Proactive prevention of false-positive conflicts in distributed ontology development
A Version Control System (VCS) is usually required for successful ontology development in distributed settings. VCSs enable the tracking and propagation of ontology changes, as well as collecting metadata to describe changes, e.g., who made a change at which point in time. Modern VCSs implement an optimistic approach that allows for simultaneous changes of the same artifact and provides mechanisms for automatic as well as manual conflict resolution. However, different ontology development tools serialize the ontology artifacts in different ways. As a consequence, existing VCSs may identify a huge number of false-positive conflicts during the merging process, i.e., conflicts that do not result from ontology changes but the fact that two ontology versions are differently serialized. Following the principle of prevention is better than cure, we designed SerVCS, an approach that enhances VCSs to cope with different serializations of the same ontology. SerVCS is based on a unique serializa tion of ontologies to reduce the number of false-positive conflicts produced whenever different serializations of the same ontology are compared. We implemented SerVCS on top of Git, utilizing tools such as Rapper and Rdf-toolkit for syntax validation and unique serialization, respectively. We have conducted an empirical evaluation to determine the conflict detection accuracy of SerVCS whenever simultaneous changes to an ontology are performed using different ontology editors. The evaluation results suggest that SerVCS empowers VCSs by preventing them from wrongly identifying serialization related conflicts.