Quote from: esuncloud on September 18, 2014, 08:38:39 PM

Quote from: jl777 on September 18, 2014, 08:30:43 PM

Quote from: esuncloud on September 18, 2014, 08:14:53 PM

I have been watching this atomic cross chain trading thing for a while, and it is interesting to see that you guys issue asset for this relatively old problem.
The reason that this ATOMIC was not implemented before is because it is difficult. However, considering you guys have gotten million dollars in valuation, maybe you could hire really smart guy with this money.

It will be helpful if you could give more details about this POE/POG, and convince us that it will really work as you expected. Meanwhile, ATOMIC assetholders will not get anything before it is finished and the price of ATOMIC will be totally speculating before that.

Take this a risk reminder of the challenges in ATOMIC development, and I would love to see ATOMIC could finally come true.

It is a difficult problem indeed and I have warned of the risks.

The POE and POG are the way to ensure that there are the most connections possible between all the various blockchains. if these do not exist, then making a cross chain tx is problematic. So think of this as the "bandwidth" for doing atomic tx. The bandwidth by itself is not enough.

Inside the MGW (multigateway) I am doing a primitive form of atomic exchanges, BTC for mgwBTC asset and vice versa. I have also done overlapped and synchronized micropayment channels using offchain exchanges of signed and validated transactions. not exactly atomic, but the performance profile is close.

So at the application level it is possible to achieve close to atomic exchanges, but there are always some issues regarding timing or some other things.

By having a "ticker tape" blockchain where all the atomic trades are recorded, this provides a single efficient place to verify the status of a tx pair involved in an atomic swap. Now if everybody were honest, then we would be done. Of course we need to prevent double spends, transaction malleability, etc.

So I am envisioning that at first an atomic tx will be having some chance of being invalidated, but as time passes then the confidence level grows, just as with any decentralized blockchain. There is a tradeoff between instant results and accurate results and the details as to what tradeoffs are chosen are not completed yet.

It is possible that the tradeoff decision is totally up to the client of Atomic. Since InstantDEX is a built in client of Atomic we will be working at the same issue top down and bottom up. It is likely I will release the InstantDEX without full Atomic support at first as it wont be ready yet, but there are several hard-coded ways to do the InstantDEX tx. The Atomic will just make my life a lot easier.

Stay tuned as we work out all these details and feel free to contribute with ideas, suggestions, criticisms, etc. We want an open process and invite all that want to help, to help!

James

Just added some details, and this is another reminder.

The closest solution for atomic cross-chain trading is as below:
https://github.com/TierNolan/bips/blob/bip4x/bip-atom.mediawiki

But this is obviously not enough, and another possible idea is to combine the idea of DPOS, and users dynamically elect agents to handle the exchanges.

MGW is more a proof of concept in this regard as the gateway function is similar but different from exchanging

with the miners available and adding some reputation system would allow for some semi-automatic method

clearly there is a tremendous need for a solution and now there is capital to fund the development, a ready mass market from SuperNET and so this is the explanation for the interest in Atomic