| This dissertation investigates the nature of economic development: how do human societies advance in economic well-being? More narrowly, it investigates the role of capital goods in this advance: what is the nature of the processes by which people improve the capital structure? The dissertation addresses these questions through examining software development--its practices, tools, and technologies, and their evolution. Software development illuminates the nature of capital development in general.;The theoretical foundation of the work is Austrian capital theory. Neoclassical growth theory, including the "new growth theory" of Paul M. Romer, is found to be of little of use here, because it abstracts away from what Ludwig Lachmann calls the structural aspects of capital: its heterogeneity and complementarity. Following Carl Menger, this study attributes human advancement primarily to the advancement of human knowledge. Capital goods embody the knowledge of many. Because this knowledge is dispersed, often tacit, and incomplete, the capital structure evolves through a social learning process.;Requirements for software applications can rarely be established before development begins. Hence software developers have evolved procedures such as rapid prototyping for learning what the software must do. Prototyping is an iterative, interactive, dialogue-like learning process. Tools, languages, and methodologies are evolving which primarily enable understanding of the complex systems being developed. Object-oriented programming systems especially improve learning by providing rapid feedback and by allowing developers to program with high-level abstractions suitable for thinking about the problems being addressed.;Because conditions change, software must constantly evolve to maintain its value. Experience with software maintenance suggests that evolvability of software systems is fostered by modularity, which improved understandability, localized changes, and reduces system complexity. Modular, object-oriented techniques also enable construction of reusable software components, and hence improved specialization and division of knowledge.;Availability of reusable components may lead to component markets, given requisite social learning in the development of standards, new distribution channels and pricing techniques. Markets themselves would foster additional social learning by generating new information.;Exponential growth is checked by the difficulty of social learning: more rapid economic development depends on learning better at all levels of economic organization. |
