Closed bquinn closed 2 years ago
At last week's meeting (11 Jan) we concluded that it's not a good idea, and any instances of DomainIncludes / RangeIncludes are probably "code smells" / anti-patterns.
@bquinn and @silveroliver will go through the existing ontology and look at each instance of domainIncludes / rangeIncludes and work out whether to turn them into parent classes or separate properties each time.
Here's a list of the existing instances of domainIncludes and rangeIncludes...
sport:competition schemaorg:domainIncludes sport:CompetitionPhase, sport:Event . sport:participation schemaorg:domainIncludes sport:Season, sport:Event, sport:Action .
I've noticed that in schema.org, they define some properties to have a range or domain (or both) that allows multiple nodes. This can be represented in RDF Schema so they invented their own properties, domainIncludes and rangeIncludes that have be multi-valued.
If we use this construct, it allows us to have properties (eg
name
) that can apply to many classes, and it also allows us to have some properties which can either have objects or literals as values (e.g. a Person or a string). (Schema.org uses this a lot.)Is it a good idea for us to follow the same approach?