Web services are software components that can be discovered and employed at runtime using the Internet. Conflicting requirements towards the nature of these services can be identified. From a business perspective, Web services promise to enable the formation of ad-hoc cooperations on a global scale. From a technical perspective, a high degree of standardization and rigorous specifications are required to enable the automated integration of Web services. A suitable technology for Web services has to mediate these needs for flexibility and stability. In this paper a new approach to the description of Web service semantics is introduced. It is a visual approach based on the use of software models and graph transformations and allows for the description of innovative services while providing a precise matching concept. An implementation using current standards and tools is available.

In object-oriented software development, requirements of different stakeholders are often manifested in use case models which complement the static domain model by dynamic and functional requirements. In the course of development, these requirements are analyzed and integrated to produce a consistent overall requirements specification. Iterations of the model may be triggered by conflicts between requirements of different parties.However, due to the diversity, incompleteness, and informal nature, in particular of functional and dynamic requirements, such conflicts are difficult to find. Formal approaches to requirements engineering, often based on logic, attack these problems, but require highly specialized experts to write and reason about such specifications.In this paper, we propose a formal interpretation of use case models consisting of UML use case, activity, and collaboration diagrams. The formalization, which is based on concepts from the theory of graph transformation, allows to make precise the notions of conflict and dependency between functional requirements expressed by different use cases. Then, use case models can be statically analyzed, and conflicts or dependencies detected by the analysis can be communicated to the modeler by annotating the model.An implementation of the static analysis within a graph transformation tool is presented.