Ownership of association ends by an associated classifier may be indicated graphically by a small filled circle (aka dot).
What impact would it have, in terms of code generation, if an association end belongs to the classifier or to the association itself?
That just means that the owned end represents a property typed with the owned classifier within the owning classifier. It is identical to having the owned property shown in the attributes compartment of the owning classifier. However, showing a association with both classifiers clarifies visibly the connection between both classifiers.
The dot notation is new with UML 2.5. It had been introduced to make ownership clear which in former UML versions was shown more wishy-washy using a navigation which actually tells nothing about ownership.
