A Versioning and Evolution Framework for RDF Knowledge Bases

We present an approach to support the evolution of online, distributed, reusable, and extendable ontologies based on the RDF data model. The approach works on the basis of atomic changes, basically additions or deletions of statements to or from an RDF graph. Such atomic changes are aggregated to compound changes, resulting in a hierarchy of changes, thus facilitating the human reviewing process on various levels of detail. These derived compound changes may be annotated with meta-information and classified as ontology evolution patterns. The introduced ontology evolution patterns in conjunction with appropriate data migration algorithms enable the automatic migration of instance data in distributed environments.