Programming Language Interoperability in Distributed Computing Environments

Abstract

Distributed computing environments, such as ANSAware, CORBA, DCOM, and DCE achieve software component interoperability through specifying all public component interfaces in a common interface definition language. Programming language interoperability is achieved by mapping these interface specifications to the programming language of choice. The language mapping defines the representation of an a priori agreed upon set of data types in the target programming language. The developer is restricted to this convention. This is, however, just the first step in achieving programming language interoperability. By focusing on the CORBA standard we illustrate this mapping procedure. We develop and thoroughly analyze several design schemas for implementing cross-language method invocation with CORBA. From this discussion it will become clear that the current standard lacks support for user extensible programming language interoperability. The CORBA standard does not allow to efficiently and effectively interface arbitrary programming languages to CORBA middleware. Point-solutions are possible, however, portable implementations of language mappings are not. We propose an addition to the current standard which would alleviate this problem and provide sufficient interfaces for other ORB extensions, such as request monitoring, debugging hooks, and costum marshalling, among others.KeywordsProgramming language interoperabilitylanguage mappingsIDL-to-Sather language mappingmiddlewaredistributed computing infrastructureCORBA