Summary

In an object-oriented system the interactions between objects provide the main structure of the program. By testing specifically for problems between objects as they are integrated into larger and more complex objects, problems involving checking preconditions and the sufficiency of objects returned as the result of a message are discovered early. There are many factors that influence each interaction. Statistical sampling techniques such as OATS provide a means of selecting an effective subset of combinations of these factors to investigate using test cases.

The design of test software is influenced by the design of the software it is intended to test. In an object-oriented software development environment, the design of the production software is guided by a set of standard patterns. By discovering and documenting the standard ways in which test objects interact with each other and with the software under test, less experienced testers can benefit from the knowledge of more experienced testers. The result is test software that is of a better quality and is more reusable.