Decentralized BTC Stock Market [Goodbye GLBSE]

[odolvlobo]

tl;dr: Bitcoin - viable.  Decentralized stock exchange - fundamentally flawed.

Double think is strong in this one. Ignorance is strength.

I tend to agree here. The implementation of distributed arbitration in the bitcoin protocol is simply genius. Likewise, distributed arbitration for other transactions could be implemented for securities, somehow. It might be difficult or non-intuitive, but you can't just claim that it would be impossible.

BTW, there is a major problem with using bitcoins (e.g. colorized bitcoins) to represent securities. There is a limited number of bitcoins, but a potentially unlimited number of securities. The solution is to create a separate block chain for securities with no limit on the number of securities that can be created.

[casascius]

The court will sign a message, declaring a coin decolorized, and send it to the blockchain. You don't need a centralized database.

And the nodes maintaining the blockchain will accept it why?  Because the court clerk controls the majority of CPU power on the network?  Because the court holds a special private key that gives it an elevated status over the rest of the network like RealSolid and his SolidCoins?  What will ensure that a duly authorized court can do this, but not you or I?  What happens if a court misuses its discretion, misapplies the law, or simply fat-fingers an entry and signs something it totally shouldn't have? (something today's courts manage through the judicial review process, i.e. appellate courts etc.)

How about when two different courts with equal ability to simply "sign" and "send" things into the blockchain staunchly disagree on an outcome and see no reason to stop "signing" it into the blockchain in a back-and-forth tug-of-war?  Imagine people edit warring on a hypothetical Wikipedia with no 3RR (3-revert-rule) and no administrators, when does it end?

You have a big mistake here: you compare decentralized stock exchange with a real stock exchange. However, the right comparison should be decentralized stock exchange vs. GLBSE. A decentralized exchange does everything GLBSE does, without the extra risk of hacking etc.

...and of course, without any meaningful accountability imposed on the issuers of security for the benefit of their shareholders either.  By pointing out the differences between GLBSE and a decentralized exchange, you suggest your understanding is that I am describing a fundamentally technological problem and meanwhile I am describing a fundamentally sociological one that would affect GLBSE and a decentralized exchange equally.

[casascius]

tl;dr: Bitcoin - viable.  Decentralized stock exchange - fundamentally flawed.

Double think is strong in this one. Ignorance is strength.

Go ahead and create it then, and let me know how it works out.  Neither of us are presently in a position to convince the other, so I'm letting it go.  To paraphrase a saying commonly applied to bitcoin: if I don't think it'll work, then I should come back in a few years and either say "told-ya-so", or start using what should be by then a more mature and robust decentralized stock exchange complete with safeguards that ensure that issuers are accountable to their shareholders for their statements and actions.

[phelix]

[...]

You have a big mistake here: you compare decentralized stock exchange with a real stock exchange. However, the right comparison should be decentralized stock exchange vs. GLBSE. A decentralized exchange does everything GLBSE does, without the extra risk of hacking etc.

...and of course, without any meaningful accountability imposed on the issuers of security for the benefit of their shareholders either.  By pointing out the differences between GLBSE and a decentralized exchange, you suggest your understanding is that I am describing a fundamentally technological problem and meanwhile I am describing a fundamentally sociological one that would affect GLBSE and a decentralized exchange equally.

What about the stock issuer making himself accountable on his own?

What about separation into a decentralized exchange and (several) third party IDing and stuff?

[molecular]

hey guys, interesting discussion!

For me the situation presents itself as follows: we need 3 things:

• decentralized ownership-tracking (e.g. by using colored coins in blockchain) and way to proove ownership (for voting purposes, for example)
• decentralized exchange (order matching)
• accountability / link to real world of asset issuer

I think the first 2 points can be solved satisfactorally given enough resources.

