The key objective of this research project is to discover principles that underlie and enable the semantic interoperation of heterogeneous information systems. Such systems comprise widely diverse and locally autonomous components with potentially intricate mutual constraints. A major technical challenge is to design activities that may execute over a number of components, but nevertheless preserve the global semantic requirements, i.e., the integrity of the information at large. This research project addresses the above challenge by (a) innovating representations that capture the semantics of data and activities in a unified mathematical framework; (b) developing techniques for reasoning about the above semantics to derive the maximally relaxed semantic constraints on data; (c) formulating design rules through which complex activities can be systematically specified; and (d) implementing the design rules in a prototype automated reasoning tool to assist in capturing and validating specifications of heterogeneous activities. A centerpiece of this project is the evaluation and iterative refinement of the developed techniques with respect to a taxonomy of problems. Continuous evaluation, greatly facilitated by the automated reasoning tool, is essential for discovering the lasting scientific principles of semantic interoperation.
