18th International Conference on Cooperative Information Systems (CoopIS 2010), Crete, Greece, October 27-29
Abstract. In this paper we deal with problems occurring in evolving interrelated Web databanks. Examples of such databanks are networks of interlinked scientific repositories on the Web, managed independently by cooperating research groups. We argue that changes should not be treated solely as transforming operations, but rather as first class citizens retaining structural, semantic and temporal characteristics. We propose a graph model called evo-graph for capturing in a coherent way the inherent relationship between evolving data and changes applied on them. Evo-graph represents changes as arbitrarily complex objects, similarly to data objects. We discuss the temporal characteristics of the evo-graph, and show how the evo-graph can provide past snapshots of the data. To uniformly express temporal and provenance queries we introduce evo-path, a path expression language based on XPath. Evo-path takes advantage of complex changes in the evo-graph in order to answer queries that interpret and elucidate data evolution.