3rd point is the hard part (I agree with casascius that if there is no real-world-link of the colored coin to some enforcable contract, this whole idea probably only makes it easier for scammers. I acknowledge the possibility of independant private dispute resolution as a service)

To solve "link to real world" problem, let me suggest a market for centralized agencies which offer the following services:

• Make accessible a list of "colored coins"/assets in existance
• Record the identity of the real-world person issuing some asset and offer disclosure to a court in case of dispute
• Safely store (in a legally binding manner the contracts associated with the respective assets and ensure these are legally usable in case of problems. Offer relevant services to courts (identify asset issuers, provide means of asset holder to proove ownership of asset at certain point in time)

I think this could solve casascius' "accountability"-objection, no?

Maybe the service would have to have the issuers actually real-world-sign some contract in which they agree to subject ownership of the asset to the mechanics of ownership-transfer using colored-coins / smart property.

[jl2012]

And the nodes maintaining the blockchain will accept it why?  Because the court clerk controls the majority of CPU power on the network?  Because the court holds a special private key that gives it an elevated status over the rest of the network like RealSolid and his SolidCoins?  What will ensure that a duly authorized court can do this, but not you or I?  What happens if a court misuses its discretion, misapplies the law, or simply fat-fingers an entry and signs something it totally shouldn't have? (something today's courts manage through the judicial review process, i.e. appellate courts etc.)

If the message is embedded in a standard and valid transaction, I can't see why miners will refuse to include it in a block. If many miners do so just because they don't like the message, they would also refuse anyone's transaction because they don't like him/her. If many miners do so, that's simply 51% attack and bitcoin is fucked up

How about when two different courts with equal ability to simply "sign" and "send" things into the blockchain staunchly disagree on an outcome and see no reason to stop "signing" it into the blockchain in a back-and-forth tug-of-war?  Imagine people edit warring on a hypothetical Wikipedia with no 3RR (3-revert-rule) and no administrators, when does it end?

The terms of the agreement could indicate which jurisdiction it is bound to.

You have a big mistake here: you compare decentralized stock exchange with a real stock exchange. However, the right comparison should be decentralized stock exchange vs. GLBSE. A decentralized exchange does everything GLBSE does, without the extra risk of hacking etc.

...and of course, without any meaningful accountability imposed on the issuers of security for the benefit of their shareholders either.  By pointing out the differences between GLBSE and a decentralized exchange, you suggest your understanding is that I am describing a fundamentally technological problem and meanwhile I am describing a fundamentally sociological one that would affect GLBSE and a decentralized exchange equally.

Technology ALWAYS moves faster than social norms/rules/laws, e.g. contraception, test tube baby, doping in sports, BitTorrent, child porn on TOR, and needless to say: bitcoin. Decentralized exchange (DE) is a technological advance on top of GLBSE. DE reduces the counterparty risk by removing the role of exchange operator, and does equally good (or bad) in enforcing the contract between issuers and shareholders. However, they allow people to invest in something that is otherwise impossible. The investors should balance the risk and benefit

[FreeMoney]

The court will sign a message, declaring a coin decolorized, and send it to the blockchain. You don't need a centralized database.

That's absurd, how will I know which key(s) belong to the court, and why will I care and how will I know who agrees?

[jl2012]

The court will sign a message, declaring a coin decolorized, and send it to the blockchain. You don't need a centralized database.

That's absurd, how will I know which key(s) belong to the court, and why will I care and how will I know who agrees?

All these information could be included in the agreement of the security

[FreeMoney]

The court will sign a message, declaring a coin decolorized, and send it to the blockchain. You don't need a centralized database.

That's absurd, how will I know which key(s) belong to the court, and why will I care and how will I know who agrees?

All these information could be included in the agreement of the security

Oh, doing it voluntarily with complete foreknowledge on a case by case basis seems fine. And leaves room for people to do whatever they want.

[killerstorm]

That's absurd, how will I know which key(s) belong to the court, and why will I care and how will I know who agrees?

Suppose colored coin represents ownership of property according to some contract. E.g. contract says "I owe money to bearer of that coin", basically.

Court might simply announce that it won't enforce that contract. I.e. for a particular coin contract is invalid.

