5.3 BLACK BOX—BASIC TESTING
5.3.1 Boundary Value Testing
Boundary value testing is a method to verify the behavior of systems at operating boundary areas by selecting test data values that lie at operating extremes. Boundary test values may include maximum or minimum values within the normal operating domain, values just inside and just outside operating domain boundaries, typically encountered operating values or specific error condition operating values.
The objective of this method is to test systems at boundaries of the operating domain where a substantial number of errors tend to concentrate. Generally, this method is applicable to software, embedded systems and systems that contain some software components. The weakness of boundary value testing is that the testing process is not exhaustive and the method is not appropriate for complete validation of a system.
The boundary value testing method is based on selecting test cases within sets of equivalence classes at the “edge” of the class rather than selecting any element at random. As a result, this method facilitates a possible reduction in the number of test cases relative to the number of detected errors. In summary, the system is not fully validated but a high proportion of errors can be found. The method entails two-step operation: (1) defining equivalence partitioning and (2) generating and executing test cases at extreme ends of equivalence classes.