Ilias Lazaridis wrote:
>
> There is a relation, but not an OO 'instance of' one.
>
> Concretely, in the diagramm V1.3 this means:

The answer is really very simple, and much as I loathe to repeat it again 
(well not really)

UML is broken!

If it cannot model Ruby correctly then it cannot be "unified".
Bug Booch and Rumbaugh about their arrogance.

ObNotetoHonorableMrMatsumoto:
Not sure if it was a typo but "mataclass" is perfect.

--
J Lambert