Then you can still trade that coin, and issuer might pay you. But if you go to that court, it would not accept your claim. Maybe some other court will, who knows...

Anyway, people who trade coins are very interested to listen to these invalidation claims. And it naturally makes sense to publish them in same place where contract is published. Because it becomes pretty much a part of contract.

So it is decentralized: whoever is interested would store this information. It is compatible with uncertainty of court system decision: even if X says contract is invalid, you can go with it to Y.

And it is compatible with colored coin system as you can re-assign different color to invalidated coins, and so your client will not accept them being the same as original.

It is ugly, though, but legal system in general is ugly.

So there is a theoretic framework to do this, but I'm not saying that it should be done. It's more like "our system supports this feature too, if you really want that".

[killerstorm]

For me the situation presents itself as follows: we need 3 things:

• decentralized ownership-tracking (e.g. by using colored coins in blockchain) and way to proove ownership (for voting purposes, for example)
• decentralized exchange (order matching)
• accountability / link to real world of asset issuer

Yes. This is what I was talking about from the start. Let's decouple it into many things and let each part to evolution independently.

3rd point is the hard part...

To solve "link to real world" problem, let me suggest a market for centralized agencies which offer the following services:

• Make accessible a list of "colored coins"/assets in existance
• Record the identity of the real-world person issuing some asset and offer disclosure to a court in case of dispute
• Safely store (in a legally binding manner the contracts associated with the respective assets and ensure these are legally usable in case of problems. Offer relevant services to courts (identify asset issuers, provide means of asset holder to proove ownership of asset at certain point in time)

Yes, this is what I was talking about. I don't think it's particularly hard, it is just a bit expensive, maybe more expensive than size of the market would support now... Essentially you need a bunch of lawyers (I assume ones which now do tax planning, offshore companies, investment banking perhaps) and a little bit of tech. The only problem is that high-end lawyer won't be interested in doing this if he gets lavish salaries on the Wall Street.

We have a lot of passionate tech guys here, but not enough passionate lawyers/investment banker

Anyway, it can work exactly like investment banking works in real world: to issue shares or debt you need to go to investment banking guys and they'll arrange it for you.

But of course different such companies will offer different level of service. Underwriting debt for a large company is different from personal loan or small startup crowdfunding.

[jtimon]

• accountability / link to real world of asset issuer

3rd point is the hard part (I agree with casascius that if there is no real-world-link of the colored coin to some enforcable contract, this whole idea probably only makes it easier for scammers. I acknowledge the possibility of independant private dispute resolution as a service)

I will try it again, but I feel I'm being ignored.

1) The only thing needed is to trust the issuer of the asset.

2) Legal binding may be necessary for that trust to exist in some cases.

3) Digital signatures can be legally liable.

4) You can just bind the digital signature used on colored coins to a legal contract.

So I see the 3rd part as really simple technically speaking. You only need to explain this to a lawyer so that he writes such a legal contract and voi la: you're legally liable for the assets you issue with colored coins, (or ripplecoin or two-phase distributed ripple).

By the way, I still see "not having to modify the current chain protocol" as the only advantage that colored coins have over ripplecoin.
Having to represent the assets with underlaying satoshis is just an artificial limitation for the system.

[casascius]

I will try it again, but I feel I'm being ignored.

I don't think you are being ignored, with all due respect, I think you are simply wrong and don't yet understand why.

1) The only thing needed is to trust the issuer of the asset.

Yes, that's correct, especially if the goal is to get scammed.  Finding people to trust issuers isn't the problem, there's a long line of those just waiting to be separated from their money by scammers, even here on the forums.  Just because you trust something doesn't mean it's trustworthy.

2) Legal binding may be necessary for that trust to exist in some cases.

No.  This is wrong.  Legal binding is necessary in all cases (except when the goal is to get scammed).

Without it, any issuer can simply say "OK everybody, bend over, you're screwed" and there's absolutely nothing you can do.  They can do that even if that's not what they originally planned when they started, before ever being tempted by an actual opportunity to grab lots of money and run wild and free.  That's not security.

