Say I need to model two use cases: "do something" via a batch file interface and via a web service call (UC1 and UC2). Batch interface will "do something" with each record in the file, while web service "does something" only once. Each UC has different primary actor and slightly different set of input/output data of "do something".
Both UCs (UC1 and UC2) share common core "do something" logic and this logic includes interactions with other, participating actors.
Do I understand correctly that the common core "do something" logic should be modeled as a separate UC
<<include>>d from UC1 and UC2?
Do I understand correctly that the common core "do something" UC will have no primary actors? It will only describe interactions with the other actors, right?
How do I model the interface of the "core" UC i.e. its input and output? Normally we do not describe what UCs receive at input, right? All UC templates I've seen have attributes like: "name", "abstract", "actor", "scenarios" etc. but not "input" and "output". If it is important to describe what data the primary actor provides, this is reflected in UC scenarios as interaction steps with the actor, right? But in case of my "core" UC there are no such interactions with the invoking actor.