^ this thread.
YA KNOW, GUYS, if you just removed the default trust list altogether and started using the forum's trust system like the OTC Web of Trust (or even better, if you grew up and learned PGP like the rest of the adults), this would be a non-issue.
The *only* bit missing is the PGP tie-in. That's it. It boggles me that there's a superior system
that can be almost completely mapped to this bogus forum trust setup, and nobody knows/cares.
So get rid of the entire Default Trust list.
Fuck them. They need to learn.
Fuck you. Who cares.
Imagine.
To further drive home the point that every single goddamn one of you who is not present in the OTC Web of Trust is doing this entire trust thing perfectly wrong, I'll
.
I. The Web of Trust is not, as the name would seem to imply, an oilfield in which trust plays the role of oil, and you deploy some apparatuses and other devices to extract the trust therewith.
Trust is not in the web, that or any other web. Trust is not in the wording, not on the paper, not in the symbols, or certificates, or seals. Trust is not in others and other things, but much like faith - for which it serves as a ready synonym - trust is within oneself.
The Web of Trust is merely the infrastructure upon which trust is built, by you, for your own use, within yourself. The same objective set of relations can result in drastically different trust in the eyes of drastically different third parties. The point of the WoT is not to make these judgements for you.
II. The WoT works by reducing the unknowns problem.
It allows the user - any user - to confidently identify the sources of information, both in the negative and in the positive. That is to say, if sources of information exist, the user may by the WoT find them, and safely assume that should no sources of information be thus found, no sources of information in fact exist. It further allows the user to judge the quality, reliability and precision of said sources, and this independent both of the direct source and of the counterparty he's examining.
III. How to use it. Let's understand what all this means with a simple example. Consider the village of Wotania, wherein there exist exactly 100 agents, all participating in the Wotania WoT, and wherein strong currency is used for all transactions. Suppose Joe wants to buy a used car from Moe. While the currency he'd be paying Moe in is strong, the car he's buying is anything but, and so Joe would like to evaluate Moe before paying him. What's he to do ?
First off, he should evaluate Moe's relevancy. Obviously since there are 100 agents in total, the highest score any one agent could achieve in the WoT would be 990. This would reflect the situation where one particular citizen was considered as perfectly known by all other citizens. Thus if Moe has accumulated a score of 33, this clearly shows that at least 4 of the agents know him, which roughly means one in 25. If Joe knows 80 agents personally, but none of the agents he knows rated Moe, this makes Moe suspect on the first pass.
Sure, it's possible that Moe is only known to the minority subgroup of 20 agents doing things with cars, and within that subgroup he's quite well known, whereas Joe is strictly a part of the Wotania web industry, and they walk everywhere. However, this is something that Joe can evaluate by himself, without having any need for Moe, and without needing to ask him anything. For instance, if the island has a total of 3 car manufacturers, and all of them are in Joe's 80, Moe's position suddenly became untenable. Sure, it's possible that used car salesmen are completely separated from car manufacturers in the manner car people are separated from web people. But it seems less likely (and the likeliness of it is, again, fully within Joe's estimative hands).
Leaving that aside, if the average rating in Joe's WoT is 3.14, whereas Moe has received his 33 trust from exactly 4 people, averaging thus 8.25, there's suddenly exposed a very strange divergence between the two groups. Sure, it is legitimately possible for Moe's subgroup to be much more tightly knit, and thus his friends much more familiar with him than is the case in Joe's group. This happens, but not without other consequences, which again add valuable information into the credibility equation.
Or suppose instead that Moe's rating of 33 was provided by 10 people, yet still none of them are in Joe's reach. Somehow it is possible that out of 10 different people, 10% of the population of Wotania, nobody had any dealings with the 80 people Joe knows. They live on the same island, they go about their daily business, yet no one's ever met. Possible, especially if one lives in the US, but also improbable, and in this improbability, informative. Because this is the point of the WoT : its factual information reduces to a pile of factually correct statements, which all work as probabilities, and it's trivial to calculate the likeliness of a fact that depends on a number of other facts with known likelinesses : you just need multiply. 0.2 here, 0.5 there, 0.66 and another 0.15 suddenly you're at 0.8% which may well be under your risk tolerance threshold.
But let's say that out of Moe's 10 raters, 3 are in Joe's WoT. One supplied 3 points, the others one point each. Joe directs his questions as to Moe to each of them :
Dear Sue, Hue, Lue :
I am considering buying a used car from Moe.
I see that you have rated him in the past. How did that go ?
To which the three are held to answer (and the treatment for non-answering is again an informative variable, whose treatment rests with Joe) in the canonical form :
Hi Joe!
I bought a pair of shocks from him April last. They were broken, but he refunded my payment without much hassle.
All the best, Sue.
or
Hi Joe!
I bought a car stereo from him. It had a big scratch on the side and some other misc damage, but he let it go real cheap.
All the best, Hue.
At this point, Joe knows, but quite exactly, what the story with Moe is : he's a small time car thief, and he's trying to make ends meet by selling whatever car parts he can get his hands on. Simple, really.
Alternatively, of course, he could be a very reputable used cars salesman. 0.8%, or as the clueless say, “it’s impossible to call it one way or the other”.
The important parts here are the easiest to overlook :
i.
All this was established with no input required from Moe. Sure, if you want to get fancy you could easily ask a few choice questions just to see how the guy answers. But at its core, simplest, most barebone functionality, the WoT does not require you to talk to your subject of interest, at all.
ii.
The process works universally : If it worked you know it worked, and of what quality its results are. If it failed you know it failed, and why and how come. It has, in other words, exactly the opposite properties to those decried by Naggum in his Lisp advocacy misadventures piece.
iii. Especially considering the alternatives,
this is mindbogglingly cheap. In fact, I am ready to argue that the savings this model brings are, both in aggregate and on a percent basis, more significant than the savings Bitcoin brings in payments, or provably-fair in gambling.
So why isn't it used more often ? Because people are stupid.