Maintaining architectural conformance through the mapping of architectural decisions to implementation practices and patterns

点击翻译按钮获取中文摘要

Architectural decisions express rules and actions that shape software architecture and have a large impact on software quality. Common architectural decisions include applying design solutions such as architectural styles, design patterns and architectural tactics, and choosing appropriate technologies. While existing technologies enable to rapidly build software systems, developers still need to understand the selected design solutions and be able to correctly implement them using the chosen technologies.****In fact, despite the progress made by existing approaches in managing and documenting architectural decisions, we still lack effective support for enforcing architectural styles, tactics and patterns during development and evolution of the system. Also, recent empirical studies revealed that developers do not always understand the impact of their changes on architecture. This is partly due to the fact that developers do not know what styles, tactics, and design patterns were adopted. Even if they knew, they may not know how to implement these design solutions using the chosen technology; i.e., existing technologies implement differently the same design solutions and they have their own design practices and patterns.****Therefore, the goal of the proposed research program is to study common existing technologies to figure out how they support design solutions (i.e., architectural styles, tactics and design patterns); and use this knowledge to develop techniques and tools that enforce these solutions' constraints and prevent architectural erosion. To do so, we propose to first build a knowledge base that contains explicit representations of the constraints and interaction protocols inherent to common styles, tactics and patterns; and explicit representations of technology-specific implementations of these styles, tactics and patterns including how constraints and interactions are mapped. Second we propose to leverage model-driven techniques to build appropriate approaches and tools that use the knowledge base to translate constraints and interactions into the context of specific projects, and enforce these constraints.****The results of this research program will contribute to a better understanding and support of design solutions including architectural styles, design patterns and architectural tactics, and enforcing these solutions during software development and evolution. By supporting developers in understanding design solutions and their appropriate implementations using common existing technologies, we contribute to enhancing the quality of software systems, increasing productivity and reducing maintenance costs.**