he Action Semantics proposal, currently being standardized by the OMG, aims at integrating semantics of an action language into the UML 2.0. The Action Semantics are programming language independent. They define operations that support the synchronous manipulation of objects, the generation and handling of signals, and the logical constructs that support the specification of algorithms. The modeler specifies these operations using a text-based action language. Various UML tools support an action language. They have been used for a long time in the embedded and real-time modeling communities, such as in the Specification and Description Language (SDL) community. The use of an action language makes UML models executable, i.e. it allows designers to test and simulate models and to fully automatically generate executable code.

Opponents of the Action Semantics proposal argue that Action Languages are turning UML into a full-fledged programming language – which defeats the purpose of modeling.  Yet, action languages are not intended to compute algorithmic functions. Low-level algorithmic or hardware interface functions are specified outside the model and used as modeling primitives. The model provides different views on how these primitives are assembled to form an application. Action languages are rather used to glue these primitives together into executable views of the system, which can be separately tested and verified.

The authors’ position is that the use of an action language within UML statecharts makes a more powerful form of AOM possible. A joinpoint model for statecharts can capture not only particular object state transitions, but also the individual actions executed during a transition. Such a joinpoint model is richer than the joinpoint models of AOP languages. Hence, AOM has the potential to address some of the shortcomings of AOP.