To Home page

Ripple

This rambling explanation fails to explain what level of anonymity is provided, or how it is provided. We want aggregative anonymity, cryptographically secured, with money flowing on chains of trust.

Fails to explain need for integration with reputation management, fraud problem, ordering, person to person transactions

How internet money should work

The idea is that an organization will issue some kind of money equivalent - promises to deliver US dollars, promises to deliver gold, shares, promises to deliver some service. Let us call the organization issuing the money Big Money Issue Corp, and the money Big Money Issue Corp OUs. 

You will be able to buy Big Money Issue Corp OUs from Bob, and use them to pay Carol for something. You will be able to do this over the internet by instant messaging - dropping some money into the instant message. Let us suppose you pay Carol to deliver some cigarettes. Carol will know you paid her, and what you paid her for, and you will be able to prove that Carol agreed to deliver cigarettes for a certain payment and that you paid to deliver some cigarettes, prove it to the whole world, which ordinarily you will not want to do, but you might want to do if the deal goes bad.

Now the most obvious way to make this work is that the people operating the financial system know everything about everyone, and all transactions are on their computers, readable by the government, who can thereby tax everyone.

We don't want a technology where everyone deals directly with Big Money Issue Corp, because then Big Money Corp will be under pressure to refrain from dealing with this or that person or group, and then it will have to know its customers, and for a big organization to know its customers is likely to be inefficient for the big organization and burdensome for the customers.

We want payments to be provable, and linked to deals, so that if something goes wrong, the customer can do something about it - but provable and deal linked payments run contrary to anonymous payments.

Cypherpunks had a vision of how anonymous funds transfer would work on the internet, which vision ran into some disastrous obstacles.  The cypherpunk vision was built around Chaumian anonymity, but this technology does not in itself provide many capabilities that are necessary for internet money, and makes implementing those capabilities more difficult.  In the Chaumian vision, everyone dealt directly with Big Money Corp, but Money Corp could not link Bob's payment to Carol receiving money. Unfortunately, this meant no recourse for Bob if he paid and Carol stiffed him, plus the fact that everyone dealt directly with a single center provided a single point of control

Ripple is a brilliant insight, great idea, but suffers from the critical mass problem that afflicts any proposal to provide competition to the existing government run money system. 

In addition to the critical mass problem, governments tend to treat competition as illegal, once the competition gets large enough to to make a dent - observe what is happening to e-gold. E-gold is attempting to comply with overbearing and intrusive regulation to the great detriment of its business and its customers, but the government has no intention that it be possible for e-gold to comply. Of course a decentralized system is a lot more resistant to this than a centralized system such as e-gold. E-gold looks like being the Napster of internet money - one that showed the way, though it was ultimately suppressed. 

Therefore, in my opinion, the best prospect for a decentralized money system such as ripple is in some activity that is already somewhat illegal, for example a file sharing system, in which the participants grant each other upload credits. 

Ripple, as it stands at present, does not really address the routing problem, (in the sense that the proposed routing algorithm sucks), and indeed does not need to, for there are very few participants. For a file sharing system, would need to be more DHT like. 

To the extent that a payment system is anonymous, tends to be used by ordinary criminals and ponzis. This gives government justification and excuse to shut it down. 

Requirements of an internet payment system:

Payments should be identifiable by the participants, but not identifiable by outsiders, and should automatically generate and maintain the appropriate records on the participants machines. If you pay a net presence for X, you should be able to prove to outsiders that you paid that particular identifiable net presence for that particular X, but outsiders should not be able to prove it except with your assistance. This requirement presupposes a mechanism for identifying net presences - Zooko based yurls, and/or authorized certificates. Since we obviously want to transfer money between people who do not necessarily have authorized certificates, have to support yurls, which means we have to first have a yurl system, which at present we do not. 

To resist shutdown, and keep costs low, a payment system needs to inherently implement "ripple", making every man a bank. Ripple, however tends to reduce the profits of the currency operator, and is some work to implement, so is generally not done. 

The problem with Chaum's anonymous cash, and Brands' anonymous cash, as generally envisaged, is that it relies on a single central server, with which everyone has a relationship - a star shaped network. Internet networks really have to be power law networks - neither star, nor entirely equal peers. A single central server is a central point of failure, a point that governments can attack. 

"Ripple" means that if the Bank of Whatsit implements ripple, then you don't need an account with the Bank of Whatsit. You can do fine with an account with someone who has an account with the bank of Whatsit. Hey presto, a power law network. 

So if you want to buy a domain name registration from Wonton, a registrar that has an account with the Bank of Whatsit, and you have an account with Joe, who has an account with the bank of Whatsit, then Joe's machine tells Bank of Whatsit to pay Wonton. The Bank of Whatsit only knows that Joe payed Whatsit. Joe knows that you paid Wonton, but does not know what you paid Wonton for. Wonton knows you paid them through the Bank of Whatsit, does not know or care who Joe is, knows what you paid them for, but cannot prove what you paid them for. You know, and can prove, that you paid Wonton, and can prove what you paid them for. 

Now what I have described is all account based - no Brand style or Chaum style anonymous tokens. But if the Bank of Whatsit implements ripple, Joe, or anyone else, *can* implement Chaum style or Brand style anonymous tokens. Further, you can pay Wonton using Joe's tokens even if Wonton has no relationship with Joe, and has never heard of these new fangled tokens, and the Bank of Whatsit has never heard of them either, for through ripple you can turn the token payment into an ordinary account payment from the Bank of Whatsit. 

Thus if we had payment guys implementing ripple, they could avoid the heat that is likely to arise from implementing Chaum style tokens, indeed be entirely unaware of either those tokens or that heat, combining the anonymity that comes from Chaum style tokens with the anonymity that comes from the faceless crowd of a power law network. 

These documents are licensed under the Creative Commons Attribution-Share Alike 3.0 License