Heterogeneous-language distributed objects with hidden IDL

A major feature of modern distributed programming is the notion that the components of a distributed program need not be implemented in the same programming language. This heterogeneous-language distribution allows individual distributed components to be implemented in the language which best suits that particular component's functionality. For example a display component could be written in Java, utilizing the platform-neutral Java API, while a component responsible for a portion of a real-time simulation could be written in a language better suited for that purpose, such as Ada.;Previous heterogeneous-language distributed systems have been developed which either use an explicit Interface Definition Language (IDL) or provide direct mappings from constructs in one language to constructs in another. IDL-based approaches require that component implementations match the output of an automated mapping from IDL to implementation language. This automated mapping may not represent the most natural implementation in the component's language. The language to language mapping approach allows server components to be cleanly implemented in their language. The implementation of the server component is the focal point of the translation process, rather than an outside definition language representation of the server component. The disadvantage of this language to language mapping approach is that it does not scale well for multiple languages.;This research focuses on developing a new paradigm for heterogeneous-language distribution which blends the advantages of the IDL and direct language approach. In this research an interface definition is used as an intermediate step in the translation process from server language to client language. This approach retains the server implementation as the focal point of the translation process while still using a single interface definition language representation. The use of a single interface definition language allows the approach to scale for multiple implementation language support.