Topic: P2P Exchange for bitcoin - page 7. (Read 42805 times)

Sure, TCPIP, HTTP, RSS - whatever.

Its not about TCP/IP. Dont worry, it will get done right.

Well... that's 5 minutes of my life I won't get back. Seriously? wow. I would start by reading books about how TCP/IP works, and then write some actual software.

The idea of one person or a small group of people having control goes against the idea of decentralization. This must be a very tiring obstacle to have to deal with. All that needed to happen was to design a software along the basic lines of bitcoin. Its free, its open source:


https://bitcoinstarter.com/projects/54      

You are not being precise at all. If an attacker would want to manipulate MtGox he could do several things, but first of all brake into the machines and steal money. MtGox has a database of users, bank accounts and Bitcoin wallets.

Is this a piece of code? What machine runs what code? How does the exchange actually transfer $ on one account to BTC in the same account? Is there a limit order book? Where is the datastructure stored? How does the matching take place? How does a user open an account? where does he open up an account? etc

I see the overall point, and agree with the sentiment, but it simply does not make sense to distribute a matching engine, because for a market price to form one needs on place where computation takes place. Or at least I can't see how this could be possibly be different. And if it is possible I would like to know how this could work. Say you have 100 buy orders and 100 sell orders at various prices going into the system at various times. If every transaction is verified by a network, it would take 1h for one order. At that rate you process 24 orders a day, i.e. at an average order size of 10 BTC ~240 BTC per day. I guess what you could have is something more like a call auction which takes place in time intervals, for example once a day. But that would mean if you have price drops of say 30%, you can not sell, but have to wait. In futures markets this is called limit down and usually means you're going to have a very bad situation.

MtGox has no control over the bitcoin protocol but still they use bitcoin to store value.
Our proposal is something similar to the bitcoin protocol, no one can control it but no one can manipulate it either.

If I would operate an exchange the last thing I want is a solution I have no control over. That's the first principle of security. You have control over what you do, otherwise how would you know everything is safe? The operator has to have control over the hardware, the software and the bank accounts. Again, the structure of the Bitcoin network has nothing to do with the structure of an exchange. Although there are possibilities, but the central question is how the interface between fiat and BTC works, i.e. the banking side of the exchange. Why should I distribute traffic over a network if I can do all calculations on a local machine? That's like sending a parcel across the street, but around the world first.


The status quo tells you little what can happen to you. The operator of the second biggest exchange could now face prison time. Not quite the same as "pain in the ass".

So one solution would be to have gateways/brokers. Instead of going to the exchange yourself you got to a broker, who can minimize costs. But then you have counterparty risk, which even for exchanges is significant (B24, Bitfloor), but higher for possible brokers. And if the brokers and exchanges move to offshore locations, counterparty risk increases. So in the end, as of today, it does not make that much sense to transact in Bitcoin in a country where you have existing banking infrastructure. The best thing would be to focus on places where the value is so much greater because the infrastructure is not there.

What makes sense to work on is to have exchanges develop which is helpful for the network overall. And the most pressing issue are questions about the legal situation.

Yes its up to the exchange operators/brokers/dealers to deal with the legal issues.
What I'm talking about is a technical solution that exchange operators would be able to use and a framework of how the exchange should work, but in the end its up to them how they handle fiat, settlement and regulation. The proposed exchange system is a hosting service which provides a service to run a virtual exchange instead of coding/hosting it your self, but the fiat is always in the bank account of the brokers/exchange operators. Its up to them to follow regulations in their state. Operating an exchange may be pain in the ass in some states, and completely liberal in others.

Your throwing around words. These trading venues are run by operators for a good reason. Sure you can have 5-10 orderbooks, which are then integrated by trading between the venues. But you still need a) a matching engine b) an operator c) gateways. So that is what we have, and that has nothing whatsoever todo with P2P networks / blockchain etc. MtGox might collapse precisely because they have too high a latency. That is the problem. You need central exchanges, because you need people in charge of bank accounts. Even if you have had a solution, this would be money laundering in probably all countries, because this would make fiat money untraceable. I've spoken with someone who has the capacity to open an exchange and it wouldn't even be much effort. But he doesn't want to face the legal challenges. I would say you need a chunk of money to reserve for clever lawyers. Otherwise you will face what B24 operator faces right now.

The regular Joe or Jane that is not online much does not know what a torrent is.











