Closed acka47 closed 2 years ago
Hmm, with this version, the schema space is basically the same thing as the identifier space, right?
I don't think so. The identifier space defines the namespace for the resources to reconcile against, the schema space defines the type of resources one has. In lobid-gnd we woud have different values:
I could imagine a reconciliation service for places in Wikidata which would use the following:
However, I do not think that we should stick to this definition for historical reasons and would be ok with sticking to the current definition in the spec.
I am also fine with making it closer to the historical version, but would ideally like a wording which would make the difference between both parameters a bit clearer. With this version, the two are really close:
The difference between "type" and "sort" is quite subtle!
The difference between "type" and "sort" is quite subtle!
One refers to the namespace of the entity identifiers, the other to the entity type. Maybe something like this?
However, it is quite unintuitive to give an entity type in a field called "schema space", I'd rather expect something like in the current draft.
Yes that would be better!
@acka47
identifierSpace: The URI namespace for the entity identifiers returned by this service. [...]
That is better but not by much, it gives information about the labeling of the namespace. It would be better to have a description that also includes what is being identified and why (it might be useful) to know or disambiguate with. I.E. I almost feel like we should describe a bit about what an "identifier" is and what is being represented by it. Maybe flipping the wording to "...for the identifiers of an entity returned by this service" ?
Pushed an attempt to join the suggestions by @acka47 and @thadguidry, tweaked for consistency: https://github.com/reconciliation-api/specs/pull/76/commits/22a7e37e35d0c79f5796c19ac410739cf479d812
Merging this to move forward, but we are happy with further tweaks!
Resolves #3