Using a model-based test generator to test for standard conformance

test generation. The process of test generation is automated by GOTCHA, which explores the state space described by the GDL model. The user has several alternative test generation strategies, including breadth-first search, coverage-directed search, and on-the-fly test generation. Breadth-first search and on-the-fly test generation algorithms are well known. Coverage-first search involves giving priority to exploring states that lead to new coverage tasks before those that lead to areas of the projection state space that have already been encountered. FARCHI, HARTMAN, AND PINTER IBM SYSTEMS JOURNAL, VOL 41, NO 1, 2002 98 The principle that underlies GOTCHA’s test generation strategy is the construction of a search tree that explores the entire state space. This is done by traversing all the reachable states of the state machine. The set of coverage tasks is constructed by observing each instance of a projection state or transition that satisfies a coverage criterion. An on-line randomization algorithm chooses a reachable representative of each coverage task encountered. GOTCHA performs a further reachability analysis, starting from the randomly chosen representative, to determine if a TC_EndTestCase condition can be reached from the specific instance of the coverage task. If a test that satisfies all the test constraints exists in the state machine graph, then the test case or path is output to a file in an XML format for describing paths. If no TC_EndTestCase condition can be reached from a particular instance of a coverage task, another representative of the task is chosen. If no TC_EndTestCase condition is reachable from any reachable representative of a coverage task, then the user is notified of the fact that a reachable task has been identified with no test cases satisfying the test constraints through any of its representatives. The abstract XML test suite comprises the following elements: ● The name of the model, for example,