Specifying graphical modeling systems using constraint-based meta models

Embedded computer-based systems are becoming highly complex and difficult to implement due to the large number of concerns designers must address. These systems are tightly coupled to their environments, requiring an integrated view that encompasses both the information system and its physical surroundings. Mathematical analysis of such systems requires formal modeling of both "sides", including their interaction. There exist a number of suitable modeling techniques for describing both the information system component and physical environment, but the best choice changes from domain to domain. We propose a two-level approach to modeling that introduces a meta-level representation. Meta-level models define modeling languages, but they can also be used to capture subtle interactions between domain level models. We show how the two-level approach can be supported with computational tools, and what kinds of novel capabilities are offered.