Experimental comparison of three system test strategies preliminary report
John H. Rowland, Y. Zuyuan · 1989
This paper reports on some experiments which compare three system test strategies which were applied to artificial software systems. The artificial systems were seeded with known defects and then tested according to the three strategies which involved random system level input, random system level input with defect removal, and modified module coverage. The results suggests that of the three strategies considered here, the modified module coverage exposes a higher number of defects, especially for large systems.