That may be a problem in the short term, but I hope that we soon have a number of competing gateways with plenty of liquidity between them. So even if you trade to MtGox IOUs because that's the deepest order book with the best rates between bitcoins and USD, you could immediately exchange those USD IOUs for ones at a gateway whose redemption policies you find more attractive.

The cost of transactions is managed by consensus much the same way as which transactions get into the ledger is managed by consensus. The system has a 'base_fee' that all other fees and reserves are based on. A node can introduce a transaction to change the base fee. If a trust-weighted majority of nodes vote "yes" on the transaction, it will be incorporated into the ledger and the fee will be changed.

Even without this, if there's a consensus that the transaction fee is too high, nodes could just propagate and vote "yes" on transactions with lower fees.

There are two things that help to prevent this:

First, Charlie would lose his network. So long as the value of his network is significantly greater than the value of the money in it at any particular time, this is a very significant deterrent.

Second, while I wouldn't expect Ripple debt to be enforceable alone, I don't see why the contracts backing it wouldn't be enforceable at entry and exit points. If Charlie got $10 cash from Jack and has a contract with Jeff that says he'll pay that money to him if the ripple system tells him to, I don't see why that contract wouldn't be enforceable. For Charlie to argue otherwise, he has to argue that he gets to default on his contract with Charlie and his contract with Jeff and keep the money.

But I absolutely agree that this is a risk that should be evaluated carefully. Using the ripple system as a medium of exchange, this risk is very low. Using the system as a store of value, this risk is higher.

To fill the donut hole, we need reliable gateways. If we don't have that, you're absolutely right, it's not usable for medium-sized transactions denominated in fiat currencies. (At least not without real community credit which, if it ever happens, is certainly far off.) We have a lot of things built into the system to encourage the creation of reliable gateways, we'll see if it works.

As for the benefits of Ripple versus something like PayPal, see the list I posted previously. Key advantages are lower fees, cross-currency transactions, the inability of a single entity to freeze your funds, higher speed, no chargebacks, the ability to easily change who handles your money without changing payment arrangements, and so on.

The big advantage of Ripple here, I hope, is competition. It would be like if you could choose PayPal and I could choose Dwolla and we could still pay each other seamlessly, with me putting money into Dwolla and you pulling money out of PayPal. We could choose the service that best covers our needs or our location. And if PayPal decides they have a problem with me because someone stole my identity a decade ago and they give me a lifetime ban (long story), I can move my business to another service without disruption.

There are several things built in to try to keep IOUs pegged to the currencies they're denominated in. For example, within the system, issuers must always accept their own IOUs at face value.

Also, we expect merchants or others who accept payments to just use a raw currency. You'll communicate a price as, say, "30 Euros". If a merchant can't maintain sufficient connectivity to allow consumers to pay them 30 Euros by using 30 Euros at a major gateway, they should consider that a failure. So they'll try to maintain connectivity to one or more major gateways such that anyone with similar connectivity can pay them 30 Euros for an amount very, very close to 30 Euros.

You may have a "junk economy" too, I suppose. I don't think people will find that terribly useful, but if they do, I guess that's good too.

One of the things that we're finishing up is a gateway between bitcoin and ripple. This will mean that both that you can cash out ripple IOUs as bitcoins and also that you can pay merchants that take ripple payments in bitcoins.

Absolutely. And I should try to be clearer about what I mean.

Someone can owe you money in the sense that the system records a particular balance between you and them. Here, the words "owe you money" don't really refer to an actual legal or moral obligation but to a particular status in the system. That status has an effect though, because the system has that status, the system will act in a certain way when it receives future transactions.

Where it gets real is when people have an actual "withdraw on demand" agreement. You could have such an agreement with a friend of yours. You could agree that the ripple system says they owe you dollars, they'll pay you dollars on demand in exchange for you cancelling out that debt. More likely, you could enter into such an agreement with a "gateway", an entity that makes it their business to enter into "withdraw on demand" agreements with lots of people.

We should make it very clear that absent an agreement that says otherwise, there is no inherent obligation to settle "debts" recorded in the ripple system. The system gives those "debts" power, and you can't stop it from doing that. But that's strictly within the system until you get to a case where you have an enforceable withdraw on demand agreement (or people are willing to settle on an ad hoc basis).

If I extend you 10 USD credit, you are under no obligation to ever pay me any actual US dollars. However, I can then acquire your IOUs within the system various ways and I can use them as a medium of exchange with anyone who can accept them. I can also give them to you and obtain other IOUs you hold within the system. This is all automated. But to get actual dollars in my hand or bank account, there needs to be an agreement outside the system.

If ripples become scarce in the future, the transaction fee can be lowered by consensus. Ripples are quite finely divisible, but if it came to that, their divisibility could be increased in the future, just like with bitcoins. That couldn't even come close to being a problem in less than 100 years though.

