Preliminaries
Let us assume that we have a smarter tester C who generates tests
using the formal specifications for W. C then tests W and evaluates
the code coverage using one of the several code-coverage criteria.
C finds that the code coverage is not 100%, that is, some parts of
the code inside W have remained uncovered, that is untested, by
the tests generated using the formal specification. C then generates
and runs additional tests with the objective of exercising the uncov-
ered portions of W. We say that C has performed both black-box
and white-box testing. C has used specification-based test generation
and enhanced the tests so generated to satisfy some control-flow-
based code-coverage criteria.
Now suppose that tester D tests W as a component of ...