Can you test Completely ? When to Stop Testing ?

t's impossible to fully test a program. In the face of an infinitely large testing task, we must be skeptical of statements that some people make that the testing project must always do this or always deliver that. In the face of an infinitely large task, everything is a tradeoff--work spent on one task is work not allocated to another. Wise allocation of resources to tasks and deliverables must be a function of the information objectives of the project at hand. Several metrics appear to check how much testing we have done or how much is left.

Coverage measures are an example; so are defect arrival rate probability models. If "complete testing" means that there are no remaining unknown bugs, then these approaches cannot measure completeness of testing. Instead, they must mean, complete according to some artificial criterion. There are predictable risks of using these metrics. People who rely on them often distort how the project is run in ways that often yield worse testing