Generalise notion of groups with targets?

[MattWindsor91]

After the most recent metamodel revision, a Target has become the single unambiguous RoboStar object that a sequence describes, and the Actors reduced to references to things accessible from that Target. This said, sequences are not the only thing that can/should have a Target:

• Activity diagrams and any other diagram form we add will need Targets for obvious reasons;
• CSP groups should be able to have Targets (which would need to be able to reference targets of other things, I think?), so that their tick-tock contexts can be specified centrally rather than once per CSP fragment;
• It'd be nice to have a notion of a group that exists only to define a Target and its Instantiation, which can then be used as the basis of other groups' targets.

Perhaps this is overengineering, which is why it's a question and not being implemented yet!

[MattWindsor91]

This has become SpecGroup. Currently, CSP groups are technically SpecGroups but not yet properly integrated (eg, they don't use their Targets) - this'll need a separate issue filing as there are design decisions to make.