Bitcoin Forum

I'm wondering what the core devs' opinion is of adding multiple units of account, atomic transfers etc to Bitcoin?  Think Colored Coins+. 

This is what I think:

It seems like the greatest risk of an alt-coin replacing bitcoin is that bitcoin does not handle different units-of-account.  This seems to be a key feature of most if not all of the non-copycat alt-coins.  The foundation of economic behavior and contract law is exchange, not transfer and so bitcoin seems very deficient.  To truly capture economic activity, and live up to its "electronic cash" moniker, Bitcoin needs atomic exchange between different units-of-account. 

The ongoing (and recent) failures of centralized exchanges have shown the difficulty of centralized approaches.  And even without failure, they had significant issues in terms of choice (can't exchange any pairs), price-and-time efficiency, and regulatory load.  This is why bitcoin was decentralized in the first place.

I understand the reluctance to make changes that would cause a fork in clients that have not updated, but this can be and has been managed.  It will be especially manageable when it becomes generally understood that the real question is not "to fork or not to fork", but whether the fork that brings the promise of digital currencies to full fruition is Bitcoin v0.9.5 or SomeOtherCoin v0.1.0.  I think that we should be confronting this today.  Once an alt-coin has made significant strides, IT will have the network-effect advantage, not Bitcoin.


Yes I know all about colored coins.  Its a good start but the key feature is exchange, not coloring.  Colored coins do not really support anonymous atomic exchange.

1. We need colored coins+ natively in the Bitcoin core wallet (satoshi client) to send the message that this concept is officially supported.
2. We need EVERY wallet to at least respect (and not touch the coins of) colored designations.
3. We need 1 satoshi transfers.  Either make an exception for color designated coins, or use the size of the txn fee to decide to relay.  The txn fee is the "right", market based way, if I'm willing to spend 1000 Satoshis to transfer 1 Satoshi, why should miners deny me?
4. We need colored annotation on every txout.  Today, wallets can operate on just the txout set.  They don't need the entire blockchain history.  We need to keep it that way or even make it simpler to help hardware wallets.
5. We need script primitives that support atomic exchange and a colored spend fee (like a miner's fee, but to the colored coin creator).  These details are better described in a full spec.

With these features companies could issue their stocks, bonds, etc natively on the blockchain.  Other companies could "back" crpyto-<thing> with real <thing>, and receive a small fee for their service. 

With these features, Bitcoin can actually DO everything the pundits have been promising...

Yes trust is still involved, but it is significantly reduced.  This is ok.  Trust is intrinsically involved for many units-of-account.  For example, holders of a mortgage instrument are trusting that the owner won't trash the house.  There is no way to avoid this.  In fact, its hard to argue that bitcoin-the-currency won't out-compete every trustless (and therefore unbacked) colored coin.  What do these other coins have that bitcoin does not?  So I think that bitcoin & coins based on some degree of trust will be the only ones with economic significance on the bitcoin blockchain.

BUT basic exchanges become places where buyers and sellers meet; they actually never touch the money.  At the same time, there is still a niche for centralized, high frequency trading.  This mirrors what is good about our economic system today.  I CAN exchange stock by hand with a random person on the street, but a centralized exchange makes it more convenient.

This is not an overlay network that is a parasite on bitcoin-the-network at the expense of Bitcoin-the-currency.  Bitcoin remains the fundamental digital crypto-currency facilitating the exchange of special-purpose units-of-account. 

Bitcoin is decentralized. There is no official. If you want it so much, fork the source code and persuade people to use your branch

That's a long road full of uncertainty for both Bitcoin and Bitcoin2.  Bitcoin the code is freely available, and Bitcoin the network is decentralized, but Bitcoin the product is very much controlled by those with commit access to the github repository.  Better to persuade them first. 

Frankly I don't see how much persuasion is needed.  Exchange is the essential part of the economy.  Alt-coins are implementing it.  Bitcoin must keep up or become obsolete.  The valley of the hype cycle is a good time to work on this kind of thing...

 

I don't think we need any of these things, in fact I think they're outright undesirable. I'm open to be convinced otherwise, but a list of demands doesn't do much for me— implementations of interesting tools do.

Generally "colored coins" have huge hype inertia, but in virtually every proposal to make them do something useful that I've seen you can generally apply some modest transformation to eliminate the colored coin and end up with the same, or better, security model and usually better scalability.

A lot of what pundits promise are pointless, nonsensical, dangerous, or outright impossible. Thats what pundits do.

Decenteralized things are great and essential and are most of my motivation in the Bitcoin space— but colored coins are usually superfluous to creating trustless exchanges and usually hurt scalability at the same time.

Regardless of all this, you aren't going to make any progress with this approach. If you go author useful tools and people use them, great. If that happens it doesn't matter what I think and I suspect that along the way you'll learn the truth of what I'm saying in a more convincing and efficient way than anything I could write here.

I completely agree with gmaxwell that we don't need these things.  The original Satoshi concept is powerful because of its simplicity.  



I disagree.  The value of bitcoin is in the strength of our community and our shared agreement that the blockchain is correct.  Any alt coin that actually shows promise can be made more appealing by relaunching it with an initial distribution of coins equal to the unspent outputs in the bitcoin blockchain.  

We are calling this concept a "spin-off": https://bitcointalk.org/index.php?topic=563972.0 (https://bitcointalk.org/index.php?topic=563972.0).  Unlike side-chains, spin-offs are independent of the bitcoin network.  

Is bitcoin becoming obsolete?  Someone should tell Bitcoin.
This "must keep up" notion is absurd on its face.  Bitcoin is not running behind a bunch of failed experiments and new attempts that may or may not work or prove viable trying to catch up.  The experiments are excellent and worthy and laudable, but it is senseless to chase every new idea.  Bitcoin is not the place for implementing rapid application prototyping.  Take some time to figure it out first before breaking one of the few things that already work.

Embrace, extend, extinguish strategy (http://en.wikipedia.org/wiki/Embrace,_extend_and_extinguish) is bad enough when done from the outside.  Lets not do it to ourselves.

thanks for your replies.  I am not advocating for any small feature but for the ability to exchange which is fundamental to economics.   Since the technology already exists in ethereum there is little value in making another altcoin tool unless there is an intention to merge to bitcoin.  but the list was certainly not meant as demands just discusssion points to explore the issue. sorry for any misunderstandings there.

gmaxwell: I am not aware of existing solutions with non parasitic bitcoin blockchain use.  can you send me link or quick summary.

It sounds like this guy might benefit from reading up on sidechains. I think it will be very interesting in the time to come to see how they affect all this alternative features hoopla. This might deflect some of the criticism of the core code, which is meant to be stable and slow moving as opposed to feature-driven sidechains where beta testing of new concepts could occur. This would meet the issues addressed by GMaxwell above where developers can demonstrate utility rather than bitch about what the core Bitcoin code is missing.

Thanks for the advice, I noticed last night that it does look like side chains are gaining some renewed attention, due to gmaxwell's clever "spend back" solution that takes advantage of merged mining.  This feature really would allow services to be layered on top of bitcoin in a manner more analogous to TCP/IP, including a chain with the multi-currency features that I am proposing.  In fact, I'm thinking of making a summary topic here since the information about this idea is hard to find and not cohesive.