Closed jesserobertson closed 4 years ago
@fils there's nothing in that schema at the moment. We probably need to define a simple geo schema too.
Also did we discuss mapping schma.org/dataset into our description schema? Is it at the right level?
Relationship types are in the IGSN docs here: https://github.com/IGSN/metadata/wiki/IGSN-Registration-Metadata-Version-1.0
There's also a partial crosswalk with DC for some of the other terms on that page
Many of these relationships involve (i) a pointer to the related sample/party/service (ii) information about the nature of the relationships - probably drawn from a codeList or controlled vocabulary (iii) other annotations
The standard/scalable way to support this is through an association-object - i.e. an intermediate node, which carries both the onward pointer to the relative, and the additional information. In PROV-O these are called qualified terms. However, because PROV is activity-centric, there is no qualified association between entities except for 'derivation'. Because of this, in DCAT we created an additional association type, the qualified relation for relationships between entities (e.g. samples). For explanation and examples see DCAT qualified forms.
Of course you could create a new predicate for every single relationship type, but using a decoration from a separately maintained vocabulary is probably more scalable, and the intermediate node also supports additional annotations.
Ok I'll create a new issue around a more scale-able way of checking relationships for after the sprint - my concern is that we may want to be specific about what relationships we want to support here?
Need to fill out
relationships.schema.json
to define related identifiers.