Mercury - Fully trustless cryptocurrency exchange - Looking for testers!

[mappum]

Could a service for automatic portfolio rebalancing be developed out of this?


Just posted this in the ethereum forum: http://forum.ethereum.org/discussion/1844/automatic-portfolio-rebalancing/p1

Would it not be nice to have a etherium service that could automatically rebalance a portfolio with a number of cryptocurrencies? With rebalancing i mean that when there is a change of the value of one of the currencies, it is bought or sold so as to keep the ratio between the currencies in the portfolio constant.

Do you know of such a service?

Best / Tim

Sure, that sort of thing would work great with Mercury. But not quite yet, I'll have to develop the REST API before applications can start being built on it.

[mappum]

In case anyone is watching this project, I still plan on releasing the alpha version of Mercury today. I've been working all week to get it ready, and it's almost there. However, it should only be used for testing purposes, don't trust it with all your money until the system has been audited and we're sure it's safe.

I'll post a link to the download tonight.

[p2pbucks]

Great project ， Come on  : )

[mappum]

The current version of Mercury (0.0.1) is now ready for testing, you can download it here: https://github.com/mappum/mercury/releases/tag/0.0.1-alpha

Remember, this is alpha software and is likely to have bugs. Feel free to test it out with a small amount of money, but it's not ready to handle large sums of money.

If you encounter bugs or have any comments, please let me know and I'll address them as fast as I can. Thanks to everyone for being interested in Mercury!

[cass]

Great Project! Do you've considered to add BitShares BTS bitUSD, bitGOLD etc!?
I'm working for BitShares. If you've interest in principle please don't hesitate to contact me via PM!

Cheers
cass

edit: BTW i really like your github activity map^^ Nice idea

[delulo]

...the OP...

Thanks for this wallet! It looks very promising.
My questions are regarding the usability and how it works in practice:
1. AT would allow (in principle) Alice and Bob to exchange their coins across two different chains. But how do do the two find each other? Does your wallet include a order book and order matching?
2. What parts of the exchange process would be manual, what automated?
3. How would the user experience be different from a centralized exchange? (centralized exchange: cash in coin x -> place order on market for coin x and coin y and wait till order is matched OR take an order that has been placed -> cash out)

[mappum]

Great Project! Do you've considered to add BitShares BTS bitUSD, bitGOLD etc!?

I'll consider any cryptoassets that have lasting value. I'm going to take a neutral approach and add assets that have sufficient volume, so I don't have to pick sides on what currencies to use. Bitshares has significant volume right now so I'll look into adding them.

Thanks for this wallet! It looks very promising.
My questions are regarding the usability and how it works in practice:
1. AT would allow (in principle) Alice and Bob to exchange their coins across two different chains. But how do do the two find each other? Does your wallet include a order book and order matching?
2. What parts of the exchange process would be manual, what automated?
3. How would the user experience be different from a centralized exchange? (centralized exchange: cash in coin x -> place order on market for coin x and coin y and wait till order is matched OR take an order that has been placed -> cash out)

1. Mercury uses an order matching server. However, no trust is placed in that server (it does not deal with the actual exchange of funds), and use of the service in optional. Two parties can find each other and make a deal on their own (e.g. through forums or IRC), then they could still use Mercury to trustlessly swap the funds.
2. Everything is automated, the user experience should be just like any of the current exchanges (except that it is from a desktop wallet instead of a website).
3. It shouldn't be different, except instead of depositing into an account, you can trade using the balance in your wallet.

[delulo]

Great Project! Do you've considered to add BitShares BTS bitUSD, bitGOLD etc!?

I'll consider any cryptoassets that have lasting value. I'm going to take a neutral approach and add assets that have sufficient volume, so I don't have to pick sides on what currencies to use. Bitshares has significant volume right now so I'll look into adding them.

