The mining of models from data finds widespread use in industry. There exists a variety of model inference methods for perfectly deterministic behaviour, however, in practice, the provided data often contains noise due to faults such as message loss or environmental factors that many of the inference algorithms have problems dealing with. We present a novel model mining approach using Partial Max-SAT solving to infer the best possible automaton from a set of noisy execution traces. This approach enables us to ignore the minimal number of presumably faulty observations to allow the construction of a deterministic automaton. No pre-processing of the data is required. The method's performance as well as a number of considerations for practical use are evaluated, including three industrial use cases, for which we inferred the correct models.

Domain-specific languages (DSLs) are popular for many reasons, such as increasing productivity for developers and improving communication with domain experts. Both textual and graphical DSLs are viable solutions with complementary pros and cons: while graphical DSLs shorten the learning curve and facilitate documentation and communication, textual DSLs aim at higher productivity thanks to more efficient editor functionalities. This paper presents the industrial experience on the adoption of a hybrid approach combining an existing textual DSL with a read-only graphical state machine representation (visualization), equipped with a selective abstraction functionality that offers user-specific, highly configurable views on states and transitions. Our approach is the result of an evolutionary process to improve the modelling experience, relying on frequent user feedback. We argue that a well-tailored visualization is a suitable way to shorten the learning curve and ease the adoption of model-driven approaches in industrial settings.