In this survey paper we consider the class of protocols for information-hiding which use randomization to obfuscate the link between the observables and the information to be protected. We focus on the problem of formalizing the notion of information hiding, and verifying that a given protocol achieves the intended degree of protection. Without the pretense of being omni-comprehensive, we review the main approaches that have been explored in literature: possibilistic, probabilistic, information-theoretic, and statistical.