Our new article on fault propagation and its relation to model-based test-case generation was just published:
The publisher Elsevier provides a free download of the original article until December 2.
The paper describes how we generate test-cases for reactive systems by
We show that this is more efficient than previous approaches that search for observational equivalences directly, i.e. skipping Step 2.
Our case study of a car-alarm controller shows that for some subtle faults we need up to nine additional interactions until a fault becomes visible at the interface.
The paper describes how we generate test-cases for reactive systems by
- injecting faults into the models (mutation testing)
- generating a test-case that triggers this fault (refinement checking)
- extending the test-case until an observational failure could be produced (ioco checking).
We show that this is more efficient than previous approaches that search for observational equivalences directly, i.e. skipping Step 2.
Our case study of a car-alarm controller shows that for some subtle faults we need up to nine additional interactions until a fault becomes visible at the interface.