Semantic Interoperability

The issue of semantic interoperability pervades distributed systems whose components are designed independently. For example, when a source intends 70 to mean “70 meters, with 20% error”, the receiver must not interpret the result as “70 feet, exactly”. We examine semantic interoperability problems in a distributed object management environment, i.e., for a system of distributed objects and services rather than for a database. A semantic interoperability service is defined, and its required functional components argument describers, conversion functions, and a planner are identified. We categorize levels of service that each component may provide, and how these choices affect system functionality and evolution. We also begin to examine the various ways to map the functional architectures to software components – sources, receivers, mediators, and request broker – and data administration tasks. Finally, we contrast with semantic interoperability for database clients, and discuss why the type system should not be the repository for conversion knowledge.