> Nice thougt, but on second thought impracticable for various reasons: > as > you pointed out, it sacrifices security, which IMO is not up for
On face value it does - but I wonder if we might find ways to compensate for that. More hops maybe. Before TOR existed I gave thoughts to a file sharing system which otherwise was similar to TOR but worked on a friend basis. My idea was for politically oppressed, no one had a complete file, but only pieces and was anonymous anyways. > discussion. Also, it would become much harder for new tor nodes to be > added to the network (I, for example, do not know any other tor > operators personally) and therefore would also sacrifice potential There are issues - but I think given the status of TOR and where its headed now it might be worth some research. People could offer to be introducers, or allow certain "new" people on at a risk.. > capacities and performance. Plus, we don't really gain any security, Actually I think we might gain performance in many cases. > since an "attacker" would only have to gain the trust of one > established > operator with one legitimate node to add the imaginary thousands of > illegitimate others to his trust list and therefore to the web of > trust. Adding someone would not add all your friends - but instead your friend relationship would allow your friend to relay through you - and then on to your friends. Your friends would never know who your other friends are. If someone is bad, you simply boot them and revoke their trust. Since you are their relay, your friends are now protected because you have booted them. The idea is that each person would have relay access to all of their friends, would relay to their friends etc.. After 3-5 hops.. the numbers could be quite big.
