Boundary and typical values in Programming:-

A systematic plan is needed for choosing the values to be used in test cases. One possibility would be to just choose random values4, but this could lead to carrying out too many or too few tests and could mean that the testing takes
far too long or overlooks some important test cases. Something much more systematic is required. A system is needed that will allow testers to choose a small enough number of test cases so that they can actually carry out the test in a reasonable amount of time, but which still gives them enough test cases to allow them to be confident that a program is correct.

There are basically two sorts of expected values that can be input to a program. Typical values are simply examples of those that the program is expected to process when it is running normally; testing with these is just a case of choosing a collection of typical values and assuming that if the program processes these all correctly it will also correctly process all other similar values.

Boundary values are more complicated. These are those that occur around a limit on the acceptable values. The program should be tested with all boundary values and also the two values immediately surrounding each boundary. To help the tester, a test plan should also indicate the purpose of a particular test case – whether the value is typical or boundary.