Thanks for this wallet! It looks very promising.
My questions are regarding the usability and how it works in practice:
1. AT would allow (in principle) Alice and Bob to exchange their coins across two different chains. But how do do the two find each other? Does your wallet include a order book and order matching?
2. What parts of the exchange process would be manual, what automated?
3. How would the user experience be different from a centralized exchange? (centralized exchange: cash in coin x -> place order on market for coin x and coin y and wait till order is matched OR take an order that has been placed -> cash out)

1. Mercury uses an order matching server. However, no trust is placed in that server (it does not deal with the actual exchange of funds), and use of the service in optional. Two parties can find each other and make a deal on their own (e.g. through forums or IRC), then they could still use Mercury to trustlessly swap the funds.
2. Everything is automated, the user experience should be just like any of the current exchanges (except that it is from a desktop wallet instead of a website).
3. It shouldn't be different, except instead of depositing into an account, you can trade using the balance in your wallet.

Regarding 2/3: ....Then the wallet would allow to automatically sign/broadcast atomic tx whenever my order is matched on a server?

[btc_enigma]

Nice ! I don't fully understand the internals how it work . But this sounds very promising . Keep up the work !

[mappum]

Regarding 2/3: ....Then the wallet would allow to automatically sign/broadcast atomic tx whenever my order is matched on a server?

Yes. The server will notify the wallet that it has found matches, and the wallet will verify they are valid then start the swaps with the other traders.

Sweeeeet!

Does that mean "instant transactions"  like at the speed of Darkcoin (10 seconds)?

Or how long would the actual transaction take to verify?

No, you will still be limited by blockchain confirm time (but a green address model can be used to create instant transactions for trades). For small trades, it can be safe to use unconfirmed transactions, so the trades will be instant. For large trades, you will want to wait for confirms to make sure the other trader can't double-spend.

[delulo]

With the current setup of Bitcoin, you would need to wait just above 100 confirmations to do secure ACCT (Atomic Cross Chain Trading). This would be 1 full round of all delegates for DPOS (about 15-20 minutes atm). These intervals don't make sense for trading. For ACCT to become viable for any chain (without waiting too much) you need a PBFT or PBFT-like real-time Consensus and Ordering Protocol (COP) on top of the main blockchain to make ACCT viable.

That, or the light client trading server does Paypal-like fraud detection and compensates counterparty risk accordingly with pretty high trading fees (anywhere from 0.5-4% of each transaction).

Source: https://bitsharestalk.org/index.php?topic=14472.msg188878#msg188878

[mappum]

With the current setup of Bitcoin, you would need to wait just above 100 confirmations to do secure ACCT (Atomic Cross Chain Trading). This would be 1 full round of all delegates for DPOS (about 15-20 minutes atm). These intervals don't make sense for trading. For ACCT to become viable for any chain (without waiting too much) you need a PBFT or PBFT-like real-time Consensus and Ordering Protocol (COP) on top of the main blockchain to make ACCT viable.

That, or the light client trading server does Paypal-like fraud detection and compensates counterparty risk accordingly with pretty high trading fees (anywhere from 0.5-4% of each transaction).

Source: https://bitsharestalk.org/index.php?topic=14472.msg188878#msg188878

I'm not sure where that information is coming from, it is incorrect. There would be no reason for atomic swap transactions to require more confirmations than any other payment transaction, so something like 6 would be a safe upper bound for even very high-value trades (and less confirmations can be used for smaller value transactions).

And any contract protocol that relies on third-parties as escrow or delegates is really a hack and not a very clean way to work with cryptocurrency. Protocols like that are also weak as any third-party can be bribed or be a sybil node.

BlockStream's design for sidechains relies on atomic cross-chain swaps with a very similar protocol to what is being used in Mercury, and their judgement is probably the most trustworthy in the cryptocurrency space. (See http://www.blockstream.com/sidechains.pdf, page 21).

[jwinterm]

With the current setup of Bitcoin, you would need to wait just above 100 confirmations to do secure ACCT (Atomic Cross Chain Trading). This would be 1 full round of all delegates for DPOS (about 15-20 minutes atm). These intervals don't make sense for trading. For ACCT to become viable for any chain (without waiting too much) you need a PBFT or PBFT-like real-time Consensus and Ordering Protocol (COP) on top of the main blockchain to make ACCT viable.

