Reputation Systems

Development
1

Has there been any discussion of whether or not this will have some sort of reputation system? I know that due to it’s basic nature of being anonymous, people could hold more than one account. But I think its more or less proven that reputation systems, while not perfect, substantially increase reliability. Has this been considered of this?

What makes this so unique is that it’s a new form of the internet where you have login credentials. So you don’t need a centralized company to monitor reputation. You could have a global decentralized pseudonymous reputation system that links across the network.

6 Likes
2

Yes, the ability to rank PMID’s (public names) and associated public data is something we have discussed many times. Similar to a reddit type ability network wide. There are issues over minority protection, but all in all we imagine its worth implementing this. It will allow you to set a minimum rank of people you wish to connect wit and sire you wish to see.

For fully anonymous data this is likely a must.

4 Likes
3

I’ve not thought this through, but am a fan of reputation systems and think this sounds like a great app, which provides a reputation API to other apps. Something like Disqus but distributed & anonymous.

2 Likes
4

The promise of the system is network validated arbitrary data storage. The first application of this is safecoin.
As David mentioned, this could be expanded to a network-wide reputation. It could also be used for app specific reputation (Reddit defines their own reputation, monitored by the network), and app specific identity verification (for voting, access rights, age verification, Sybil defense, etc). All of this with the same security as the currency, without all the back-flips like colored coins needed in Bitcoin. Since this is all as secure as the currency, the data can be used in smart contracts as well.

The beauty of this is that the reputation and identity system is built into the core infrastructure. Creating aggregate ranks based on multiple reputations and complex access rights systems become not just feasible, but easy.
And there is no central control. You always have control over who can see your identity, and which identity you use at any given time.

7 Likes
5

I wonder if the is research on trust networks?

SAFE could provide self validating trust for entities (people, apps, repositories, items of content/data) and the qualities/properties of those entities (popularity, accuracy, reliability, relevance, identity, trustworthiness! Or suitability for x where x is some user selector for which a trust measure has been added).

If we can code a generic trust schema that stakeholders can configure themselves it would magnify the value of SAFE a million times.

3 Likes
6

A reputation system should have:

  • A small payment 1EURO to get a reputation badge, this would reduce a Sybil attack and go towards further development of the rep system
  • An Up or Down vote (Reddit) an 0,01EURO per vote could also prevent missuse
  • How many successful transaction you’ve got
  • How many successful delivery of physical goods you’ve got
  • If you own a physical store & chamber of commerce number
  • Social fingerprint (Facebook, Linked In, Twitter etc)
  • Comment field & the average feedback time after an customer complain

These are just a few markers, but they should get a small impression

7

Is this something to be baked in or as an app? I’m quite fond of keeping the core basic and allowing for as many 3rd party opt-in situations as possible. Choice between reputation systems seem just as important as having a reputation system in the first place.

4 Likes
8

I vote app, but it could be a valuable take-up feature, so since there will be apps and apps, maybe one important enough for MaidSafe to be involved in or create. I imagine @dirvine has lots of ideas competing for MaidSafe time once they get going in this area though, and can’t wait to see what they might be :-).

3 Likes
9

An example of a distributed reputation system is getMarked

Part of Project Bitmark

getMarked is a distributed reputation system, like karma or kudos which follows people around the web.

Marks as Reputation

getMarked is something simple which your friends and family will use.
Reputation is given by the process of ‘Marking’, giving someone a Mark for something they have created or shared.

For international readers: ‘Marking’ is like ‘Scoring’ in English, and a ‘Mark’ is a ‘Score’.

Usage is similar to clicking Facebook’s ‘Like’ button, no complicated crypto currency program to be seen.

Reputation systems such as scores, karma, and likes have proven successful, Marks can follow a user around the web, and also real life.

The process of marking associates the feeling of earned value with the term Mark.

As the marking system is used, Marks become more distributed and scarce.

This adds a layer of competition and incentivization for users, they will compete for marks as they become more scarce. Natural milestones exist, to get 1000 Marks, 100 Marks, and later just 1.

Marks as Currency

The novel part is that Marks are literally currency as ‘1 Mark’ is ‘0.001 Bitmark’. This comes later, when a user realises that they can spend their earned marks on something.

