Towards the practical adoption of LIDL: a toolchain for modeling human-machine interface software interactions

Abstract

The design and implementation of safety-critical human-machine interface (HMI) software typically follow the model-driven development process to guarantee secure reliability pragmatically. Systematic and comprehensive modeling for interactions involved in HMI software is essential when applying the model-driven approach. LIDL Interaction Description Language (LIDL) is an interaction description language well suited to HMI software interaction modeling. Several problems limit the practical adoption of LIDL. LIDL does not have a visual modeling language or debugging tools, resulting in error-prone and inefficient modeling. In addition, there is no toolchain to connect it to the subsequent steps in model-driven development, such as verification and implementation. This paper presents a toolchain developed for LIDL to alleviate the problems it faces in practice. The toolchain consists of three tools, a visual modeling tool, a compiler translating LIDL code to Lustre, and a LIDL debugger. We design the visual modeling tool and debugger to improve LIDL modeling efficiency and the compiler to support the debugger and integrate the LIDL modeling into the whole model-driven development process of the HMI software. Specifically, the modeling tool uses tables and diagrams to visualize LIDL code, pragmatically reducing the possibility of introducing modeling errors while maintaining the semantics of LIDL code. We add management features to LIDL models to form a library of model components that modelers can reuse easily. The modeling tool can also automatically convert LIDL models created by modelers into LIDL code. The debugger helps modelers debug LIDL code to locate problems in the model and correct them quickly.