Real-world services-that is, non-software-based services-differ significantly from Web Services, usually defined as software functionality accessible and configurable over the Web. Because of the economic, social, and business importance of the service concept in general, we believe it's necessary to rethink what this concept means in an ontological and computational sense. Most current semantic approaches focus on the computer science and software aspects of services insofar as they lie within the Web environment. Such a view is incomplete and unnecessarily limiting, because current Web Services ignore significant elements of the service context. A Semantic Web approach can-and should in our view-represent and reason about what we call value webs: services that have a grounding in the real world, beyond their Web elements and references alone.The OBELIX (Ontology-Based Electronic Integration of Complex Products and Value Chains) project has therefore developed a generic component-based ontology for real-world services. This OBELIX service ontology is first of all a formalization of concepts that represent the consensus in the business science literature on service management and marketing. In addition, our ontology embodies a number of systemstheoretic notions that specify how diverse service elements, seen as individual "Lego blocks," can connect to each other to form a larger service system: a service bundle. Here, various topological connection and typing rules play a key role, similar to the design of complex engineering systems.Furthermore, we express our service ontology in a graphical, network-style representation, and we've developed support tools that facilitate end-user modeling of services. Then, automated knowledge-based configuration methods let business designers and analysts analyze service bundles. We've tested our ontology, methods, and tools on applications in realworld case studies of different industry sectors.
An interdisciplinary approachWe believe that ontology research and application can cover and handle more than current semantic approaches do. First, Semantic Web Services research still lacks convincing realistic industrial use cases that are the basis of practical use and empirical validation of ontologies and problem-solving methods.Second, ontology research and application requires an interdisciplinary approach to services that draws not only from computer science and AI but also from economics, systems theory, and business practice. Such an approach will lead to ontological descriptions that capture much richer service profiles, even of hitherto unseen services. For example, the need exists to more clearly distinguish requester-(customer) and supplier-oriented service descriptions. Suppliers and customers have different roles in, and viewpoints on, a service, which leads to different ontological commitments and descriptions. Furthermore, service industry practice shows that "nonfunctional" aspects (such as quality features) are empirically important to service selection, compositi...
