You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Since STA does not have a formal RDF vocabulary, is it better to develop a mapping from STA JSON elements (it has not formal schema) to SOSA - i.e. a JSON-LD context and regard this instead as an implementation? Otherwise we are bring a range of alignment languages into play which may be hard to manage. Alternatively we could provide such a mapping as an informative annex?
in SOSA, as an RDF language, we can explicitly use polymorphism at the cost of maintaining and publishing RDFS or OWL Class definitions. when using JSON-LD annotations we can declare multiple properties to be of type "@type" - with the constraint that the values is, or can be mapped to, a URI (either with an @base or a CURIE using a declared prefix).
I think with the BuildingBlocks approach to publishing STA elements and STA-SOSA bindings we could declare two different profiles with mappings using JSON-LD without introducing a new SOSA element - noting soft-typing will always be present in the real world and maybe SOSA should not try to cover this explicitly.
When aligning STA with SOSA and OMS, we run into 2 different profiles:
If we pulled OMS:ObservationType into SOSA core, we could define rules on which ObservationType leads to which mapping profile
The text was updated successfully, but these errors were encountered: