EDIT

This is strongly related to #10322. There we already split SpecimenOrObservationBaseDTO and subtype DTOs into relatively pure DTOs and loader classes. But there are still a couple of issues left.

• sort order for derivatives => implemented for sort on the short label
• loading of the instances still partly takes place in the service classes, therefore some of the computing steps (computing summary label, boolean hasXXX values and updating the data after adding a derivative, ...) still takes place in the DTO but should be triggered by the loader class
• for map computation there is still a separate call to separate remote method taxonOccurrencesFor which internally does the call to the loader again. We need some similar structure as distributionInfoFor which combines DTO base values as well as map parameters => wrapping occurrenceInfoDto implemented, but not yet including map parameters
• do load not via model objects
• adapt DTOs so that they work with jackson (see #note-5, see also #6992)
• remove all model classes from DTOs (...) => this is mostly done, are there stillopen issues?
• implement paging (exists for model objects in OccurrenceService.pageRootUnitsByAssociatedTaxon() maybe used in old derivate path view in dataportal and (without paging) in TaxEditor
• create tests for DTO loading