Transforming Abstract Interpretations by Abstract Interpretation - New Challenges in Language-based Security