Our understanding of software design does not yet suffice to enable economically efficient production of new software-intensive systems, either commercial-scale or the ultra-large-scale (ULS) systems emerging as priorities for government and industry. New design methods are not enough. Also needed is a science of the phenomena underlying design. Elements of such a science appear within reach, based on the integration of knowledge from software engineering and financial economics. A premise of this project is that there is essential uncertainty in the technical and economic outcomes of design activities, therefore designs and design activities must be engineered to account for and exploit this risk. Modularity in the design of software artifacts, processes, and industry structures is one key approach. This project aims to develop and evaluate value-oriented scientific approaches to modular system design in order to provide the knowledge needed to significantly reduce the economic inefficiency of current software practices. Specific goals include the following: developing methods to represent design structures at many scales, based on the notion of design rules of Baldwin and Clark; mapping modular design structures to economic value using options theory; and devising practical ways to develop and assess architectural strategies for both commercial-scale and ultra-large-scale systems.
