Redefining Regression Testing and Details under It Regression for Fix and Regression for Risk
Daniel L Asfaw · 2014
Regression testing has two shared goals. The prime one would be to help building a bug free application, while the other equally important goal could be to maintain the healthiness of an existing system. If the emphasis is on the first objective, that testing would conventionally be named as regression for fix, and if otherwise, it is termed as regression for risk. The former implicates activities destined to verify the entire software development process and validate the end product. The later centers on retaining the wellbeing of an existing product by periodically checking on its quality, to make sure if it still meets standards and requirements. This article, which is a part of an ongoing PhD dissertation in computer science and software testing in an Inter-University Program at Universidad Azteca and Universidad Central de Nicaragua (UCN) in Mexico City - Managua Nicaragua, attempts to redefine regression testing focusing on the aforesaid two details. To this end, related literatures have been reviewed. Best industry practices have also been assessed and analyzed. 50 expertise from various levels and walks of skill sets have been involved through questionnaire, discussions, and interviews. The finding reaffirms regression for fix as a functional testing directed to examine if the fix is done properly, and if the defect is resolved and has no aftermaths. By the same token, it redefines regression for risk as a recurrent functional testing routed to check on the ability of an existing software to meets standards and expectations. Regression testing, therefore, is an umbrella term encompassing, at least, both of the above.