AN EXPERIMENTAL COMPARISON OF THE EFFECTIVENESS OF BRANCH TESTING ANDDATA-FLOW TESTING

An experiment comparing the effectiveness of the all-uses and all-edge s test data adequacy criteria was performed. The experiment was design ed so as to overcome some of the deficiencies of previous software tes ting experiments. A large number of test sets was randomly generated f or each of nine subject programs with subtle errors. For each test set , the percentages of executable edges and definition-use associations covered were measured and it was determined whether the test set expos ed an error. Hypothesis testing was used to investigate whether all-us es adequate test sets are more likely to expose errors than are all-ed ges adequate test sets. All-uses was significantly more effective than all-edges for five of the subjects, and appeared guaranteed to detect the error in four of them. Further analysis showed that in four of th ese subjects, all-uses adequate test sets were more effective than all -edges adequate test sets of similar size. Logistic regression analysi s was used to investigate whether the probability that a test set expo ses an error increases as the percentage of definition-use association s or edges covered by it increases. The evidence did not strongly supp ort this conjecture. Error exposing ability was shown to be strongly p ositively correlated to percentage of covered definition-use associati ons in only four of the nine subjects. Error exposing ability was also shown to be positively correlated to the percentage of covered edges in four (different) subjects, but the relationship was weaker.