It is a truism that software testing is never complete. Testing can only confirm the presence of bugs but says little about the defects that undoubtedly remain, even in the shipped product. However, no one is throwing up their hands and saying that we should forget about testing because it is ultimately a futile effort. In fact, the opposite is true, since it is an activity not of absolutes but of probabilities.
The more testing that is accomplished, the higher the probability that customers will be pleased with the final product. No one expects any software to be completely bug-free, but they do expect it to be 99.5 percent useable, to perform well, to not crash, to retain their data and to play nice with other apps.