I'm currently writing up a design for a digital reputation capital system. The intent is not to provide a framework for licensing or formal endorsement system, but instead, allow people to automatically discover the opinions of others about various entities. I'd like to know how much people would want anonymity in a system like this. My preferred solution would be to allow anonymity through the established services of remailers. This has the advantage of having people who use a nym constantly (and well) get more respect when doing it then those who use a nym occasionally. The reason this solution is preferred is that it allows a fully distributed system to exist, with no centralization needed at all. Is losing that distributed characteristic of the system worth gaining a system that supports anonymity? (It might be possible to design a work intensive system to handle distributed anonymity, based on Merritt's protocol for voting without any central facility (Applied Crypt section 6.5), but the amount of work involved is quite high, thus the system wouldn't work in a production environment.) Adam