Reply to Craig Newmark

I found a post about an interview with Craig Newmark and replied to it here

The following is in response to Trust and reputation systems: redistributing power and influence

1. Reputation need not be contextual. You only need to trust someone to adhere to one principle: karma neurtal (including honesty). everything else follows from that. For example, I have a friend who is most unreliable at keeping appointments. But he knows this, and even advertises the fact of his notorious unreliablity in this capacity. So we work around it. If we have agreed to meet on a particular day, he has to call me on that day a little before, to confirm that he will be able to keep the appointment.
2. Persistent and verifiable identity is not a prerequisit, nothing needs to be proved at all. look at ripple for example.
3. history is unnecessary - see ripple
4. Unvarnished does not track trust, it tracks opinions
5. no ethical standards are required whatsoever, laws would just get in the way - again, look at ripple.


Reply to Charles H. Green:
1. this one has been dealt with I believe
2. this does not happen in my network system
3. what if a takes responsibility for the indiscretion of b, and b takes responsibility for the indiscretions of c? If you are unable to think of anyone you trust enough to take personal responsibility for the consequences of their indescretions (if they were to default), then there is no conceivable efficient network system that would serve your needs.
4. The object of trust is correction of indiscretion. If a seller is completely useless, and you complain about it and they apoligise, give your money back, with compensation for wasting your time, then they have corrected their indescretion. If you have an (indirect) trust connection to amazon, then you trust who they trust, and if they trust an author, then you trust that author not to waste your time. Then authors will tend to work very hard to accurately describe their book and the target audience. Any attempt to obtain more readers than the book is intended for, for example by over-tagging, will result in complaints and therefore expense. The key component of the system I propose is what has been named by others as a system of "distributed lawsuits".