Enabling Active Ontology Change Management within Semantic Web-based Applications. Mini-thesis: PhD upgrade report

Enabling traceable ontology changes is becoming a critical issue for ontology-based applications. Updating an ontology that is in use may result in inconsistencies between the ontology and the knowledge base, dependent ontologies and applications/services. Current research concentrates on the creation of ontologies and how to manage ontology changes in terms of mapping ontology versions and keeping consistent with the instances. Very little work investigated on-the-fly keeping track of ontology changes while update (active ontology versioning) and using these information to control the impact on dependent applications/services, which is the aim of our research presented in this thesis. The approach we propose is to make use of ontology change logs as a check-point to analyse changed entities related to the requested services via end-user’s incoming queries (RDQL/SPARQL) and amend them as necessary to maintain the validation and continuousness of the dependent application. Firstly, We build up Log Ontology I as the concept structure to organize and construct the change information, develop our prototype system to demonstrate how the change information retrieved from Log Ontology I could be used to control the impacts brought by the ontology changes on the dependent applications and services. And then, by analysing the limitations and difficulties of our prototype system in maintaining the services related to the more complex ontology changes, we identify that the problem which fails the system facing the more complex ontology changes is the inabilities of Log Ontology I to represent complex change information in a semantic fashion. Therefore, we retract to put more focuses on Log Ontology I to enable the implementation of the mechanism to on-the-fly keep track of ontology change information, forming Log Ontology II, in order to reserve the semantics of ontology change from the beginning of ontology update process. Finally we discuss the future direction in terms of how the improved Log Ontology II enables the better service validation and continuousness maintenance of changing-ontology-based applications.