Over three years ago, I published the following paper on my weblog. The idea was to move the discussion on new gTLDs forward by finding common ground on a core set of principles. In 2002, the ideas struck some observers as radical. Now, in retrospect, they seem almost quaint. Over time, I think most of these principles have become generally accepted. Since we're again talking about new gTLDs, I thought it was time to republish the paper.