Currently, currencies are just a three-letter code and a sequence number (in case a currency is re-issued). We have plans for custom currencies but they're not supported yet. If someone created a new currency, you would never see that currency unless you elected to transact in it. You can't hold an IOU you didn't agree to hold, and the currency is part of what you have to agree to.

I'm not sure if you mean nodes you might connect to or nodes you might trust. The list of nodes you can connect to is managed just like Bitcoin does it. Nodes keep lists and offer those lists to their clients. We don't currency have a DNS anchor, but will probably add one. You can have a list of known node IPs in your configuration.

As for nodes you might trust, anyone who wants to can maintain a list of such nodes. Servers can choose root trust points and then those root trust points can list additional trust points.

For clients and servers not trying to actually process transactions, you only need to use trust to determine what the current consensus ledger is. From that point on, you're just walking hash chains.

The order book consists of entries in the ledger. There are entries that hold actual offers and index entries used to track the "tips" of the order books. Since the order book is in the ledger, the same scheme that ensures everyone operates against the same ledger does the same for the order book.

To create an offer, you have to do two things. First, you have to consume XRP for the transaction that places the offer. Second, you have to have sufficient XRP in your account to cover the increase in the reserve. The reserve drops when the offer is removed or fully taken.

I do too. And what's stunning is how simple our solution is. However, there is a lot of subtlety underneath the simplicity.

For example, one key thing we did was organize the ledger as a hash tree so the entire ledger can be stated with a single hash. This makes it easy for all participating validating nodes to sign a statement that they believed that the entire database had a particular state at a particular time.

Another thing we did was organize the ledger so that simple proof chains can be extracted from it using only hashing operations. So a client that sees signed receipts to convince them that the ledger had a particular hash can be shown a chain that proves a particular entry is or isn't in the ledger.

That's a fair point. I agree that we need to put much more effort into documenting. There just are only so many hours in the day.

There are a variety of advantages of the distributed database model:

1) You can easily exchange the balances for balances at different gateways. This means people who use different gateways can still pay each other instantly.

2) You can easily exchange balances for different currencies.

3) The market for exchange balances across gateways gives you a reliable real-time view of how trustworthy gateways are.

4) A gateway can't stop you from exchanging IOUs or enforce policies on anything but deposit and withdraw operations.

5) Costs are kept low because no entity can set the system's fees.

6) The gateway can't really freeze IOUs, destroy IOUs, force chargebacks, or the like. (Strictly speaking, they can. But for practical purposes, it's extremely unlikely that they will because they'll make their own customers innocent victims. It's the same reason Mt Gox had to accept tainted bitcoins from their merchants.)

7) People can easily operate small gateways in areas not served by larger gateways and offer 1-to-1 exchanges of their IOUs for a major gateway's IOUs.

8 ) Longer term, ripple may facilitate community credit.

9) You can change what gateway you use without having to notify everyone who pays you or have a change in how you integrate with the payment system.

I think the complexity argument is a red herring. Every payment system can be looked at from a high level at which it's simple and a low level at which it's complex. Ripple is no different. As for the electricity use, that strikes me as a very strange argument, but assuming it's true, I hope the benefits will significantly outweigh the costs.

If you only trust theymos, then he is the only person who can owe you money. You cannot hold any other kind of IOUs. Nobody can give you an IOU that you haven't chosen to accept. Internally, the ledger holds a "ripple balance node" for a pair of accounts and a currency. So if you trust theymos to owe you up to 50 USD, there will be a "vampire/theymos/USD" node in the ledger with a credit limit in the "theymos owes vampire" direction of $50. If that's the only trust you have extended, the only way I can pay you $10 is to make theymos owe you $10 or to you make you owe $10 less to someone you already owe money to.

People generally form a ripple payment transaction by knowing who they want to pay and how much. "I want to pay vampire 10 US dollars". To do that, I have to make someone you've chosen to trust owe you $10 more than they do now. I can do that by exchanging any IOUs that I hold or any credit that's available to me. But you decide which IOUs I have to get to you. I can, of course, always pay you with your own IOUs if I can find them. So if I can make you owe $10 less to anyone you currently owe money to, that's a $10 payment to you too.

They couldn't issue unique IOUs to you unless you had agreed to accept them. You can't give anyone an IOU they haven't agreed to accept (other than their own).

Right, but that's because PayPal can default on just you. Do you worry that they'll choose to shutdown their business while your money is there? Probably not, especially considering there's an entire economy that only exists because of them. But if you did worry, the solution would be to keep as little money with them as possible, not to miss out on the economic opportunities PayPal has created.

A ripple gateway can't stop you from trading IOUs that you hold to others who wish to accept them.

This is the nice thing about the way Ripple IOUs work -- unless the gateway stops paying everyone, their IOUs should hold value. And if the gateway stops, or slows, redeeming, it will become quickly obvious as the exchange rate on their IOUs changes.

Yes, that's right. For any particular currency, between any two accounts, there's only one balance. While you must acquire other people's IOUs before you can trade them, you can produce an unlimited supply of your own IOUs, limited only by whether other people are willing to hold them.

