Software design decisions are usually made at early stages but have far-reaching effects regarding system organization, quality, and cost. When doing design, developers apply their technical knowledge to decide among multiple solutions, seeking a reasonable balance between functional and quality-attribute requirements. Due to the complexity of this exploration, the resulting solutions are often more a matter of developer’s experience than of systematic reasoning. It is argued that AI-based tools can assist developers to search the design space more effectively. In this chapter, the authors take a software design approach driven by quality attributes, and then present two tools that have been specifically developed to support that approach. The first tool is an assistant for exploring architectural models, while the second tool is an assistant for the refinement of architectural models into object-oriented models. Furthermore, the authors show an example of how these design assistants are combined in a tool chain, in order to ensure that the main quality attributes are preserved across the design process.
Co-Evolution of Object-Oriented Software Design and Implementation
