iMeMex: From Search to Information Integration and Back

Abstract

In this paper, we report on lessons learned while building iMeMex, the first incarnation of a Personal Dataspace Management System (PDSMS). In contrast to traditional keyword search engines, users may not only search their collections with iMeMex but also semantically integrate them over time. As a consequence, the system may improve on precision and recall of queries in a pay-as-you-go fashion. We discuss the impact of several conceptual and architectural decisions made in iMeMex and outline open research challenges in combining search and information integration in order to build personal dataspace management systems.