Eventually marking something could be synonymous with paying for something. Marks are to be earned, each mark earned by an individual adds value to all.

Consider marks to be spendable karma, an amusing post on a social network could pay for your coffee, marking a video of the crisis in syria could pay for aid on the ground, marking an article about a mistreated animal could pay for it’s shelter, and marking this idea could pay for it’s creation.

getMarked also acts as a faucet, distributing free Marks to the first users. 1,000,000 Marks will be distributed by it, 500 to each new user.

This distribution model creates an initially exclusive invite only network, powered by 2000 people.

Additional marks to send can be acquired by people by depositing them from the Bitmark block chain.

Each users reputation is the sum of all Marks they have been given.

Their available balance to give as marks to others, to withdraw, or to spend, is the balance of marks they have been given minus any marks they have sent.

Marking Web Addresses

When a user marks something with a web address, they are not only scoring it but also bookmarking it.

Each user then has a store of all the things they have marked on getMarked.

When you mark something you can choose public or private, public is the default.

If it is public then it is published on getMarked, and you become it’s marker.

Each thing listed on getMarked can be marked by other users.

A measure of marks received over a period of time gives each marked thing a rank, the most marked things for a given time period are the most popular and the most viral.

Think: reddit.

Distributing Marks and Pulling People In

The marks given on getMarked are given to the marker.

Anybody who further shares the viral thing elsewhere will receive a portion of the Marks it recieves via their share. This gives each user more marks to further mark with, distributing marks between far more people, and making marks even more scarce.

If the thing marked has a getMarked button on it, then it is claimed by the content creator, the owner of the button. When this is the case any marks given on getMarked are split between the marker and the creator.

This encourages content creators to join getMarked to ensure anything new they create receives it’s marks.

It also encourages companies to join getMarked and use our button

Consider a music store where every song can be marked by our users. Each user that marks a song provides a small amount of revenue, if a song is good it can go viral and receive a large amount of marks.

Now consider the self publishing music artist who shares a song for free, his friends mark it, their friends see it, potentially all of getMarked’s user base sees it, and if it is good it receives many marks. Which pays for the song and let’s everybody hear it for free.

This model can be applied to anybody who makes any thing, they simply share and if it is good they get marked.

Bootstrapping Business

Each person or service who receives marks en mass, will reasonably need to withdraw and deposit, to save them or use them elsewhere. The process of doing this is through the Bitmark system, wherever getMarked is present you will find Bitmark powering it and keeping it secure.

Every service which adopts getMarked will see that it is a revenue stream, as getMarked adoption grows so does their revenue stream.

getMarked adoption make Marks more scarce, each person may reasonably have less, making each individual mark more valuable, such that divisions of marks can become the norm.

Anybody can create a service where items are markable and receive revenue from it, the Marks are divisible so small fees can be taken by service providers and websites to cover their costs. This may reduce the need for adverts on websites as a source of revenue.

Content promoters receive marks given to their articles, whilst the thing being promoted may also be marked providing revenue to it’s creator, and anybody else who may subsequently share either, and will receive marks too.

Each person can therefore earn and spend marks when relaxing, when working, and by doing anything they enjoy, at home and at work.

Each creator of things can receive marks as a form of revenue, and use them to purchase other things.

As services and providers acquire more marks, they may well wish to use them to pay for services they require, and encourage their associated businesses to accept marks as payment.

Consider a small organization which rescues children from streets and gives them shelter, now consider a video about them being marked by our users, the good cause alone will generate many marks, and pay for the organization to continue their work.

getMarked provides instant one click viral crowd funding for everybody in the world at web scale.

It is reasonable to think that at this level of adoption marks will be scarce, and that a single mark may have a fair value.

This multifaceted approach is proposed to bolster and bootstrap Bitmark adoption within the general population, whilst lowering the barrier to entry, and ascribing a notion of value to each Mark.

3 Likes
10

A reputation system for companies, organizations and services would be great. I’m doubtful about a reputation system for individuals however. People can much more easily “spend” their trust points or go bonkers than companies. I will use the distrust everyone approach. :smiley: And then rely on trustless automatic systems which I believe will be the predominant model in the future.