The fact that some people still get their money back despite the lack of legal protection has nothing to do with whether it's necessary.  You can also close your eyes and drive through a red light and sometimes nothing will happen and you'll get where you're going earlier than normal, that doesn't mean stopping for red lights is unnecessary.  In the case of ponzi schemes, scammers deliberately pay some people back plus generous profit to attract more investors/victims with the illusion that everyone will be paid back with generous profit.

3) Digital signatures can be legally liable.

Not if you can't identify the person who made it, nor prove that the private key belongs to them and only to them.  Even then, if the person denies making the signature, the cost of proving otherwise is not insignificant, and the person/people making the decision (judge/jury) will likely know nothing about PGP or cryptography.  You'll need to pay an independent third party expert to testify to the authenticity of that signature for it to be believed.

4) You can just bind the digital signature used on colored coins to a legal contract.

If the legal system disagrees with the conclusion of ownership represented by the colored coins, the legal system will disregard the meaning of the colored coins and substitute its own judgment, rendering the colored coins meaningless.  We don't need to wait for the future to find out if this is for real: lawyers and the legal system can already tell you this today if you ask.

So I see the 3rd part as really simple technically speaking. You only need to explain this to a lawyer so that he writes such a legal contract and voi la: you're legally liable for the assets you issue with colored coins, (or ripplecoin or two-phase distributed ripple).

And then, since you've clearly identified yourself, you're also on the hook for selling unregistered securities to the public and become a takedown target - exactly the reason why the decentralized exchange would be invented in the first place.

[jtimon]

1) The only thing needed is to trust the issuer of the asset.

Yes, that's correct, especially if the goal is to get scammed.  Finding people to trust issuers isn't the problem, there's a long line of those just waiting to be separated from their money by scammers, even here on the forums.  Just because you trust something doesn't mean it's trustworthy.

2) Legal binding may be necessary for that trust to exist in some cases.

No.  This is wrong.  Legal binding is necessary in all cases (except when the goal is to get scammed).

It depends on the asset. Ripplepay, villages, LETS, CES and other mutual credit systems IOUs work perfectly without legal liability.
Not sure other private currencies like miles or linden dollars are linked to a legal contract, yet people accept them.
For stocks you're probably right, you need legal liability to be able to trust the issuer.

3) Digital signatures can be legally liable.

Not if you can't identify the person who made it, nor prove that the private key belongs to them and only to them.  
...
You'll need to pay an independent third party expert to testify to the authenticity of that signature for it to be believed.

Obviously the legal contract needs to identify the liable actor. That independent third party can operate really cheaply, it only needs legal digital signatures. In some cases it can be the state itself. As said, in Spain all citizens have and ID card that enables digital signature and is legally binding.

4) You can just bind the digital signature used on colored coins to a legal contract.

If the legal system disagrees with the conclusion of ownership represented by the colored coins, the legal system will disregard the meaning of the colored coins and substitute its own judgment, rendering the colored coins meaningless.  We don't need to wait for the future to find out if this is for real: lawyers and the legal system can already tell you this today if you ask.

I'm talking about a legal contract that makes colored coins MEANINGFUL AS OWNERSHIP CERTIFICATE. I don't need to ask any lawyer, without that legal contract, there's no legal ownership. What we need to ask lawyers is what should such a contract contain.

So I see the 3rd part as really simple technically speaking. You only need to explain this to a lawyer so that he writes such a legal contract and voi la: you're legally liable for the assets you issue with colored coins, (or ripplecoin or two-phase distributed ripple).

And then, since you've clearly identified yourself, you're also on the hook for selling unregistered securities to the public and become a takedown target - exactly the reason why the decentralized exchange would be invented in the first place.

Yes, you can't have both. Either you're legally liable or anonymous.
But the great thing is that the market gets decentralized. But as always you need to trust the issuer. This is not like bitcoin where the issuer is the chain itself.

[casascius]

I'm talking about a legal contract that makes colored coins MEANINGFUL AS OWNERSHIP CERTIFICATE. I don't need to ask any lawyer, without that legal contract, there's no legal ownership. What we need to ask lawyers is what should such a contract contain.

