Using Informal Knowledge for Improving Software Quality Trade-Off Decisions

To deliver high-quality software, in a software development process a variety of quality attributes must be considered such as performance, usability or security. In particular, quality attributes such as security and usability are difficult to analyze quantitatively. Knowledge about such quality attributes is often only informally available and therefore cannot be processed in structured and formalized decision-making approaches to optimize the software architecture. In this paper, we have defined a framework in order to make use of informally available knowledge in automated design decision support processes. We connect qualitative reasoning models with models for quantitative quality estimation to optimize software architectures regarding both knowledge representation models together. By our approach quality attributes for which no quantitative evaluation model is available can now be used in automated software architecture optimization approaches. For evaluating our approach, we demonstrate its benefits using a real-world case study and an example that is related to a real-world system.