linking use-cases and components

[ngong]

What is the best practice to express that components shall be aware of certain use-cases?

What I did for a quick solution, is to create a composite diagram for the use case (that is not detailed any further) in order to paste the regarding components into it as Simple Links and surrounded them with a Boundary named "Affected Components".

Is there a better or more-common method?
Will it be possible to get a report, inverting that information: What use-cases affect component XY?

Any comment welcome.

[Paolo F Cantoni]

Hi Rolf,

FWIW, I'd put a dependency between the Component (client, origin) and the Use Case (supplier, destination), stereotyped «affected».  That way you have a more amenable relationship that can be queried more effectively.

An added bonus is that both ends of the relationship can be obtained using the add related components (or using my Neighborhood Diagrammer Add-In).

HTH,
Paolo

[ngong]

Yes, thank you Paolo, your suggestion seems to be much better.

And it is easy to convert from my first attempt:

I'll keep the composite diagram (explaining the use-case in more detail), drop the use-case there (simple link), use the dependency "realize" and stereotype it by "affected".

This will tell me that the components realization is affected by the existence of the use-case.

Or to put it in more strict words: "If a component is not affected by a use-case, any change to the use-case will not affect the realization of the component in no way."

 

[neil_albiston]

Your Neighborhood Diagrammer Add-In sounds interesting. Is it available to the  community or is it commercial?

I've just started working on an addin to merge multiple diagrams. Happy to share the final results but it is at a very early stage at the moment.

Neil