Output details
11 - Computer Science and Informatics
University of York
ARTOO : adaptive random testing for object-oriented software
<09>Demonstrates new inter-object distance measure to drive adaptive test input selection for OO software. Comparsion against leading randomised strategy (ART) demonstrates (average) 5-fold (sometimes two orders) reduction in test cases needed to find first fault, albeit with greater computation cost (average 1.6 times). Where low number of test cases is more important than execution time (e.g. when oracle is manual, most often the case) potential savings are huge. ARTOO is also less sensitive to seed choices. ICSE acceptance rate was 15%. Paper is widely discussed in the literature. PhDs (Ciupa, Leitner) now Software Test Managers at Credit Suisse and Google.