Key points are not available for this paper at this time.
When large programs are tested, a distinction is made between module testing and integration testing. During module testing, one concentrates on testing algorithms and their data, whereas during integration testing, one concentrates on testing the relations and the interfaces between the modules. The authors use the well-known criteria for module testing to define similar criteria, based on control and data flow, for the purposes of integration testing. An advantage is that it is possible to use strong test data selection criteria, such as all-loop-iteration-free-paths, when testing a single module because, generally, a module will be small. For integration testing, data flow criteria are especially useful since the aim of integration testing is to show the correctness of the interfaces between the modules which, in the case of procedures, are determined by the data flow between the calling and the called procedure.>
Linnenkugel et al. (Wed,) studied this question.