Defining Requirements for Vehicular Software Integration Testing

Sammanfattning: In every software/system development life cycle, requirement representation is one of the crucial phases. The future success of the development process and end product quality very much depends on the quality of the requirements. Often, ambiguous and inconsistent requirements lead to a rework during the later stages of the development process. Furthermore, requirement specifications are used as an important input for testing activities, such as generating test case specifications. This thesis presents a Systematic Literature Review (SLR) of studies related to requirement representation (definition) for software integration testing. The purpose of this research is to improve current knowledge in requirement representation for integration testing in order to identify the state of the art in the area. We are interested in identifying possible suggestions for requirement representation, such as approaches, tools, lesson learned, and techniques that can be used for integration testing. The search strategy primarily retrieved 1060 studies in total, out of which 56 were considered as primarily relevant studies. However, during the data extraction phase of the SLR, 40 papers were found irrelevant according to inclusion/exclusion criteria. These studies were divided into different categories: model-based, executable specifications, natural-textual language, and combinational language. The systematic review aims at identifying probable gaps in research about requirement representation (definition) for software integration testing. The thesis work presents the results of performing data analysis and synthesis regarding all final selected studies. The proposed suggestions for requirement representation and their relation to integration testing are summarized. Although not many studies were found that addressed requirement representation for integration testing, this issue has been an interest of scientists as well as practitioners for many years. The majority of reviewed publications used a model-based approach in order to solve the issue of requirement representation for integration testing. On other hand, one of the most problematic ways of performing requirement representation is using natural-textual language. Expressing requirements in natural-textual language may introduce ambiguity or inconsistency, and these ambiguities/inconsistencies can be noticed in the later stages of the development process. The later they are identified, the higher the cost of eliminating these ambiguities or inconsistencies, which can significantly increase the development costs. Thus, the authors recommend using some support tool or methodology when expressing requirements in natural-textual language. Using a support tool or methodology can help to identify and eliminate ambiguities or inconsistencies in the early phase of the system/software development process. In the other publications, the authors combined two different requirement expressions together in order to improve the requirements' quality and perform virtual integration testing in the early phases of the development process.