Follow @nxtchg

Author Topic: A Strategy for Success  (Read 7892 times)

draco

  • Minion
  • *
  • Posts: 25
  • Respect: +3
    • View Profile
Re: A Strategy for Success
« Reply #30 on: March 01, 2016, 11:01:45 pm »
0
Quote
Not sure what you are describing here, but it doesn't seem to solve the problem of last tx.

If it fails, it expires. All previous expired too. So you're out of luck getting your funds.

It could work if the expiration could be much slower than 10 sec, for example 24 hours, long enough to be able to broadcast transactions with a high probability. But only the transaction with the highest version number is included in the blockchain. So the version number replaces the time-based expiration in my earlier idea. (The version number is also the original Bitcoin idea, but it had problems, in Bitcoin at least).

NxtChg

  • Overlord
  • *****
  • Posts: 1114
  • Respect: +61
    • View Profile
Re: A Strategy for Success
« Reply #31 on: March 01, 2016, 11:22:25 pm »
0
It could work if the expiration could be much slower than 10 sec, for example 24 hours, long enough to be able to broadcast transactions with a high probability.

This means nodes now need a mempool, same as in Bitcoin, then they need to decide how to pick txs to vote on, etc. It's just too big of a departure from the architecture optimized for fast confirmation time. No, that's not gonna work and is not worth it.

It's way easier to implement AC rules that allow claims in the form { amount, destination, signature } - these claims are not time-bound.

It's also way more flexible since you can now create channels to several people/services from the same account.

Should probably implement it, but later :)
Tentacle Overlord, The Deranged Genius of The Abyss

draco

  • Minion
  • *
  • Posts: 25
  • Respect: +3
    • View Profile
Re: A Strategy for Success
« Reply #32 on: March 02, 2016, 01:11:28 pm »
0
It's way easier to implement AC rules that allow claims in the form { amount, destination, signature } - these claims are not time-bound.

Of course, if there is an easier way to implement the same functionality, that is even better. But for the trading use case it is important that the amount paid can also go lower, not just higher. I'm not sure if this kind of "claim" can support that.

NxtChg

  • Overlord
  • *****
  • Posts: 1114
  • Respect: +61
    • View Profile
Re: A Strategy for Success
« Reply #33 on: March 02, 2016, 01:25:16 pm »
0
But for the trading use case it is important that the amount paid can also go lower, not just higher. I'm not sure if this kind of "claim" can support that.

That's not possible with any kind of system. Once you gave somebody a key, you can't take it back, unless you write something into the blockchain and that ruins the whole idea.

This can only be achieved with bi-directional channels, i.e. two channels in opposite directions: when you buy, you send a microtx to the broker. When you sell, he sends you one.
Tentacle Overlord, The Deranged Genius of The Abyss

draco

  • Minion
  • *
  • Posts: 25
  • Respect: +3
    • View Profile
Re: A Strategy for Success
« Reply #34 on: March 02, 2016, 01:57:21 pm »
0
That's not possible with any kind of system. Once you gave somebody a key, you can't take it back, unless you write something into the blockchain and that ruins the whole idea.

I don't think it is impossible, but it is a hard problem.  Whatever the solution is, I don't think it would have to work in all theoretical cases if it works with a high probability in normal cases.

This can only be achieved with bi-directional channels, i.e. two channels in opposite directions: when you buy, you send a microtx to the broker. When you sell, he sends you one.

In the system I have in mind, a new microtx is sent every time the price changes. If the price follows a random walk, both parties have to sign bigger and bigger transactions, even though the price could have returned to the same value it was at the beginning. So I don't think this works well enough to be a practical solution.

It is a difficult problem. But there might be a solution which is practical enough, even if it is not 100% secure.

NxtChg

  • Overlord
  • *****
  • Posts: 1114
  • Respect: +61
    • View Profile
Re: A Strategy for Success
« Reply #35 on: March 02, 2016, 02:54:09 pm »
0
I don't think it is impossible, but it is a hard problem.  Whatever the solution is, I don't think it would have to work in all theoretical cases if it works with a high probability in normal cases.

It is. This is the whole idea behind solving the double-spend and the only solution we know is blockchain.

In the system I have in mind, a new microtx is sent every time the price changes.

