Object Imaging

• Authors: Satish R. Thatte
• Venue: ECOOP '95 Proceedings of the 9th European Conference on Object-Oriented Programming
• Year: 1995

As heterogeneous and distributed object systems become more common, interoperability of system components is becoming a pressing issue. One important challenge in making such component-based systems work is support for object transport. Earlier data transport standards such as XDR focused primarily on marshalling RPC parameters and results for transport between different processor architectures. Effective use of the potential benefits of heterogeneous distributed object systems will require object transport at many levels of abstraction and granularity, ranging from small objects exchanged via RPC to large multimedia documents transferred across environments via the world-wide web. To emphasize the transformations involved, we refer to the problem of producing copies of objects across different environments as object imaging. In this paper we propose a very general notion of object imaging in which a flexible knowledge-base in the form of a dual logic of conversion and conformance is used to represent elements of imaging knowledge as composable combinators. This approach has the potential to seamlessly cover the entire range of imaging problems from simple primitive datatypes to complex object structures, and supports a variety of notions about conformance of object and interface types, including the multiple-inheritance single-interface object model of CORBA and the inheritance-averse multiple-interface model of COM.