On the Appropriateness of Domain-Specific Languages Derived from Different Metamodels

In model-driven development domain-specific languages (DSL) are often considered models while the description of DSLs are expressed using various metamodels. To estimate the influence of a metamodel on the quality of DSLs derived from it, it is appropriate to measure functional suitability. As defined by the standard ISO/IEC 25010 (SQuaRE), functional suitability consists of completeness, correctness, and appropriateness. Among these issues, only appropriateness can be evaluated without specifying the domain. This paper is a study of a relationship between (a) the metamodel's expressive power regarding the syntax of the DSLs derived from the metamodel and (b) the appropriateness of those DSLs. In this regard two metrics are defined. The first metric evaluates a metamodel and produces the estimation of the derived DSLs' appropriateness. The second metric incorporates the domain and further assesses the quality of a DSL in terms of appropriateness. Both metrics are based on abstract syntax trees of programs written in the derived DSLs, and demonstrated using examples on two different domains.