- Define relevant fields, extract content and sort into a structured data type (class, interface, json whatever) - Should give insight into the data differences (content, semantics) being provided by the different Trias APIs - Best case make it a starting point for implementing RDF/OWL annotations