Record standards-compliant provenance for datasets in a legacy data catalogue

RDA Group: 
Provenance Patterns WG
Contributor: 
Nicholas Car
Actors: 
Catalogue manager
Dataset contributor
Goal: 
Record provenance for datasets that can be interpreted as PROV-DM information using legacy dataset metadata models
Summary: 
A Catalogue Manager or Dataset Contributor wishes to record PROV-DM-compliant provenance for datasets in their catalogue but is constrained to using a non-PROV-DM and likely non-RDF dataset metadata model. In this Use Case, the actor will use a legacy dataset metadata model to record provenance but will reply on a mapping from that model to PROV to allow the provenance to be interpreted as PROV-DM provenance. This, in turn, relies on the legacy dataset metadata model having elements able to record provenance/lineage/dataset history that can be mapped to PROV. The Actor will then record the metadata in the legacy provenance format, ensure that the mapping to PROV-DM is available and indicated for use for this instance, and publish the provenance. Interpretation of the legacy metadata into PROV-DM-compliant provenance using the mapping may or may then be undertaken automatically by the catalogue or some additional tool or might not be undertaken here at all and instead left to users to undertake.
Preconditions: 
The legacy dataset metadata is able to be mapped to PROV-DM
The PROV-DM mapping is published
Postconditions: 
The legacy format to PROV-DM mapping remains available
Steps: 
Actor records the provenance of the dataset in the legacy format
Actor ensures that a legacy format to PROV-DM mapping is published and available for use
Actor indicates that the particular mapping should be used for the dataset's (this may be automatic)
Actor publishes the dataset metadata
Alternative Steps: 
System actor automatically interprets the published legacy format provenance as PROV-DM using the mapping and caches or publishes it