SharpLudus revisited: from ad hoc and monolithic digital game DSLs to effectively customized DSM approaches

This paper describes our experience in improving an ad hoc approach for creating domain-specific languages targeted at digital games, replacing it by a customization of more structured approaches in the domain-specific modeling literature. We give special focus on the benefits of partitioning the target game domain into prioritized sub-domains, as well as on promoting game engines to domain frameworks that can be more seamlessly consumed by generated code. A case study for the arcade games domain is also presented for illustration and evaluation purposes.