Most of the complexity theoretic research on crypto-protocols has concentrated in two main directions. Developing general protocols to compute any multiparty function, and developing efficient protocols for specific functions. The general approach sharpened understanding of what is the exact meaning of 'secure' protocol, and resulted a rich new area of complexity theory, but in most cases resulted solutions which are currently too complex for any real usage. The second approach, which may be termed, the 'applied' approach benefited from the insight and the beauty of those new theories, but in most cases had to go its own way, not only in dealing with specific functions, but also in many cases in dealing with specific restricted adversaries (rather than 'any random polynomial time adversary'). The author illustrates some examples from these areas.
展开▼