Some methods are able to be implemented using existing methods, but just in a non optimized way. For example removeAll can be implemented using remove but just invoking remove for each key. Obviously this is not the long term preferred option, however it can get parts of the API working now and be able to test the behavior.