Clone this wiki locally
Initially, in the test system, it is on the honor system.
The 1 person 1 ID is a little like Ethereums mining mechanism in that we want to really look at the best one we can get in the timeframe, not particularly being hung up on a particular one. It is different in that changing the OnePerID doesnt require a 'hard fork'. However, such a decision also needs to be decentralized; likely by delegative vote, it has to take enough votes such that an attacker has no chance to change the OnePerID in his favor.
It needs to have few enough 'multiply counted people' so it doesnt significantly affect any votes.
Note that for some of these, if non-people from attackers get an ID, they attacker may get more power to add more non-people. I.e. it stays working because it is working, and it has to keep attackers away from tipping points.
This forum thread. Don't know if anything there works, probably not, but there is a space to explore.
Basically, people are indicated by other people to be persons, and once they have enough reputation, can indicate more people.
Note that if a lot of it is public, these could have a Facebook-like effect of mapping relationships of people. It is preferable to avoid that as much as possible.
Hiding secrets "geocaching" could be a form of low level vouching.(and a 'game' that computers cannot play) Another approach is Face check, combining a reputation system to see if two people in videochats are the same, there the conclusion might be longer-range, however.
Note: (by Jasper) i dont see it entering the blockchain easily. Possibly humans already in the system could look at identifying physical aspects of humans.
Using the fact that people cant be in two places at a time, when gathering people check that the others are who they say they are, and indicate that in that system.
Synchronized pulsed captchas
Computer-hard problems like captchas are used to help against sometimes. They are not fool proof. However, if you synchronizedly do it at the same time, it is harder to hire people or computers to emulate it. If we get CPU mining, and someone attacks it but mines outside the pulse, it may show up in a temporary decrease of mining power. This blog post details a way one might implement captchas.
It might be good to have some way to simulate which methods might work. (TODO read about some criterions somewhere)
Trustlet.org has suggested data sets.
Centralized (ugh) services
Estonia has one, apparently, however it is far from the best solution. Although it might bring in information that isnt taken at face value.(just like passports/ID cards might)