Visual data flow languages with action systems

Due to recent advances in multicore processor technology, parallel computing has received much attention lately. In this context, the need for a natural parallel execution paradigm makes a visual data flow programming language (VDFL) attractive. VDFL programs are mainly used in measurement and controlling tasks where accuracy and robustness of the system are fundamental. The verification of these systems are, however, difficult and testing processes cannot fully expose all the potential errors. This problem is typically tackled with formal methods, including formal specification, model-checking, and refinement. In this paper, as the main contribution, we formalize the basic concepts of VDFL with action systems. We show that there is a natural mapping between VDFL and action systems that supports future work on reasoning and refinement of VDFL programs using action systems.