UML and use case modeling have the last few years received much attention. Use cases are used to describe the functional requirements of a system and a use case diagram shows the relationships among use cases and actors within a system. The semantics of use case diagrams are, however, often unclear. The aim of this dissertation is to point out ambiguities and limited expressiveness present in the current use case notation of UML and to indicate in which directions remedies to these deficiencies may be sought.

Problems reported by researchers were identified and presented to a set of use case practitioners. Then a proposal of possible improvements of the use case notation, which should reduce the ambiguities and increase the expressiveness in the notation, was developed. To place some measurements on the value the proposal can have in practice, it was also presented to a set of practitioners.

The proposal provides a suggestion of how to model various tricky situations in a use case diagram. Furthermore, the dissertation gives insight on how the use case technique is used in practice. Practitioners argue that the use case specifications are more important than the use case diagrams. They also argue that the purpose of use case diagrams is to describe a system's main functionalities and not details, and that as few use case relationships as possible should be used in use case diagrams.