IMPLEMENTATION OF NEW APPROACHES OF SOFTWARE TESTING

In this thesis, the proposed experiment, and results of three experiments conducted to study the evaluation techniques, are presented. The results of the first experiment indicated that continuous testing has a great potential to complement manual testing activity. In fact, for defects associated with long message sequences and specific input data, continuous testing seems particularly appropriate to detect states that could potentially give rise to them. Evolution- mutations may contribute to the discovery of defects that are difficult to reveal, which would probably go unnoticed under goal-oriented and randomized testing. The results of this study showed that whenever interaction oncology has a non-trivial shape, the oncology-based generation method achieved a higher coverage of oncology classes than manual test case derivation. It also overcomes manual derivation in terms of the number of faults encountered and the portion of the input location discovered during testing. The level of automation achieved by our equipment eCAT allows automated test case creation at negligible additional costs.