I don't understand this at all. Why would txs be sent on price changes and not on buying/selling? What would it give you?

Plus, again, you have to trust your broker. In normal micro-payments model the trust is minimized, because the cycle "money<->goods" is much shorter.

In case of trading, you won't get whatever it is you're buying, like USD, gold or stocks.

Besides, if price changes not in your favor, you can simply refuse to re-sign tx, and the broker won't get his money.

So I really don't get it.
Tentacle Overlord, The Deranged Genius of The Abyss

draco

  • Minion
  • *
  • Posts: 25
  • Respect: +3
    • View Profile
Re: A Strategy for Success
« Reply #36 on: March 02, 2016, 04:58:39 pm »
0
Maybe this is not important if you are convinced that it can't work. But anyway:

I don't understand this at all. Why would txs be sent on price changes and not on buying/selling? What would it give you?


To prevent a client with a losing position from disappearing.

In case of trading, you won't get whatever it is you're buying, like USD, gold or stocks.

Correct. That is also common in the traditional markets. For example a futures exchange for agricultural products doesn't have agricultural products, and neither have most of the traders. If an american trades euros on the forex market, he never actually gets any euros. You can sell corn or cattle futures without having corn or cattle. 

Besides, if price changes not in your favor, you can simply refuse to re-sign tx, and the broker won't get his money.

The broker would immediately broadcast the last valid transaction. He would also need to charge a small fee to prevent somebody from systematically abusing the system.

NxtChg

  • Overlord
  • *****
  • Posts: 1114
  • Respect: +61
    • View Profile
Re: A Strategy for Success
« Reply #37 on: March 02, 2016, 05:12:13 pm »
0
To prevent a client with a losing position from disappearing.

Uhm, the most reliable way to do this is actually to take money upfront, like it works now.

Then you know the user has the money and won't run. And the user already trusts the broker.

Correct. That is also common in the traditional markets.

That's not the point. The point is users will have to trust the broker. This is no different than how it is now. No need for micropayments.

The broker would immediately broadcast the last valid transaction. He would also need to charge a small fee to prevent somebody from systematically abusing the system.

Prices can jump quite significantly. On a big position the broker can lose a lot.

I think you're trying to solve a non-problem. The current system works pretty well.
Tentacle Overlord, The Deranged Genius of The Abyss

NxtChg

  • Overlord
  • *****
  • Posts: 1114
  • Respect: +61
    • View Profile
Re: A Strategy for Success
« Reply #38 on: May 08, 2016, 07:41:10 pm »
0
"And yet, standing here today, I cannot point to a blockbuster company that has been built on blockchain technology. There has been no lack of trying. There is no lack of funding. I am a true believer and yet I can’t help but admit that despite all the hype, all the effort, all the capital invested, there has not been anything truly transformative to society that has been built on the blockchain, except perhaps Bitcoin itself and likely Ethereum. But these are enablers. What have they enabled other than grey and black market activities?"

Fred Wilson's talk, http://avc.com/2016/05/coin-center-keynote/
Tentacle Overlord, The Deranged Genius of The Abyss

NxtChg

  • Overlord
  • *****
  • Posts: 1114
  • Respect: +61
    • View Profile
Re: A Strategy for Success
« Reply #39 on: May 16, 2016, 10:39:04 am »
0
That's a good source of ideas: http://dapps.ethercasts.com/
Tentacle Overlord, The Deranged Genius of The Abyss

Titulaer

  • Jr. Minion
  • **
  • Posts: 88
  • Respect: +14
    • View Profile
Re: A Strategy for Success
« Reply #40 on: May 16, 2016, 12:03:12 pm »
0
That's a good source of ideas: http://dapps.ethercasts.com/

Wow... I didn't know there were so many Ethereum dapps in development! There seems to be a lot of confidence in Ethereum.

NxtChg

  • Overlord
  • *****
  • Posts: 1114
  • Respect: +61
    • View Profile
Re: A Strategy for Success
« Reply #41 on: May 16, 2016, 02:36:11 pm »
0
Wow... I didn't know there were so many Ethereum dapps in development! There seems to be a lot of confidence in Ethereum.

Replied in a separate topic: https://simtalk.org:444/index.php?topic=178.0
Tentacle Overlord, The Deranged Genius of The Abyss