Talk:Inferred Trust Ideas

Funny! Morgan and I had a chat about this when walking through Montreal yesterday night. :)

I think there should be more to it. The problem simple multiplication of trust and trust of judgment is that the values will quickly end up close to 0%. It depends on what the values mean, 0% = distrust, 100% = full trust. It probably makes more sense to make the numbers go closer to 50% (neutral trust), or to use a two value system. E.g. a triangle, where the bottom means "I don't know anything", the top left "distrust" and the top right "full trust". guaka wikitalk 15:06, 22 September 2007 (PDT)
 * That's an interesting point. Does 0% mean distrust, or no trust. That's a question I'll give some thought to. Can you positively distrust somebody? Can somebody have a negative trust value? Interesting... --CallumTalk 14:54 (GMT+1), 24 Sep 2007
 * I totally agree with Kasper here, actually I think we both agree with Josang trust algebra! ;-) Check the demo applets linked from the page for understanding better what we mean with triangles of trust. About meaning of 0, I argue that a trust statement of value 0 in [0,1] means distrust, while if you don't know or don't wnat to express trust in user A you simply don't express a trust statements, i.e. there is no arrow from you to A in the trust network. But this is a very open question. In particular expressing negative relationships open a whole new issue of issues (?!?). See A survey of trust use and modeling in current real systems if you have a lot of time and are prepared to an hard-to-understand English glibberishing (no idea what glibberishing means of course ;-) --PaoloMassa 09:52, 24 September 2007 (PDT)