Using semi-formal methods for detecting interactions among smart homes policies
Science of Computer Programming
| Hi-index | 0.00 |
Finding ways of detecting interactions between requirements is essential in order to develop a set of clear requirements, which serves as a foundation for successful software development. Detecting requirements interactions as early as possible helps avoid high repair costs. This thesis presents IRIS, I&barbelow;dentifying R&barbelow;equirements I&barbelow;nteractions using S&barbelow;emi-formal methods, which is a semi-formal approach for detecting requirements interactions. IRIS is a systematic six step approach that uses tables, graphs, interaction detection scenarios, and subjective judgment to detect requirements interactions in software systems. IRIS has the advantage of not only being domain independent but also customizable towards a specific domain in order to enhance its performance. IRIS helps reduce the number of necessary pair-wise comparisons between requirements that have to be performed informally by a human expert. This reduction is achieved by discarding irrelevant comparisons that will not lead to interactions. A general requirements interaction taxonomy was developed for identifying when two requirements are considered interacting. This requirements interaction taxonomy provides interaction detection scenarios that are used within IRIS for detecting interactions. To validate IRIS, it was applied to three different case studies from different domains. In the first case study, the lift system, IRIS was able to detect 7 interactions as opposed to 6 interactions that were detected by another approach reported in literature. IRIS was also able to achieve 17.6% reduction in the number of comparisons. The second case study analyzed telephony features and IRIS was able to detect 21 interactions with 17.9% fewer feature comparisons. This result is very good as other approaches that detected 22 interactions all used formal methods. The third case study looked at smart homes policies. IRIS detected 83 interactions with 19.3% fewer policy comparisons. The smart homes case study is a major contribution as the results from it serve as the first fully documented analysis of interactions between smart homes policies in literature. To facilitate the application of IRIS, a tool was implemented. IRIS-Tool Support (IRIS-TS) is built as an add-on module for DOORS which is a well-known commercial requirements management tool.