I hear you and agree with you, but then have to ask what is the purpose of the decentralized colored coins?  The purpose is defeated if the ownership certificate defines the ownership and not the colored coins.

The apparent purpose of the decentralized colored coins is to allow the asset to be exchanged and owned anonymously.  If you've followed E-Gold, the powers that be don't like that, which you may be aware of, since that's why you'd be discussing inventing a decentralized exchange.

Suppose I announced, "hey everybody, my name is Mike Caldwell, and I have X ounces of gold here in Sandy, Utah, USA.  All Satoshis numbered <insert list here> are hereby colored and represent legally binding claims on my gold... PGP signed, me.  Happy trading."  Then I'd have a functioning system exactly as you suggest.  And then someone from the government would be kicking down my door and taking the gold and wiring me with a smart ankle bracelet for the same reasons they did the same to E-Gold.

http://en.wikipedia.org/wiki/E-gold

[jtimon]

It would be something more like...

my name is Mike Caldwell, and I have X ounces of gold here in Sandy, Utah, USA.  All Satoshis numbered <insert list here> coming from this bitcoin public address Addr1 are hereby colored and represent legally binding claims on my gold... PGP signed, me.  Happy trading.

It seems the US is not the "freest country in the world" anymore though. You should issue your eGold somewhere else.

The powers that be may launch a rainbow bomb if we win on the internet, but should we stop writing free software with that fear in mind?

[Energizer]

I am currently experimenting/designing a new asset exchange protocol that will take place on top of the block chain for assets issuing/buying/selling/...

My design is radically different from any current colored coins design. In colored coins the asset starts being tracked from a root transaction where all coins in such transaction is considered to be colored. In my design there is no such a thing called colored coins.

The protocol will treat btc addresses as nodes in a network. Where transactions are gonna be used as network events being sent from one node to another "BTC address to BTC address". All communication is going to go through a single point of interaction which is the "issuer's BTC address". The history of all transactions made to the issuer's BTC address will be parsed as issuing/ask/bid/buy/sell/motion/buyback, in a manner that after parsing the whole transactions history made to and from the issuer's BTC address the application can determine which btc addresses owns how many shares, which btc address is selling shares and @ what price and so on. In simple words the design will relay on tiny payments being sent to the issuer address as protocol events/service fee. The payments made to the issuer's address are gonna be treaded as messages/events where the issuer can benifit from it as operation fees. There will be no worries about assets being mixed "multiple colored coins transfered in one transaction". Cause the design by its nature will hook all the asset activities to one BTC address "the issuer's". In fact there will be no colored coins at all. A BTC address can hold 1000 shares of an asset where its btc balance is = to zero.

[jtimon]

I am currently experimenting/designing a new asset exchange protocol that will take place on top of the block chain for assets issuing/buying/selling/...

My design is radically different from any current colored coins design. In colored coins the asset starts being tracked from a root transaction where all coins in such transaction is considered to be colored. In my design there is no such a thing called colored coins.

The protocol will treat btc addresses as nodes in a network. Where transactions are gonna be used as network events being sent from one node to another "BTC address to BTC address".

I was assuming colored coins did exactly that all along.

There will be no worries about assets being mixed "multiple colored coins transfered in one transaction". Cause the design by its nature will hook all the asset activities to one BTC address "the issuer's".

I don't think it is really important. You only mix colors if you want, and that's just stupid.

In fact there will be no colored coins at all. A BTC address can hold 1000 shares of an asset where its btc balance is = to zero.

Mhhmm. Without modifying the protocol?
Can you elaborate on this?
That sounds basically like ripplecoin.

[killerstorm]

By the way, I still see "not having to modify the current chain protocol" as the only advantage that colored coins have over ripplecoin.
Having to represent the assets with underlaying satoshis is just an artificial limitation for the system.

Well, another advantage is that it is easier to implement atomic coin swaps in one blockchain than implement cross-blockchain trade.

