[ANNOUNCE] Zero Reserve - A distributed Bitcoin Exchange

[anu]

I believe that the losses you take with such fraud will be way below the inflation tax and the interest payment to the govt-financial complex.

What makes you say that?

There are various studies about consumer interest payments coming from various angles. The thing is that virtually all money comes from debt on which interest is due. Even if you yourself are not indebted, you still need to pay interest which is hidden in the price of products, services and taxes.

There was a study for my home town (Munich/Germany) about that. With a fairly typical life style you need to own about EUR 2 Million to break even - taking a profit from the debt system. Everyone else is a net interest payer.

I don't think I understand the system fully though. Could you clearly explain, what happens if Alice buys 1 BTC of Bob, but in the sending of the fiat down the chain, one node is dishonest and doesn't forward the fiat, but pockets it instead, what effect does that have on the chain, and who loses money and how is dishonest behaviour punished?

The transaction will not proceed and will time out if he tries to keep his ledger in sync with his friends but does not stick to the protocol.

If he chooses to stick with the protocol, but simply does not update his ledger, then your ledger will not agree with his. The next time the nodes try to reconcile, a QDialog will pop up, telling you that something is wrong. Your friend will find from his TX log that his ledger is correct and yours is wrong.

As you see, you can't cheat that way without getting caught red handed either immediately or shortly thereafter

As said, ZR is relying on Bitcoin to prevent the seller from cheating: A 2/3 multisig. First the seller must place the Coins out if his reach, then payment is made, then the seller releases the coins from escrow.

Its similar to the torn bank note. The escrow is necessary because there is a possible attack vector if its only 2/2: You can initiate  a buy on the entire order book and then simply walk away. Doing so must be costly.

[aminorex]

Is it really even necessary to eliminate the risk of a failure in the trust chain?  I mean, if I owe money to Alice, and I'm planning on paying it back with money which Bob promised to pay me, but Bob fails to pay in time, I will still pay Alice, because Alice is a friend. I may not trust Bob in future, but I certainly don't want Alice to cease trusting me.  Then I'd have no hope of escaping the friend-zone.  A major reason hawala networks work is that the participants are willing and able to cover for each other to some extent.  Bob's mom needed a hernia operation, so we all deferred his obligations, lest we be deemed heartless bastards, and since who knows when your own mom is gonna get a descended...whatever.  That distributed compensatory mechanism is what makes the network anti-fragile.

[anu]

Is it really even necessary to eliminate the risk of a failure in the trust chain?  I mean, if I owe money to Alice, and I'm planning on paying it back with money which Bob promised to pay me, but Bob fails to pay in time, I will still pay Alice, because Alice is a friend. I may not trust Bob in future, but I certainly don't want Alice to cease trusting me.  Then I'd have no hope of escaping the friend-zone.  A major reason hawala networks work is that the participants are willing and able to cover for each other to some extent.  Bob's mom needed a hernia operation, so we all deferred his obligations, lest we be deemed heartless bastards, and since who knows when your own mom is gonna get a descended...whatever.  That distributed compensatory mechanism is what makes the network anti-fragile.

That is how it is supposed to work. There is no need to settle all debts all the time, just like in the banking system. In the banking system, they went so far as to make settlement almost impossible, after they started to de-monetize gold. What does "settlement" really mean? You swap a debt obligation of a friend against a debt obligation of your bank. Calling this settlement is IMHO a joke.

A breach of trust doesn't happen during a transaction. The way it happens is this: Someone uses all available credit with all friends to the limit and then refuses to recognize that debt. That default would remain local - it doesn't affect the network. Its no different to paying your friend's bill in a pub because he forgot his wallet and then the friend later refuses to pay back.

[bach]

I guess the amount of credit one would be willing to extend to someone depends on the amount of other credit already being extended to them.

Does that sound reasonable? Does Zero Reserve take this into account somehow?

[anu]

I guess the amount of credit one would be willing to extend to someone depends on the amount of other credit already being extended to them.

ZR does not know the credit of friend-of-friends. But yeah, that will be the rule: I guess Bill Gates and his buddies won't even bother to open the Credit Dialog for a mere $1,000 while college students would consider it irresponsible to grant that much credit to their friends.

Does that sound reasonable? Does Zero Reserve take this into account somehow?

Its definitely reasonable, but ZR does not make any efforts to encourage / enforce responsible credit granting. As said - you know how much credit your friends grant you, but you don't know how much they grant each other. It is the philosophy of ZR to keep data strictly local. So ZR's ability to do so is very limited, even if I thought it a good idea to try - which I don't.

[bach]

I guess it is about the arrows pointing away from Bill Gates (i.e. to those extending him credit / receiving IOUs from him).

Presumably the people to whom those arrow point don't need to be friends with each other and so know nothing about each other.

