Model-driven development of a person-centric mashup for social software

Based on the success of social software, modern information and communication systems are continuously moving from an information-centric data perspective to a more person-centric view. Easy access to federated activity streams of colleagues and other awareness information that is aggregated from different distributed intraand extra-organizational systems become more and more important for the daily knowledge work. The increasing number of platforms every person uses requires a flexible data integration solution that keeps track of the connections between the pieces of information and the persons involved in their creation in order to create a unified and aggregated view for work groups, teams and communities. This unified data collection is especially important for social network analysis and data mining as individual profiles and activities are meanwhile typically distributed over various source systems. In this paper we present the Community Mashup, a person-centric multi-user data integration solution for social software and similar systems that facilitates data aggregation and filtering while retaining the link to the pieces of information in the source systems. To support continuous evolution and flexible integration of frequently changing heterogeneous APIs and interfaces, we apply a model-driven development approach based on a therefore created person-centric data model. In addition to the conceptual design of the CommunityMashup, we present a reference implementation based upon open source components. Our overall goal is to build a multi-user mashup middleware for social software that offers an universal entry point in combination with unified data access for different client devices and can be used in various application scenarios with regard to individually specified service levels, e.g. continuous availability.