Atomic coin swapping can be implemented with today-enabled features and it is inherently 100% secure.

Cross-blockchain trade requires support for 'contracts' and non-traditional scripts, and it has different security considerations:

1. You want 'the payment' to get several confirmations before you finalize the trade. Otherwise counterparty can pull off a double-spend. So it is inherently slower, i.e. you need at least 10 minutes for the trade.

2. If counter-party did not finalize the trade, your money will hang until time-out expires. So a sort of DoS attack is possible where attacker would make many trades without finalizing them, thus paralyzing counterparty funds for some time.

3. Since it depends on timeout, there are probably some sorts of race condition attacks. For example, attacker might finalize the trade around the timeout, so he will get counter-party's funds via the trade and his funds back via timeout.

But in a long term, I guess, we would want a separate chain just for colored coins, and perhaps it should follow ripplecoin semantics. But we don't know yet all requirements for "colored coin chain", so it's better to wait with it.

There is no reason why several implementation can't work in parallel: e.g. Colored Bitcoins for those who want to trade in Bitcoins, and Colored Ripplecoin for wider spectrum of things.

[killerstorm]

No.  This is wrong.  Legal binding is necessary in all cases (except when the goal is to get scammed).

This is bullshit. The fact that many agreements are honored without legally binding contacts prove that legally binding contracts are not necessary.

Maybe you're saying that statistically people have better luck with legally binding contracts than they have without them, but then you need to show us this statistic, otherwise it is simply unfounded claim.

I don't know any such statistics either, but from personal experience, I never had problems with informal agreements. I work mainly as a contractor, and whenever clients agree to pay me for work they do. I never had any single client which didn't pay, and I never signed anything in advance. Likewise a lot of my friends work in a similar fashion and have no problems either.

I was scammed just once in my whole life, and ironically I even had an IOU from this person: that person was a criminal, as it turned out, and I didn't want to get beaten.

Without it, any issuer can simply say "OK everybody, bend over, you're screwed" and there's absolutely nothing you can do.

That's the problem with your reasoning: there will be consequences for issuer. At very least nobody will trust him again.

In the worst case... Have you ever heard about internet lynch mob? If somebody knows your real name and you've pissed them off it might be very, very bad for you.

Maybe it doesn't happen in your alternative reality, but it happens in the real world.

I'm not saying that it is a right way to resolve conflicts, but it surely is a possibility, and you have to acknowledge it.

The fact that some people still get their money back despite the lack of legal protection has nothing to do with whether it's necessary.
  You can also close your eyes and drive through a red light and sometimes nothing will happen and you'll get where you're going earlier than normal, that doesn't mean stopping for red lights is unnecessary.

But here we know that being careful and following the rules is statistically much safer than doing otherwise.

But it's far from certain in case with legally binding contracts since people lose their money despite these contracts ALL THE TIME.

It has been demonstrated that sometimes removing traffic lights is better for everybody, both in terms of security and in terms of convenience: http://www.youtube.com/watch?v=lwHfibl1AoI

It would be foolish to simply abolish all rules, everywhere. But sometimes experiments make sense and they produce positive results.

Not if you can't identify the person who made it, nor prove that the private key belongs to them and only to them.

Somehow certification authorities have no problem doing extended validation (http://en.wikipedia.org/wiki/Extended_Validation_Certificate). It's pretty much same thing.

Even then, if the person denies making the signature, the cost of proving otherwise is not insignificant, and the person/people making the decision (judge/jury) will likely know nothing about PGP or cryptography.

Well, hello? People use digital signatures to interact with banks and tax authorities, but for bond market it is a problem?

You'll need to pay an independent third party expert to testify to the authenticity of that signature for it to be believed.

And it is a problem why?

If the legal system disagrees with the conclusion of ownership represented by the colored coins, the legal system will disregard the meaning of the colored coins and substitute its own judgment, rendering the colored coins meaningless.

Legal system can disagree with the conclusion of ownership represented by paper signed contracts, and? Does it make paper signed contracts meaningless in general?