A model for decision maintenance in the WinWin collaboration framework

Cost-effective engineering and evolution of complex software must involve the different stakeholders concurrently and collaboratively. The hard problem is providing computer support for such collaborative activities. The WinWin approach being developed and experimented at the USC Center for Software Engineering provides a domain independent solution for the stakeholders to cooperate in the requirements engineering phase of the software lifecycle. A major problem confronted in the WinWin framework is aiding decision coordination-coordinating the decision making activities of the stakeholders. A key element in supporting decision coordination is decision maintenance. As decisions undergo evolution, the effects of such changes on existing decision elements must be determined and the decision structure appropriately revised. The paper presents an approach to addressing the problem of supporting decision maintenance. The key ideas involve a) defining an extended ontology for decision rationale, that models the WinWin decision space and their states, b) formally describing a theory based on that ontology that specify conditions for states to hold, and c) defining an agent that utilizes the theory to determine revisions and coordinate with other agents to propagate revisions in a distributed support framework