Databases Reference
In-Depth Information
the most important components in a testing environment is test data
generator. Manually generating test data is too costly as it takes maximum
time in testing phase. So, the solution to reduce the high cost of manual
software testing is to make the test data generator automatic. It is found
that automatic test data generation has achieved some success by using
evolutionary computation algorithms such as genetic algorithms. Basically,
a test data generator system consists of three parts:
program analyzer
,
path
selector
and
test data generator
. The source code is run through a program
analyzer, which produces the necessary data used by the path selector and
the test data generator. The job of path selector is to inspect the program
data and find the paths leading to high code coverage. For the test data
generator, the input is a path and outputs are the test data that exercise
the given path. In this chapter, we focus on
path selectors
and
test data
generators
. Some test data are more proficient at finding errors than others.
So, testers have to choose the tests carefully to generate a good test suite. A
proficient test case is one that has a high probability of detecting an as-yet
undiscovered error.
One test case is not sucient to satisfy all the test requirements.
Usually, a suite of test cases is required to satisfy all known possible
requirements. In every step of testing, a new test case may be generated and
added to the test suit. Thus, a test suite may undergo a process of expansion
in practice. Most of the times, it is found that a test suite contains more
than enough test cases for satisfying the target test requirements. So, some
test cases in a test suite are redundant. If those redundant test cases are
removed from the test suite, the resultant test suite may still satisfy all the
test requirements that can be satisfied by the original test suite. A good test
suite is one which contains no redundant test cases. Therefore, finding a sub-
suite of an existing test suite that can satisfy the same test requirements as
the original test suite becomes a research problem. This problem is usually
referred to as test suite reduction and the acquired sub-suite of test cases is
called the representative set. If no subset of a representative set can satisfy
all the requirements, the representative set is called optimal representative
set or minimum representative set. As the aim of test suite reduction is to
minimize the test suite, it is also referred to as test suite minimization. An
optimal testing strategy selects the best test cases each time during testing.
Since a large number of test cases can be designed, and an effective test
suite needs to choose out of these, test case design can be considered as
a typical search problem. A search algorithm must decide where the best
