It is a well-known fact, that a good test is one that has an acceptable probability of discovering an error and that it is impossible to perform an exhaustive input testing of a program. Therefore, testing a software program you are limited to using a small subset of all possible input data. In this case, without doubt, you would want to choose the most suitable subset which is a subset with the highest probability of discovering the majority of errors.
One way of identifying the right subset is to understand that a really good test-case additionally has two characteristics:
- It covers a significant part of other possible test cases, which to some extent indicate the presence or absence of errors before and after applying this limited set of values of the input.
- It reduces by more than one, the number of other tests that must be designed to achieve a pre-defined goal of “acceptable” testing.
These properties, despite they seem to be alike, describe two quite different positions. First, each test should evoke as many different input conditions as possible, in order to minimize the total number of tests required. Secondly, you need to partition the program input region into a finite number of equivalence classes so that you will be able to assume (certainly, not absolutely confidently) that each test – representative of some class is equivalent to any other test of another class. Put it differently, if one test in the equivalence class finds an error, then we should expect that all other tests in this equivalence class will find the same error. Vice versa, if the test fails to detect an error, then we should be ready that no test in this very equivalence class will detect errors (in the event that some subset of the equivalence class doesn’t fall within any other equivalence class, because equivalence classes happen to overlap each other).
These two considerations make up the basis of the black box methodology, called also as the equivalence partition. The second consideration helps to develop a set of “attractive” conditions that must be tested, while the first – to develop a minimum set of tests to cover these conditions. If you want equivalence testing to be performed to a high professional standard, turn to Cherkassian offshore software testing team and they will exceed all your expectations concerning the workmanship.
If an error is not detected by a test of an element of the given set, then it is unlikely that the error will be detected by a test from another set. Simply put, in this case, testing time is better spent on something else (in other equivalence classes).
And, finally, test-case development by equivalence class partitioning goes through two steps:
1) the identification of equivalence classes;
2) the definition of test cases.
Add Comment