document somewhere or is this something we have decided? I didn't find anything at the two ontologies but I may have missed it. The reason I ask is that we are mapping two classes that are defined in other vocabularies. I don't think we should do this. We can get what we want (i.e. reusing properties) by saying that ms:DatasetDistribution is rdfs:subClassOf frbr:Manifestation and rdfs:subClassOf bibo:Document.
3. We should add the relations we have at https://lexbib.elex.is/wiki/LexMeta and the article, i.e.

• ms:LexicalConceptualResource and frbr:Expression (again, I suggest using rdfs:subClassOf)
• see above for ms:DatasetDistribution and frbr:Manifestation
• for frbr:Work and dcat:DatasetSeries, my hesitation is that we are stepping at others' vocabularies. Yet, at MetaShare/LexMeta we don't have yet a class for this. We will need to introduce one - on my to-do's.

I suggest we add these two properties for future use. The numberOfEntries is ok for dictionaries, but there will be cases that providers will wish to differentiate between different size units (e.g. triples, concepts, etc.). Since the properties are optional, that shouldn't cause any problem if not filled in.

Is there a way to tell ontoology that it should list the values for these three properties in the documentation? For terms, we have lots of multilingual labels, and for classes/props, we also have labels and defs.

skos:definition values are listed; rdfs:comment are not. Should I use the former also for props and classes instead of the latter, so that the value appears in the documentation, or can ontoology be set to consider also rdfs:comment?

Please, remove it; it has been merged with "ms:accessLocation" because it created confusion and programmatically the two made no difference.

range of ms:language should be dcterms:LinguisticSystem.
ms:Language is defined as a subclass of dcterms:LinguisticSystem.

What's the difference between them? no domain and no range

Similar to has been declared with the ms prefix and the IRI the prefix point to:
Change
<http://w3id.org/meta-share/meta-share/ms:SimilarTo
to
ms:SimilarTo

There are various properties re-used from other vocabularies without a range (e.g. dcterms:publisher) and others that do (e.g. dcterms:creator). Not sure if this is done on purpose. I think we should be consistent (to the extent possible) for documentation purposes at least.

I think we decided that we can use values from other namespaces in a concept scheme (see https://www.w3.org/TR/skos-primer/#secextension). Therefore, there's no need to redefine values such as accessibleThroughInterface in LexMeta, as this already exists in MetaShare.
The only reason for adding a new value is if we want to narrow it as pertaining only to dictionaries by assigning new relations. For this value, the skos:broader lexmeta:onlineDictionary does not support corpora that are accessible via an i/f. Thus, if we want to keep this skos:broader relation, I suggest we add a new value dictionaryAccessibleThroughInterface and add also skos:broader ms:accessibleThroughInterface

There are properties that have as domain and/or range a broad class (e.g. schema:bookEdition has domain frbr:Endeavour) and others with a narrower (e.g. ms:resourceName with ms:LexicalConceptualResource). What's the rationale behind this?

All classes and properties appear at the LexMeta documentation as named individuals (metashare.ilsp.gr/ontologies/lex-meta/lex-meta.v1.0.0.prelease/documentation/index-en.html#namedindividuals). This is not explicitly asserted in the ontology (at least I can't find it). Trying to figure out why this happens.
One idea might be that this is due to the relation lexmeta:wikibaseEntity, which has no domain or range; could it be that wikibase entities are defined as named individuals, and this causes the same for the equivalent item in LexMeta?
Any other ideas?

There's an object property http://iflastandards.info/ns/fr/frbr/frbrer/P3057 in the FRBR domain. Seems like an error