Intelligent analysis of software architecture rationale for collaborative software design

摘要

A set of principle design decisions drives architects to design software architecture for a system satisfying requirements. The design decision making process involves a group of stakeholders exchanging their viewpoints to address various concerns and reach a consensus. Often the architecture rationale behind various design decisions is not fully captured and hence affects the maintainability of software systems. In this paper, we identify three research challenges to analyze the architecture rationale in various perspectives to provide stakeholders with a more detailed view that aids them in decision making. Firstly, we determine collective opinions of a group on different viewpoints and detect viewpoints which have gained a significant attention into the online discussion. Secondly, we propose a method to develop a traceability matrix that links various software architecture elements to its related software requirements. The requirements traceability helps to maintain software systems and address change management. Thirdly, We perform textual analysis of stakeholders' views to determine the topics that are most discussed. In order to capture and maintain the software architecture rationale for analysis, we briefly present the design of an intelligent software architecture rationale capture system that enables stakeholders to participate in an online discussion to resolve a design issue collaboratively. The system captures a structured design rationale which maintains its links to software requirements and architecture elements. Finally, a comprehensive empirical study is presented. The results from the study indicate that our system is effective in capturing and supporting stakeholders in a collaborative decision making.