Practitioners' Expectations on Automated Test Generation

Software testing is a labor intensive and expensive task that generally accounts for a significant amount of software development cost. To reduce the high cost of manual testing and increase the reliability of the process, practioners have tried to automate it. This paper examines both the theoretical and practical role of testing in software development, then goes on to survey the purposes and limitations of testing as presently conceived. Various methods of automating test-data are presented, then the focus is reduced to program-based test data generation: using structural testing approaches such as data flow testing and control flow testing to demonstrate a systematic procedure for selecting valid and reliable test data. Various attempts have been made to automate the generation of test-data from an actual program but there are still unresolved problems to date. Despite flaws in the current methods, it is a step forward from manual testing and practioners endeavor to use these automated test tools in conjunction with manual ad-hoc testing to reduce cost and increase the reliability of software systems, while researchers endeavor to resolve the existing issues. Automated testing is an important topic in software engineering and the amount of research in the field suggests, both practioners and researchers believe that -Automation is the Key!