Add id field to MetamodelElement and Metamodel

[enikao]

We want to be stable against renaming of Metamodels, Concepts, Features etc.
Thus references to them should use an id, not their name.

This is not an implementation detail, as these ids would be used in the protocol.

Relates to #31.

[ftomassetti]

From my point of view would be important to support both the models obtained from projectional editing and from parsing. In the first case, IDs could be generated, while in the second case should be stable, maybe based on the position of the node on the tree, maybe on the name of entity (e.g., the name of the class, the name of the function, etc.).

Should ID be strings? Should ID be just an interface with possibly multiple implementation? I think this is a topic with a lot of implications and we should discuss it during one of the next meetings

[dslmeinte]

I think they should be strings, because they ought to be serializable without fuss. Anything more complicated, and chances are that other concerns (such as namespace identification and such) are leaking into their purpose.

[enikao]

Shall we have this discussion at #31? I think that's more suitable.