Principles, Patterns, and Process: A Framework for Learning to Make Software Design Decisions.

摘要

Software system design is commonly acknowledged to be difficult: the problems are ill-formed, complex, and lack complete information. Expert designers employ strategies developed and refined over years of experience, while novices attempt to apply methods they have learned in the classroom. Existing research on software design education is scarce and focused on identifying the characteristics of student designers versus the attributes of expert designers. I am interested in studying how to bridge this gap through the novel adaptation of a model of complex systems development and evolution. This adaptation is realized in the Principles, Patterns, and Process Framework ( P3F ), a suite of learning tools designed to help software design students learn and apply key aspects of expert designers' strategies and behaviors.;This dissertation presents an experimental validation of the P3F through a triangulated research approach. Collected data measured the research subjects' perceptions of their design skills and strategies, identified and categorized actual design behavior observed while the subjects worked on complex software system design problems, and evaluated the quality of the solutions they produced for these problems. The results of this study suggest that the P3F can help a novice software designer adopt and apply expert-like design strategies, recognize the development of these strategies as part of their own design process, and contribute to an improvement in the quality of their designed solutions.;Three significant contributions are made by this dissertation. It provides a comparative analysis of expert and novice design strategies and behaviors in a form not found in the literature. It also demonstrates a triangulated experimental model that features three research methodologies and analysis techniques. Most importantly, it presents and validates the Principles, Patterns, and Process Framework.