Practical expressiveness of internal and external domain-specific modeling languages

Abstract

During the long history of modeling and simulation, many answers have been given to the question of how to specify simulation models. Many of these approaches can be perceived as domain-specific modeling languages offering a syntax and a semantics. However, the individual languages are often vastly different. A central distinguishing aspect is the classification as external or internal domain-specific language. External and internal domain-specific languages are characterized by specific trade-offs regarding syntactical flexibility, computational efficiency, and amount of implementation work. We present a case study of alternative approaches to implement domain-specific languages for a small modeling problem in supply chain management. We illustrate the influence of using an external or internal language on different aspects of language performance, in particular the practical expressiveness, which we identify as one of the central properties of modeling languages.