1 Like
11

[quote=“Anders, post:10, topic:152, full:true”]
A reputation system for companies, organizations and services would be great. I’m doubtful about a reputation system for individuals however. People can much more easily “spend” their trust points or go bonkers than companies. I will use the distrust everyone approach. And then rely on trustless automatic systems which I believe will be the predominant model in the future.
[/quote] Agree! This kind of labeling for people is just another enclosure control scam.

12

When those “points” are meaningless, yes.
If they’re hard money (or expensive cryptocurrency or token), no.

Automatic systems will (in their essence) be the same point systems that you see today, just without the worthless points. Instead they may be “reverse valuable points” (aka surety bonds or “good behavior insurance”) that those who get cheated and can prove it will be able to claim and approaches which share the common characteristic that they’ll cost individuals who screw up.

1 Like
13

I’m not convinced about mixing currency and reputation:

  • it allows reputation to be traded
  • it breaks with the natural model - trust is not “spent” but accrues based on behavior that is mutually beneficial (to a pair, group, the whole) and vice versa. Like the ranking mechanism of SAFE nodes

We can do a lot with currency/tokens and these ideas are good to play with, but I think trust and reputation requires a different model.

Reputation Relies on Identity

A key to trust is identity - reputation built up with one identity remains with that identity. The, identity can be anonymous, but so long as its trust/reputation can’t be acquired except through known trustworthy behaviors, and can be damaged/lost through known untrustworthy behaviors, then it becomes safe to risk a relationship/transactions based on trust - trust that this identity will continue to behave in accordance with past trustworthy behaviors.

Reputation Is Mirrored Not Transferred

In the real world, recommendation provides a method for reputation transfer of a sort. A kind of proxy or mirror: I trust X, so if you trust my judgement, you are more willing to trust X. If my recommendation turns sour, my “mirror” is tarnished, for you at least, even if you trust it was well intentioned.

Decentralised Reputation

I notice here that trust is not just about X, but about how you and I regard X. So reputation is a distributed quality, not a property attached to an identity. This decentralisation of reputation makes it much harder to acquire fraudulently (you have to fool many, not one), and also harder to lose all at once (more forgiving, chances to learn, modify behavior, optimise for the whole).

SAFE Reputation

I’d like SAFE to enable ways for humans to extend our natural ability to handle reputation and trust, which we do automatically in personal relationships, to the very large networks we encounter today through travel and communications.

4 Likes
14

Okay, let’s see:

  1. How much would you trade your reputation on this forum for (in BTC or US$, please)?
  2. If you wouldn’t trade your reputation, but you suspect that “others” would, do you think your standards are higher compared to your fellow netizens?
1 Like
15

For me the problem is not whether I would sell my reputation, but the question of what reputation is worth.

To me one of the advantages of reputation is that it makes everybody equal. If you keep your promises whether in terms of hosting data or in giving accurate information, then your reputation will reflect that. That is you don’t have to have lots of resources to have a good reputation. If you allow reputation to be bought and sold then this playing field is tilted in favor of the rich.

Which to me defeats the advantages of a reputation system. The SAFE network already has a wealth asset in the form of safe coin. If this is transferable in a currency type sense, why bother?

1 Like
16

Some would, some wouldn’t, some would if the price was high, some would if they were in dire need.

You make a very simplistic argument here, as if my answers to very personal questions have a bearing on the arguments I made to support the statement you quoted.

Personalising your argument makes it less meaningful IMO. Whatever you, or I, would answer to those questions has very little bearing on the issues.

Your questions seem as much aimed at pushing me towards a position of either conceding or making myself out to be superior in some way. I resent that personally, and find it distasteful as a way of debating.

1 Like
17

This is another problem with monetizing reputation, it means that we are disincentivizing honest behavior. That is a person with a good reputation through hard work will not need to buy reputation, but a person with money can go out and do things which earn them a bad reputation and just paper it over with money.