Anyone can create a new exchange:

https://bitcoinstarter.com/projects/54      

radiumsoup - the issue of trust have been discussed a lot, and my solution to this problem is virtual exchanges (alliance of brokers).
What Im talking about is not something like ripple, but rather isolated exchanges. The customer has to examine and judge which exchange he trusts most before he opens an account, or maybe open account at several exchanges to diversify the risk.

Each exchange must verify and accept the brokers that will be part of that exchange, by sharing certificates with the brokers.

Each virtual exchange issues exchange-IOUs (for BTC, USD, EUR, XAU, XAG etc...) that the brokers buy/sell to their clients. The exchange should require a deposit (in form of BTC or fiat) in advance from each broker as a insurance in the event of a broker defaulting. Any possible loss will hit the exchange, if the exchange has enough deposits the other brokers and traders will never lose money because of a defaulting broker.

The only thing all virtual exchanges will have in common is a blockchain containing IOUs which makes it possible to also send IOUs directly to other people instead of using exchanges.

You might be right that this system don't change anything in real life, but at least it enables more competition by providing the same framework to everyone.

The system isn't the complex part - the complex part is the maintenance of the list of trusted entities, which happens on a per transaction basis, and is purely an abstract relational system that has value that differs for each and every person using the system. Every person is responsible for figuring out the level of trust they place in each and every token generated. (Think MtGox codes, BTC-E codes, etc.)

Inherent in the actual transaction of fiat for BTC is trust that the person or entity you're using for the exchange is going to do what you agree to do (perform the transaction at the agreed rate.) That's easy enough in person because you're meeting someone face to face for the transaction. It's almost as easy with a service as big as Gox because they have a history of performing as promised and they have a well understood business model. The proposed system introduces a trust token that can be traded easily, but anyone can generate their own token - meaning the list of entities CLAIMING trustworthiness is going to get very very big very very fast. The real life problem created by the system, then, is that the complexity to the average person trying to figure out who is trustworthy and who is not just got a lot more difficult. Their path of least resistance is to continue dealing with entities they already know and have a reasonable expectation of trusting, so this system will not benefit them at all; they will continue as they already do, preferring face to face transactions or transactions with large well-known commercial entities.

So if there is no practical benefit to end users, the only other benefit I can imagine would necessarily be between exchanges. The same system of trust must still be established between exchanges as the example above. No exchange will automatically trust any other exchange's trust tokens without the establishment of a trust relationship. There is, in practical terms, no difference between the trust an end user places in an emerging exchange and the trust an established exchange places in that same emerging exchange. The new exchange must still prove their trustworthiness, and each trusting entity must still make their own determination of whether or not to trust the new guy, independent of the decisions of others to trust them. Again, the path of least resistance is to continue using already established relationships with large well-known commercial entities. There are already mechanisms that provide this service, like the aforementioned MtGox codes and BTC-E codes; they are not often used because it's easier to just deal with the exchange directly in most cases, or in Bitcoin in others. Sure, they have a small place in the market, and their use is already established without the need for a separate system of trust tokens. I would not be surprised at all if the exchanges didn't already have some sort of agreements between them that stipulates transfers between them get settled at certain thresholds or at regular intervals, and they trust each other to pay as promised when called. But they don't have a need to transfer those agreements (if they exist) out to third parties - they should already have BTC/USD/whatever to be able to do those other transactions as promised.

This system is, in essence, a lot like introducing to Bitcoin the idea of floating checks made out to CASH. Someone creating a trust token is essentially saying to you "this is redeemable for Bitcoin, I promise." If you trust them, that's a perfectly valid transaction if you want to delay the receipt of BTC, but when you try to give the "check" to someone else who may not know who wrote the "check", and ask them to trust that YOU trust the person who gave you the "check", and then they spend it asking the next person to trust them to trust you to trust the person who wrote the "check", and so on until it's ultimately redeemed. All that comes back to each individual person maintaining a list of entities each person trusts. The "check", in the end, is only as good as the entity's word that it is ultimately redeemable, despite how much trust the Nth person has in the Nth-1's trustworthiness, and so we're back to large, well-known commercial entities being the best source of these trust tokens. A Cashier's check from Bank of America is going to be much more easily passable than a post-dated personal check written in a shaky hand with grease stains and a torn corner. Cash, (actual Bitcoin), is still much better, because there's no actual need for trust that it's redeemable for Bitcoin... it *is* Bitcoin.

Like I said, there may be a small niche need for this, such as the relationship between large, well-known commercial exchanges; but it's not really any benefit to the end user or to anyone wishing to become an exchange, as the trust relationships must be established just as it would need to be without this system.

