Abstract
Software systems have become complex and challenging to develop and maintain because of the large size of test cases with increased scalability issues. Test case prioritization methods have been successfully utilized in test case management. However, the prohibitively exorbitant cost of large test cases is now the mainstream in the software industry. The growth of agile test-driven development has increased the expectations for software quality. Yet, our knowledge of when to use various path testing criteria for cost-effectiveness is inadequate due to the inherent complexity in software testing. Existing researches attempted to address the issue without effectively tackling the scalability of large test suites to reduce time in regression testing. In order to provide a more accurate way of fault detection in software projects, we introduced novel coverage criteria, called Incremental Cluster-based test case Prioritization (ICP), and investigated its potentials by making a comparative evaluation with three un-clustered traditional coverage-based criteria: Prime-Path Coverage (PPC), Edge-Pair Coverage (EPC) and Edge Coverage (EC) based on mutation analysis. By clustering test suites, based on their dynamic run-time behavior, the number of pair-wise comparisons is reduced significantly. To compare, we analyzed 20 functions from 25 C programs, instrumented faults into the programs, and used the Mull mutation tool to generate mutants and perform a statistical analysis of the results. The experimental results show that ICP can lead to cost-effective improvements in fault detection.
Highlights
Despite the huge advancement in agile test-driven development, the problem of inexhaustible testing continues to pose a major challenge in software quality assurance
A better understanding of which structures in the programs contribute to the expense might help to choose when to use Prime-Path Coverage (PPC)
This led to an argument that the expense of Incremental Cluster-based test case Prioritization (ICP) is worthwhile because it will help the software testers find more faults
Summary
Despite the huge advancement in agile test-driven development, the problem of inexhaustible testing continues to pose a major challenge in software quality assurance. Owing to the complexities in software testing, developers and testers need ways to evaluate their testing efforts in terms of their ability to detect faults in order to make intelligent decisions about testing. The ability, given a test suite, to predict whether it is effective at finding faults is essential to rational testing efforts. In retrospect, using the set of defects discovered during a software product’s lifetime, the quality of a test suite could be evaluated by measuring its ability to detect those faults. Many efforts have been dedicated to code coverage criteria to monitor the thoroughness of software tests [2] [3] [4] [5]. Various comparisons have been made of the theoretical relations between coverage methods [8] [9] [10] [11] [12]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
