What do we mean when we say some piece of cryptography is secure? What are cryptographic assumptions? How does modelling security fail in practice?