This paper addresses the issue of discerning properties of a dynamic system that emerge due to interactions that take place in the system. Usually these are emergent properties that cannot be discerned by studying static snapshots of the problem domain.
In building paradigms for the above, we consider a specific problem - that of analyzing transactions among a set of autonomous actors in order to identify possible emergent properties. The application context is that of designing a workflow that handles transactions and logistics among a set of autonomous organizations. In this paper, we outline the general principles used to carry out such an analysis process. A more comprehensive characterization of the problem and the approach may be found in .
Interaction properties are identified along three ``views''. Firstly transactions are analyzed to identify interaction patterns among the actors. A set of ``interesting'' patterns are then selected for discerning further properties of the interactions. Specifically, we address two other kinds of properties - behavioral, which identifies orderings or dependencies that could exist among the events in the pattern; and functional, which identifies possible semantic states in an interaction pattern.