[Editorial] terminology usage related to blank node mapping in normalized dataset
iherman opened this issue · comments
At the moment, the normalized dataset includes:
the canonical issuer — mapping identifiers in the input dataset to canonical identifiers
and this is what is returned at the end of the algorithm. Strictly speaking, that is an overkill. The canonical issuer is defined to be an identifier issuer which refers to a BNode Identifier Issuer State which includes additional information that is irrelevant to the caller of the algorithm. The only information that is of importance is a simply mapping from input bnode ids to the canonical ones (essentially the relevant issued identifier map).
I propose to replace the canonical issuer reference in the normalized dataset to an issued identifiers map in the definition of the normalized dataset.
Yes, all that's really needed is the map from the final state of the canonical issuer.
Yes, we should just use the map from the canonical issuer.