A Framework to Improve the Process of Software Regression Testing Open Access
Downloadable ContentDownload PDF
Software testing is the most expensive and time-consuming aspect of software development. It is estimated that as much as 50% of development time and money is devoted to various forms of testing (Ramler, 2006). Therefore, it is crucial that all testing conducted must be performed efficiently, quickly, and accurately. Automated testing has been introduced in the past few years to help with the testing process. In theory an automated testing system can execute steps faster and more accurately than even the most experienced test operators. An automated test system does require a large investment to acquire and implement. It requires many hours for an organization to learn how to operate the system, write a proficient test script, and learn how to debug scripting problems. This often takes many hours of trial and error (Ramler, 2006). However, if used appropriately, it could be worthwhile for any organization that performs an extensive amount of software regression testing (Dustin, 2009). Yet, automated testing in certain situations does not operate perfectly. During an automated test some level of configuration alignment errors, loading errors, disconnection between avionics systems, minor intermittent glitches, and system crashes can occur (Pinheiro, 2010). These kinds of testing errors can be a major setback for a program. Therefore, the software regression testing would be more effective if an engineer or test operator is made available to oversee the process and address issues manually. The first part of the research will analyze and compare the two current testing methods against each other to see if there are significant differences in testing errors and/or time spent on testing and analysis. The second part of the research will analyze the hybrid approach against automated and manual approaches to see if there are significant differences in testing errors and time spent on testing.