It would seem injudicious to make a decision on how much to lend someone without knowing about their other liabilities (a bit like what credit reference agencies facilitate banks in doing now).

[Voodah]

Interesting. Watching.

[phelix]

How are things going?

[anu]

How are things going?

Thanks for asking. Here is the status:

For now, libBitcoin has been repaced by the Satoshi client. I am aware that this is not the best solution, but for now, I need a working testnet mode and I could not make this work on libBitcoin. Before I even think about releasing something that deals with real money I need to make sure that it works with reasonably large networks with bogus money.

The wallet can't do much at this point - it simply shows the total balance and the balance of the addresses. But a "proof of concept" release is getting very close now.

[GianniDalerta]

Anu: did you see this on reddit:

BitMarket
over bitmessage.
http://bittext.ch/8wApvESl6y

[anu]

Anu: did you see this on reddit:

BitMarket
over bitmessage.
http://bittext.ch/8wApvESl6y

Very interesting idea. If I understand it correctly, this is a de-centralized Bitmit. I am not understanding how they deal with fraud, however - solving that problem is cruicial. Also how to deal with Vandals who just call it fun to disrupt a system? If they can solve this, they are a very serious competitor to ZR.

I am not entirely happy with my solution to the Vandal problem: I put a price tag to DOS attacks, but for someone who is willing to pay the price, they can do more damage than it costs them.

But then, maybe I am taking it too serious. Seems like nobody attempted to DOS the Bitcoin network so far, even though it seems quite possible.
(Some people called Satoshi Dice a DOS attack....   )



EDIT: I called them a potentially serious competitor. That does not mean I don't welcome this. My purpose is to help making the Bitcoin ecosystem immune to Govt/Bank action. I applaude everyone who does the same.

[GianniDalerta]

Ideas to share accross the board.

[Voodah]

Anu: did you see this on reddit:

BitMarket
over bitmessage.
http://bittext.ch/8wApvESl6y

That could be very interesting.

I would very much like for bitmessage to become more popular and services on top of it will certainly help.

[anu]

How are things going?

Thanks for asking. Here is the status:

For now, libBitcoin has been repaced by the Satoshi client. I am aware that this is not the best solution, but for now, I need a working testnet mode and I could not make this work on libBitcoin. Before I even think about releasing something that deals with real money I need to make sure that it works with reasonably large networks with bogus money.

The wallet can't do much at this point - it simply shows the total balance and the balance of the addresses. But a "proof of concept" release is getting very close now.

There is now a bounty for making libbitcoin work on Windows:
https://bitcointalk.org/index.php?topic=358298.0

I added a bounty to make testnet work:
https://bitcointalk.org/index.php?topic=358298.msg4560062#msg4560062


It seems that it was a good idea to make the Bitcoin implementation a compile time option and separate it with an abstraction layer.

[Mike Hearn]

You could also investigate doing what the Hive folks have done, and use bitcoinj via JNI. There are tools that can autogenerate JNI wrappers into C++ to make the bindings easier. bitcoinj supports not only testnet but also regtest mode in Bitcoin master, which lets you run a private testnet where you can mine blocks instantly. I've switched to regtest mode for most of my testing, it eliminates waiting for blocks or trying to find faucets completely.

[anu]

First successful "Trade". The Biteasy Blockexplorer shows the associated Multisig Address on the TestNet:
https://www.biteasy.com/testnet/addresses/2N4w1rAVcQ4gWSm3h34w6YJeC6AkZEDEkYc

[smoothie]

Curious, how do transactions sent in at the same time are handled?

What method is used to verify that person A's transaction came before person B's transaction? Assuming that they were both sent at the same time.

This is likely the one of the bigger issues when verifying an exchange transaction. Collisions will happen and dealing with them is tricky.

Let me know. Thanks

[anu]

Curious, how do transactions sent in at the same time are handled?

What method is used to verify that person A's transaction came before person B's transaction? Assuming that they were both sent at the same time.

This is likely the one of the bigger issues when verifying an exchange transaction. Collisions will happen and dealing with them is tricky.

Let me know. Thanks

First come, first serve. Messages are bound to come in serially, even if it is just milliseconds, because they come through the network stack. An incoming message causes reservation of an order, or a part of it. So there are 3 cases:
1. Both requests can be filled, NP
2. The first request gets filled, the second one only partly
3. The second request can't be filled - the seller sends a VOTE_NO message which causes the TX to TransactionManager::abort().

ZeroReserve very much looks like an exchange, GUI-wise. But under the hood it is OTC. Reason is that every ZeroReserve instance maintains an order book and does the matching. Orderbooks of different nodes are bound to differ.

[91porn]

Interesting. Watching.

great
Will try the latest version.

[phelix]

OK, the latest version does not crash any more on win32, seems stable so far. I had to remove the -rdynamic option from zeroreserve.pro, though.

I'd like to buy a tnBTC for 500 zimbabwe dollars!