Prof. Dr. Jesse H. Poore

"Use of mathematical models in specification and testing"

The presentation will review a project conducted at the Fraunhofer IESE that worked systematically from a requirements document to fully automated testing. The project made use of two mathematical models. The sequence-based specification method uses automata theory to elicit a finite state machine from informal requirements. The method resolves ambiguities, errors and omissions in the requirements and produces a specification that is complete, consistent and traceably correct. A Mealy machine follows directly from the specification and its graph forms the structure of Markov chain. With probabilities on the exit areas of the directed graph, we have a Markov chain, which is the basis for statistics for test planning. The Markov chain is used to generate cases which drive a test runner. The test runner executes and evaluates test cases. Thus we automatically generate, execute and evaluate test cases. A second Markov chain records testing experience and is the basis for statistical analysis of the results of testing.