5. Conclusions

The primary conclusions from the initial QARCC-1 experiment were: • QARCC-1 can help users, developers, customers, and other stakeholders to identify conflicts among their software quality requirements, and to identify additional potentially important quality requirements. In the experiment, QARCC-1 found both of the quality conflicts previously identified, plus five others considered significant. • QARCC-1 needs further refinement to avoid overloading users with insignificant quality conflict suggestions. In the experiment, QARCC-1 suggested three potential conflicts considered insignificant. We are refining the knowledge base to address more detailed quality attributes in a more selective fashion. Two additional conclusions derive from feedback based on demonstrations of QARCC-1 to a number of USC-CSE's industry and government Affiliates. There was a strong consensus that QARCC-1 provided a useful framework for stakeholders to systematically resolve software quality attribute conflicts, and that the semi-automated approach provided a good way to balance human skills and computer tools in addressing quality tradeoff issues. Two additional conclusions are more tentative, awaiting the development of and larger-scale experimentation with QARCC-2. These are that the QARCC approach is usefully scalable to large systems with many quality conflicts, and that the effectiveness of the QARCC approach is largely domain-independent.