Selecting an appropriate RequirementsBased Test Case PrioritizationTechnique

Abstract: Context: Regression testing is done to ensure that the modified code of the software is working as it is supposed to. The simplest strategy for regression testing is to re-execute all the test cases (TC), but due to constraints such as time, cost and resources this strategy is inefficient to implement. To reduce the cost of regression testing and to increase its efficiency, Test Case Prioritization techniques are developed. Test case prioritization (TCP) techniques reduce the cost of regression testing and at the same time keeping up its quality. Various procedures for prioritizing test cases were developed to endeavor their viability to meet some prioritization goals (PG). However, the effectiveness of a TCP technique varies for each software product according to their product attributes, change attributes, their interactions, test suite characteristics, and resources for testing. Therefore, one TCP technique for every software product is impractical and leads to misusing the resources. Moreover, the existing research on how to select a TCP technique for a given software product under test is not much focused. Hence, it is necessary to find ways to select a TCP technique to ensure that the PGs of a software product under test are met. Objectives: Firstly, to identify different requirements-based TCP techniques in the existing literature and in software organizations, along with their PGs. Secondly, to identify similarities and differences between the identified requirements-based TCP techniques. Finally, finding a way to select a requirements-based TCP which ensures its PGs for a software. Methods: For this research, a systematic literature review (SLR) was conducted to gain insights into research on Requirements-based TCP techniques in the existing literature. Semi-structured interviews are conducted with the personnel working in software testing domain to acquiring insights of requirements-based TCP techniques utilized in the software organizations. Depending on the objectives of this research both open-ended and closed-ended questions are formulated in the interview questions. Data obtained from the SLR and interviews were analyzed to answer my RQ’s. Conclusions: Results from this research state that selecting an already proposed requirements-based TCP technique from the literature for regression testing a software product is not an optimum solution, instead, selecting requirement properties for a requirements-based TCP technique based on product demands and available resources is the optimum way to ensure successful regression testing. Therefore, for selecting appropriate requirements-based TCP technique, firstly, PGs for a particular requirement/feature being tested is formulated based on the testing phase of the project and available resources. In the second step, the essential aspects which affect the quality of the requirements/features under test are identified, this is done by analyzing the 3 requirement properties which are customers perceived priority, developers perceived priority and tester perceived priority of requirements. The identified critical aspects are prioritized with respect to their criticality and availability of resources. In the final step, with the available priorities of the critical aspects, the remaining relevant requirement properties can be selected to cover test cases of all the important aspects of the requirements under test in that testing phase in a test suite focusing on ensuring its PGs. Thus, the PGs are ensured, and the resources allotted for testing the project can be effectively utilized. Therefore, for an appropriate requirements-based TCP technique, 3 requirement properties, customers, developers and testers perceived priorities, should be selected. Depending on the product demands and available resources if necessary, the 3 requirement properties choose remaining requirement properties and other TCP approaches.