provided that that copies bear this notice and the full citation on the first page. A Research Agenda for Mainstreaming Service-Oriented Software

A service-oriented approach to software has the potential to provide a better model for software components than a strictly object-based model. The service-oriented client is focused directly on the services it needs instead of on the class structure of objects that implement these services. This results in much more modular software. But the service-oriented model is in danger of remaining a “niche ” specialty for software developers. To bring the advantages that can be derived from it into use by mainstream software developers, we need to be able to teach it in terms of conceptual roots, rather than as a coding style couched in terms of objects. Doing so requires that we clarify, separate, and exploit the concepts that are fundamental to a focus on services, and express them in programming-language terms. This paper outlines research issues to be addressed in doing so. 1