Infrastructure

My understanding is that there shouldn't be a requirement to add a

parent_metadata_object cito:documents child_resource_map

statement. To me, this documentation isn't correct:

??A data package in DataONE is composed of at least one science metadata document describing at least one data object with the relationships between them documented in a resource map document.??

See https://releases.dataone.org/online/api-documentation-v2.0/design/DataPackage.html#synopsis

The six requirements of a DataONE Data Package that are over and above the OAI-ORE spec are found here:

https://releases.dataone.org/online/api-documentation-v2.0/design/DataPackage.html#generating-resource-maps

None of these require a @cito:documents@ statement, and so I think this documentation needs to be updated. Likewise, the @resourceMapSubprocessor@ code needs to be reviewed to remove any hard dependency on the presence of a @cito:documents@ statement. Currently, the @resourceMap@ field in Solr represents the @ore:aggregates@ statements, and I think that that only should be used to indicate participation in a resource map. In fact, I'd even prefer having the @aggregates@ and @isAggregatedBy@ Solr fields instead to show both directions of the relationship in the index.

An example is where a data manager calls @MN.create()@ on a bunch of data objects, and then calls @MN.create()@ on the resource map with these objects aggregated. Later, when the data manager is able to get the metadata from a scientist, they may call @MN.create()@ for the science metadata document, and @MN.update()@ on the resource map, adding the science metadata to the aggregation, and inserting @cito:documents@ statements.