Change management and validation for collaborative editing of RDF datasets

Abstract

The dynamic and distributed nature of the Semantic Web implies that datasets are often the result of collective participation rather than isolated works. Change management, provenance tracking and validation of changes performed by contributing agents are all requirements of systems for collaborative dataset development. Different scenarios may as well require mechanisms to foster consensus, resolve conflicts between competing changes, reversing or ignoring changes etc. In this paper, we perform a landscape analysis of version control for RDF datasets, emphasising the importance of change reversion to support validation. Firstly, we discuss different representations of changes in RDF datasets and introduce higher-level perspectives on change. Secondly, we analyse diverse approaches to version control. We conclude by focusing on validation, characterising it as a separate need from the mere preservation of different versions of a dataset.