BIBLIOGRAPHIC NOTES
adequate with respect to the all-uses
criterion.
In Frankl et al. study, it is important to
note that the reported difference is small
and becomes apparent only at very high
levels of mutation adequacy. That mutation
is more effective than data flow does not
imply that one should use one technique
over the other. In fact, it is recommended
that tests that are adequate with respect to
data flow based or any other path-oriented
code-coverage criteria must be evaluated
against a mutation-based adequacy criteria
for further enhancement.
Researchers have often used mutation as
a means to study the fault-detection
effectiveness of other test-adequacy
criteria. Daran and Theévenod-Fosse [108]
and Andrews et al. [23] have studied the
representativeness ...