Regresar

In search of what we experimentally know about unit testing

Abstract:

Software engineering is a relatively young field and experimental software engineering is even younger, so undisputed facts are few and far between. Nevertheless, experimental results can help practitioners make better decisions, and the authors have aggregated results derived from individual unit-testing experiments published in high-quality journals and proceedings. Most of the experiments focus on two important characteristics of testing techniques: effectiveness (number of faults found) and efficiency (effort required to apply the technique). Some experiments study the quality of test-case sets according to different criteria. Although the aggregation results are far from ideal, they identify grounded information that's potentially useful for testing practitioners in test-case generation, test-set evaluation, and test-case selection. © 2006 IEEE.