Testing Techniques

Overtime the IT industry and the testing discipline have developed several techniques for analyzing and testing applications.

Black-box Tests
Black-box tests are derived from an understanding of the purpose of the code; knowledge on or about the actual internal program structure is not required when using this approach. The risk involved with this type of approach is that .hidden. (functions unknown to the tester) will not be tested and may not been even exercised.

White-box Tests or Glass-box tests
White-box tests are derived from an intimate understanding of the purpose of the code and the code itself; this allows the tester to test .hidden. (undocumented functionality) within the body of the code. The challenge with any white-box testing is to find testers that are comfortable with reading and understanding code.

Regression tests
Regression testing is not a testing technique or test phase; it is the reuse of existing tests to test previously implemented functionality--it is included here only for clarification.

Equivalence Partitioning
Equivalence testing leverages the concept of "classes" of input conditions. A "class" of input could be "City Name" where testing one or several city names could be deemed equivalent to testing all city names. In other word each instance of a class in a test covers a large set of other possible tests.

Boundary-value Analysis
Boundary-value analysis is really a variant on Equivalence Partitioning but in this case the upper and lower end of the class and often values outside the valid range of the class are used for input into the test cases. For example, if the Class in "Numeric Month of the Year" then the Boundary-values could be 0, 1, 12, and 13.

Error Guessing
Error Guessing involves making an itemized list of the errors expected to occur in a particular area of the system and then designing a set of test cases to check for these expected errors. Error Guessing is more testing art than testing science but can be very effective given a tester familiar with the history of the system.

Output Forcing
Output Forcing involves making a set of test cases designed to produce a particular output from the system. The focus here is on creating the desired output not on the input that initiated the system response.