That, or the light client trading server does Paypal-like fraud detection and compensates counterparty risk accordingly with pretty high trading fees (anywhere from 0.5-4% of each transaction).

Source: https://bitsharestalk.org/index.php?topic=14472.msg188878#msg188878

I'm not sure where that information is coming from, it is incorrect. There would be no reason for atomic swap transactions to require more confirmations than any other payment transaction, so something like 6 would be a safe upper bound for even very high-value trades (and less confirmations can be used for smaller value transactions).

And any contract protocol that relies on third-parties as escrow or delegates is really a hack and not a very clean way to work with cryptocurrency. Protocols like that are also weak as any third-party can be bribed or be a sybil node.

BlockStream's design for sidechains relies on atomic cross-chain swaps with a very similar protocol to what is being used in Mercury, and their judgement is probably the most trustworthy in the cryptocurrency space. (See http://www.blockstream.com/sidechains.pdf, page 21).

I think he meant to say Bitshares, not Bitcoin, thus all the talk about DPOS and the bitsharestalk.org...

[delulo]

With the current setup of Bitcoin, you would need to wait just above 100 confirmations to do secure ACCT (Atomic Cross Chain Trading). This would be 1 full round of all delegates for DPOS (about 15-20 minutes atm). These intervals don't make sense for trading. For ACCT to become viable for any chain (without waiting too much) you need a PBFT or PBFT-like real-time Consensus and Ordering Protocol (COP) on top of the main blockchain to make ACCT viable.

That, or the light client trading server does Paypal-like fraud detection and compensates counterparty risk accordingly with pretty high trading fees (anywhere from 0.5-4% of each transaction).

Source: https://bitsharestalk.org/index.php?topic=14472.msg188878#msg188878

I'm not sure where that information is coming from, it is incorrect. There would be no reason for atomic swap transactions to require more confirmations than any other payment transaction, so something like 6 would be a safe upper bound for even very high-value trades (and less confirmations can be used for smaller value transactions).

And any contract protocol that relies on third-parties as escrow or delegates is really a hack and not a very clean way to work with cryptocurrency. Protocols like that are also weak as any third-party can be bribed or be a sybil node.

BlockStream's design for sidechains relies on atomic cross-chain swaps with a very similar protocol to what is being used in Mercury, and their judgement is probably the most trustworthy in the cryptocurrency space. (See http://www.blockstream.com/sidechains.pdf, page 21).

I think he meant to say Bitshares, not Bitcoin, thus all the talk about DPOS and the bitsharestalk.org...

Yes he was talking about the option of AT for Bitshares. Sorry if this caused confusion. I thought thought that he meant to generalize his statement which is why I posted it here ... Maybe he can clarify it himself.

[mappum]

Now it should be much easier to run Mercury, there are releases for each platform:

• Windows (.exe)
• OSX (.dmg)
• Linux (.jar)

Try it out and let us know what you think!

[sanfranciscobitcoin]

This is really cool, just a heads up - the QR codes that are generated don't seem to be valid...

I'm really looking forward to Mercury, thanks for all of your hard work so far!  

+1 for NuBits

[redsn0w]

Interesting new "concept" maybe in the next hours I will try it and write here my opinion. From the screenshot it appears "good" now I should test a few "exchange" between doge <> litecoin <> bitcoin.

Good work guys!

[dwma]

Interesting project, we need more of this sort of blockchain application and less currency based apps.

[jkoin]

Please add Primecoin, is'nt a "2.0 coin" but is important for the searching of largest primenumber.

The network primecoin has already several prime numbers discovered:

http://en.wikipedia.org/wiki/Cunningham_chain

http://en.wikipedia.org/wiki/Bi-twin_chain

[mappum]

This is really cool, just a heads up - the QR codes that are generated don't seem to be valid...

I'm really looking forward to Mercury, thanks for all of your hard work so far!  

+1 for NuBits

Thanks for letting me know, I've fixed the QR codes. (All I needed to do was give them a white border, apparently they don't work on black backgrounds).