Sometimes we use IOU terminology and sometimes we use balance terminology. In the system, there are really just balances. So if we say "Jeff holds $50 in Jack IOUs", that's the same as saying "Jack owes Jeff $50". If Jack then pays $50 to Jeff leaving a balance of zero, this can also be described in IOU terms as Jack destroying the IOUs Jeff holds.

Pretty much. If you literally don't trust anyone to owe you anything, then you can't transact in fiat currencies using Ripple.

It depends just how risk averse you are. You could stick to just regulated financial entities insured by major governments. You could stick to only individuals personally known to you with whom you have "withdraw on demand" agreements. Any one of them wouldn't be able to do you much harm.

This really wouldn't work. You wouldn't extend significant trust to anyone unless their trust network was worth more than their total trust. So anyone who did this would do it at their own detriment and to your benefit. For example, say you've used Amazon for a few years. When you pay Amazon for an order, do you worry that they'll run off with your money and not ship your order? Even forgetting the possibility of a chargeback, you know that Amazon's business is worth much more than the total value of all the money they're holding at any one time. Plus, even the cost of them betraying you on one typical order will still leave you with a net plus for having trusting them because of the accrued benefit of hundreds of fulfilled orders.

The biggest threat would be a gateway going out of business due to a theft or an insider running off with all the customer's money. If you're very risk averse, don't use the system as a store of value for fiat currencies.

We've tried to design the system so that gateways have very strong incentives to run a solid, reliable business.

There's really no difference. Checks are used as payment and all they do is trade debt.

I think you misunderstand how people pay you. They pay you by making entities that *you* choose owe you money. It's just like the regular financial system -- when you pay me $50, you do it by making *my* bank owe me $50 more. The only entity I ever have to collect from is the one I've chosen to trust.

The idea is more or less to be a nearly-free, decentralized system like PayPal, except with no chargebacks, no entity that can force arbitrary policies on you, easy cross currency transactions, and so on.

I'm one of the people working on it. There's lots of explanatory material in the wiki: https://ripple.com/wiki

Ripple IOUs are just balances, just like a bank balance. If you have a bank balance of $50, that means your bank owes you $50. For me to pay you $50, I give you a check. The system processes that check and then my bank owes me $50 less, your bank owes you $50 more, and you consider me to have paid you $50. Most likely, you would only hold your bank's IOUs, so the only worthless IOUs you have to worry about is if your bank fails.

I would argue that a more appropriate model than the federal reserve is the post office, which starts out with all the stamps. Opencoin plans to begin giving away most of if it shortly. 1,000 ripples is enough to create an account, set up a trust relationship with several gateways and fund thousands of transactions. Unfortunately, making it free for everyone forever is not really possible because of malicious transactions causing denial of service attacks, ledger bloat, and so on. The use of a peer-to-peer model with no central authority should make it possible for it to get pretty close.

The first part is correct. There are 100 billion XRP in the genesis ledger and XRP cannot be created. It's not quite correct that 200 XRP must be held in an account at all times. Rather than having fees for creating accounts or maintaining ledger entries, Ripple uses a reserve. A reserve is XRP that you cannot transfer but that can be used to pay transaction fees. This also makes it harder to jam yourself into a situation where don't have enough XRPs to perform a transaction. The reserve is enough to cover 2,000 transactions at current rates.

They can be bought and sold, but their primary purpose is to pay transaction fees to protect the network from denial of service attacks and spam transactions.

In theory, a critical mass of users of any system can change the rules. However, there is no mechanism to create XRP other than a fundamental change in the system. While this could be done in principle, just as Bitcoin could be modified to keep the block reward at 25 forever, this would severely undermine confidence in the system and it's hard to imagine everyone needed to make such a change being willing to do so because of the damage it would do.

https://ripple.com/wiki/Ledger
https://ripple.com/wiki/Hash_Tree

Feel free to ask for details. I'm happy to improve the wiki. From the wiki:


This is spread throughout the wiki. You're right that's not concentrated in any one place.
https://ripple.com/wiki/Continuous_Ledger_Close
https://ripple.com/wiki/Ledger_Cycle

Once they're distributed, those who hold them can do whatever they want with them, and nobody can create new ones. Large numbers of them will be given away soon.

There were several sets of technical discussions about this. Some of them are here: https://ripple.com/wiki/Unedited_Notes

I'm happy to answer your questions and improve the wiki.

There are a lot of reasons you might prefer Ripple to the existing banking network, even if you assume that Ripple "only" provides the same functionality as the existing banking network. You can easily transact in multiple currencies and get exchange rates without fees. The rules are the same globally, well-defined, and nobody can enforce an arbitrary policy on your money. You have precise control over who you trust. And so on. The idea is to provide rules and functionality that is very similar to Bitcoin but with amounts denominated in existing fiat currencies.

There are two types of schemes. One where each holds part of a single key and one where each holds a key and multiple keys are needed. The accountability problem is a problem with schemes of the former type.