Sequence diagrams, systemsuse actors or components

[newarchitect]

Hi all,

Looking for some advice please. I've modelled some of our systems using UML components. I also have some actors (e.g. customer etc.) which I use on some of the component diagrams. These actors may be reused for modelling use cases in the future.

I want to create some sequence diagrams. I can use the existing actors but just wanted to know what I should be doing for systems. Should I use the existing systems (which are components) or create actors for the systems as well?

[qwerty]

Create lifelines and link them to the existing components (use Ctrl-L to select the classifier).

q.

[newarchitect]

Cheers qwerty

[newarchitect]

Sorry, silly question. What is the difference of dropping the component into the diagram as a lifeline as opposed to creating new lifelines and linking them to the components?

[qwerty]

Dropping an element onto a diagram will usually create a link and not an instance (except you changed the setting or apply that explicitly on drop). Whilst creating a lifeline creates a new instance which is then linked afterwards (so it's more safe doing it this way).

Now the un-asked question: why do I need instances? Well, it's hidden somewhere in the documentation. If you use XMI ex-/import links will clobber your sequence diagram and instances won't. Unfortunately EA does not hinder you to use non-instances in SDs  

q.