And the problem is not that you or I or anyone in particular would engage in such dishonest and unproductive behavior, but that we all know that SOMEONE will. What this means is that in any system where there is a possibility of doing something like this, that system just becomes a race to the bottom, and the “reputation” system just becomes an extension of the wealth tracking/transfer system, which as I said above the SAFE network already has.

1 Like
18

You will find out when you attempt to sell it or allow others to bet that you’ll fail to honor your promisess.
The market will let you know.

Right, so if there’s already a proxy method to trade them, why should we be concerned if reputations were to become tradable?

Okay, let’s try again:

Trustworthy behaviors on the network are basically those that make you pay your bills and deliver on your commitments (basically, services and goods you provide to others) on time.
If you stand to lose more than you stand to gain from untrustworthy behavior, you are more likely to behave properly. It’s quite simple.

I don’t trust “recommendations” because they carry zero monetary commitment. Like when a politician makes a “commitment” he’ll do something. What if he doesn’t? Nothing.
Some movie star may be “recommending” one of two brands of equally good types of bottled water because she got $100K to tweet about it.
It is extremely unlikely her reputation could be tarnished by some spectacular product safety issue related to that brand of bottled water. But still there’s zero reason for me to trust her “recommendation” regardless of her reputation that she “earned” (for free, basically) in “the community” (a bunch of pseudo-anonymous citizens/voters (yikes!) who may not even be real people, but bots).

That’s simply false. Since there’s no “World government” to ensure that all MaidSafe or Bitcoin users are actual people and that one person doesn’t operate more than one account or address, identities on the network are basically meaningless just like (or more so!) they are on Facebook (at least Facebook or Google may ask you to verify your mobile phone number).
It will depend on the value of the network (if demand for MaidSafe accounts or reputation is low, it’ll be cheaper than it otherwise would be), but it is likely to be cheap to open thousands of accounts and make each trade with all the others and give them best reputation rankings possible. Why would anyone want to rely on such a system for reputation and trustworthiness information?

Also, decentralized reputation makes instant reputation loss possible and that’s a good thing. In the EU today one can ask a search engine to remove a link to a factual article that (in one’s opinion) ruins his or hers reputation. The article will still remain online, but it’ll be harder to find.

With distributed reputation systems one’s reputation worldwide will be diminished or ruined within couple of confirmations from the network. And that is good because everyone will have the same access to up-to-date reputation information from the public ledger they use (Bitcoin or other). So if somebody steals one’s pin/password and tries to fool all of his business contacts, he’ll probably fail to fool more than one.

I could ask about “one” instead of “you”, but that wouldn’t change anything.

Your … erh, one’s credit score (or “rating”) is an example of how one’s reputation for being known to deliver services and goods on time (pay bills on time) is tradable on other markets: it is sometimes used in hiring as a proxy indicator of someone’s reputation and trustworthiness.
It is possible to create bets or various financial instruments based on one’s trustworthiness and so quantify the score in financial terms or units of crypto- or other currency. Once one knows the value of these contracts or can derive it from more liquid examples similar to his, he knows the value of his reputation.

Another issue is that among two people with the same level of trustworthiness the more trustworthy of the two is the one that has more liquidity. And a 60% trustworthy person with $100K in his pocket can be trusted up to $60K while a 99% trustworthy person with $50K in his pocket can’t.
And in this example I’d rather trust anyone (whatever his reputation score is) who agrees to put $60K in escrow with a mutually agreed upon notary until the deal is complete, which is “good behavior insurance” I mentioned in an earlier comment (above).

19

Again, the possibility of selling your reputation, simply means that this “reputation” system as a whole adds no value to the network. Its just another way of saying that you have the money to hide your dishonesty. This is one of those areas where direct market action serves no purpose.

I think @happybeing’s idea of reputation mirroring might add value, but simply allowing it to be traded, means that its just a system for tracking money.

1 Like
20

What do you mean by “proxy method?” If you mean that reputation which has all the characteristics of money, including free liquidity, is simply a proxy for money, then that is precisely the point I was making.

My point is not that we should be concerned, but that if something doesn’t add value and doesn’t function on the principles on which it is ostensibly based (like a so-called reputation system which is just certain type of monetary unit, basically interchangeable with all other monetary units), why bother?