A Logic-Based, Value-Oriented Science of Design

University Of Virginia Main Campus, Charlottesville VA

Investigators

Abstract

Abstract 0438898 Kevin Sullivan University of Virginia Main Campus Architectural decisions define search spaces within which designs evolve. Architectural commitments decisively influence economic and technical success, yet we lack a validated account of key relationships between architectures and such outcomes. Baldwin and Clark have proposed an account of the nature of architecture in terms of design rules, and of its economic value in terms of real options. Sullivan has developed options-based accounts of the value of flexibility in design and tested the ability of Baldwin and Clark's models to explain information hiding and aspect-oriented design. A major open question remains: Can this approach not only describe but support development of complex, software-intensive systems? This research will perform one experiment to test the hypothesis that it can. If successful, the work will demonstrate support for the modeling of complex architectures in terms of design decisions, design rules, dependence relations on design decisions, and modularity in the dependence structure; and account for key sources of value in the system design, judged by the ways it evolved and was exploited to create value for users. A theoretical issue to be addressed is the possibility of defensible estimates of the sigma parameter of the real options valuation model.