That is not true, the FX markets are traded decentralized at different OTC-exchanges, but due to high liquidity you will rarely see differences between FX-exchanges, any difference will be used by algos to make arbitrage profits.

Sorry, but have ever traded with a LOB? As a trader I will got to the liquidity and low latency and low fees. One trade at MtGox takes 500-2000ms to get through. Actually with a proper run exchange this should be <50ms. A decentral LOB just doesn't make sense. What makes sense is to have brokers who in one market place. In the world there is one price for gold, copper, oil, EURUSD and USDBTC. No need for more than one price. Just think what happens if you two prices at two locations. Why would anyone sell for 100$ if he can sell for 130$? So what you have is a bunch of LOs at various levels which meat at one point, the mid-price.

What can and should be done is to create an exchange alliance, which interacts with community.

The goal is not to make anything complex, but rather something simple and if possible with existing technologies.
Bitcoinx for example uses the bitcoin protocol, so everything dont need to be invented.

Do you think the iPhone or their app-store was a revolution? No, they just packaged different technologies in a way that attracted the mass market.
We could do the same for exchanges, creating a framework that many small exchanges can use so they can attract customers, without the need to code everything.
The goal is an exchange of exchanges, anyone should be able to create their exchange regardless if they got a dealershop in a village or a huge exchange with millions of users.
P2P-technology makes it scalable.

My simplified proposal:

Storage & Transaction of contracts
-New alt-coin called IOU-coin with the purpose of being colored for transactions with bitcoinx.
-Huge inflation to prevent hoarding this coin, since its main purpose is to act as a digital contract. This coin will not compete with bitcoin, but rather be a complement to bitcoin.
-Blockchain stores transaction of IOUs (legally / morally binding contracts) and prevent double spending. transactions are verified in minutes instead of 1 hours in bitcoin.
-Client-software can send colored coins to other clients.
-Settlement of IOUs for real bitcoins, real dollars or real gold is up to each exchange to handle.

Exchange software
-Each exchange is created with a bittorrent file, anyone can create a new exchange.
-Each exchange contains one or many brokers
-Each broker contains one or many traders
-The server software is run by 1 or many hosts that process orders
-Each exchange runs orders in memory and torrent file for fast trading speed.
-The net value of executed orders are sent to the IOU-coin blockchain regulary.

Beliefs are powerful things. Those who believe that bitcoin markets should match bitcoin are growing by the day. I certainly am one of them and so are the people who post here.





How about giving away free open-source software? ( Just like bitcoin. )

honestly, I don't see the market need for this... everyone agrees that exchanges are needed, and the market clearly isn't satisfied with MtGox being the single biggest player by such a wide margin. But the easiest answer to the problem of one massive exchange isn't a complex P2P implementation of real-world transactions, it's simply giving the big exchange better competition. And we have that now in the form of competing exchanges with arguably better uptime and more features, albeit with varying degrees of success... but there are a few very strong competitors already.

I mean, more power to you guys, really - but I don't think this project is going to revolutionize anything. At best, you'll make a semi-anonymous exchange with a requirement for decentralized trust of distributed individuals for the actual, real-world exchange of fiat. And management of that trust of individuals is a VERY difficult thing to do. So difficult, in fact, that I don't see how "the masses" will flock to this system.

The more I read on this, the more I think it's a solution looking for a problem. If it works, great - I'm sure it'll fill a niche need, and there's nothing wrong with that; godspeed, really. But I'd keep the expectations on adoption of this idea reasonable if I were you.

I am willing to incorporate anything that will further the decentralization of bitcoin markets. A free bit of software like this:

https://bitcoinstarter.com/projects/54

seemed perfect to me, but like you said, its a group decision. I am willing to include other ideas into my idea, but a programmer will have to work that out. As it stands, this  idea (https://bitcoinstarter.com/projects/54) is ready to implement. It just needs making.

I started writing about decentralized markets back in February:

https://bitcointalk.org/index.php?topic=145389.0;all

Gollum started this one off in April, this month.

Right! Work together or fall apart. What is the decision?

This subject has been discussed a lot in several threads including this and several good ideas have repeated themselves.
Im not the right person to judge which proposal has the best prospects to succeess.
Some of the proposals may be combined, and some may not.

I will summarize different proposals based on the threads mentined in the opening posts so we can choose our path:
-Vote on the proposals
-If there is not a clear vinner: Have final discussions and let consensus decide

We should be flexible to find the best possible solution instead of letting our egos decide.
United we Stand - Divided We Fall.

Lots of talking, no actual doing?