Topic: How Can We Create a Decentralized Exchange? (Read 966 times)

"Our blockchain" is already bloated with stuff like shatoshi dice.  And these transactions would include mining fees, so that would be helpful.  And like I said, just using colored coins doesn't guarantee that the shares will actually be paid for. Since I could send a colored coin and then not get payment - the other person would have the colored coin and I'd be boned.

Other then that, my proposed system could be implemented just as easily with litecoin as bitcoin.   I think mainly it would be used to move shares between exchanges as well as person to person OTC trades.   

With my system, though you could still have an "exchange" where people just post bids and asks along with their addresses, and then remove them when they're filled.  The exchange wouldn't actually need to hold any BTC or shares, it would just be like a message board.  Or you could write an application that used an IRC channel to post trades, as bitcoin originally did.

So a huge question here is: Is transaction rate an issue? What is the extent of influence of the time-rate of transactions on an exchange? Are faster or slower time-rates better or is there a middle ground?

Obviously, bots can trade much more quickly than a human can manually. Would slower exchange rates be more advantageous to manual trades? Would this dampen the severity of panics?

Lots of questions here.


I am not familiar enough with colored coins at the moment to comment on this aspect.

Has anyone considered maybe using one of the secure alternative coins (the only one I can think of is Litecoin, every other one is venerable)

Instead of bloating our blockchain, use Litecoin as a coloured coin system.

They could be, they could serve as tokens for shares.   i.e. 0.00001btc could represent one share. The problem is you wouldn't have a two way transaction - if you pay first, you have to trust that the person will return the coin.  If you send the shares first you have to trust the other person will pay.  It might be possible to configure it correctly with the scripts, but then you'd need to be able to write scripts that know about coin 'color'.  Which I'm not sure is yet supported.  In fact, I'm not even sure what scripting features are currently supported.

My idea would only need normal scripts with data included, that the asset issuer would look for on the blockchain.

I was thinking about this the other day, it could be done on the blockchain pretty simply.

Suppose you have user A and user B.  A owns the stock, and B wants to buy it.

A transaction could work like this:

1) B sends A a small transaction, say for 0.001Btc, (and including mining fees) with a randomly generated token, as well as a request for an amount of stock and a price in the script section of the transaction.

2) If A decides to sell or transfer - they send another message back to B basically saying that they agree.

3) B sends A the bitcoins to buy the stock.

4) The asset issuer sees the transaction on the blockchain, and changes the bitcoin address associated with their transaction and updates their internal database of who owns what shares.

Optionally, you could have the issuer send a notification to user B confirming the order.

The nice thing about this system is that it would work entirely with the existing bitcoin protocol.   The only thing needed would be a tool to configure the correct script to include in the transaction - and these wouldn't need to be 'live' scripts, just scripts containing information that the issuer would look at.

Of course, this would be slow . The primary use for a technology like this would be to transfer shares between exchanges.  It would be easy for people to setup exchange sites that would allow quick transfers of shares for fast trading, or people who want to buy and hold could keep their shares in their own wallets.

Obviously I know there are other systems out there, like OpenTransactions and so on. But this would have the benefit of working with bitcoin so people who already understand bitcoin will understand exactly how this works and will know it's reliable.

Plus, the people who make money off of this will be bitcoin miners.  Which would be nice for the miners

What about these coloured coin ideas? Can they be used as shares for decentralized exchanges?

OK... but I was mainly speaking to the speed of the transactions (i.e., buys/sells) and their confirmations.

OK then, forget .onion. Lets just assume .bit. Even though it is difficult to access, obviously, when people hear about the exchange they are going to find an explanation on how to access it. I don't see that as a huge problem - any more than downloading a wallet and reading some instruction on h9ow to transact BTC. How can the transaction speed problem be dealt with? or is this even a huge deal?

I think it is already a mistake that the various Crypto exchanges aren't utilizing .bit.

Or easily accessed by the 99% of the world that doesn't run the chrome/firefox plugin. It's just obscuring a website, not hiding it or preventing seizure.

First problem is trust.

Is there any exchange idea where the only person you have to trust is the security owner/manager of the security you are buying? If an exchange can be trustless for the system itself and the only failure point is the securities (which will never be trustless by nature) I think that would be a good idea.

But alas, I can't think of a way that could happen.

With the announcement of the closure of BTC-T over potential legal fears I began to wonder why it would not be possible to create a decentralized exchange, possibly even placing it on the .bit or .onion DNS. Anyway, is this a possibility and what would be the difficulties and limitations associated with this?

One downside that I thought of could be the speed of buys/sells due to the potential for "double-spending".