Bitcoin Developer Guide
bitcoin.org
bitcoin.org
Edit: thanks all for replies, learned a lot.
Bitcoin exposes "scripting" of the blockchain to enable decentralized consensus for arbitrary transaction types. With scripting, developers can leverage the blockchain to store a decentralized consensus of their custom transaction types. Often times this is how escrow-like services are implemented. It's also possible to modify the proof of work scheme itself (as we did with torcoin, albeit in theory only thusfar.)
The extensible part of Bitcoin is the blockchain, which is useful because it's a decentralized, publicly verifiable, mathematically provable consensus.
> Bitcoin doesn't actually solve the problem of needing trust it does solve the problem of needing to trust your escrow agent with your money but then again if you don't trust your escrow agent with your money why do you trust them to make a fair decision.
So I'm still confused here. What new kind of contracts could I build using the blockchain? What existing contracts are made easier because of the blockchain?
Two people want to exchange funds for service or goods. As usual, the seller wants to know that the funds will be available when the goods are delivered. Both people hire an escrow representative (similar to how houses are sold today) and all four create a transaction that binds the buyer's money into a contract; if any two of the remaining parties either enter complete or exit the transaction, then the money is either disbursed to the seller and service providers, or the money is returned to the buyer.
This way, the buyer's money is bound into the contract and the seller can determine this to be the case before releasing their product or service. The escrow representatives are trust companies and can see for themselves whether the transaction is progressing as promised or not due to their connections with the buyer/seller and with each other.
Have you ever wired money before? Either you haven't and are believing someone else who has told you it takes weeks or you are misrepresenting it here. I've done wires internationally between AU, CA, UK, and US and internal to each country between a variety of banks and in the past 10 years I've never had a wire transfer take more than 3 days(one off) with all the rest completing the next business day at the destination or 2 days later when I've sent it after business hours at the source.
And fees. $40 on a $1000 wire is %4. $1000 is a small amount of money to wire. If you're moving around $millions then sure, you don't care. I don't move around $millions.
Here we are talking about turn around times in under an hour, and fees under %1, even for small transactions.
What are these extra steps you're talking about? If I wire money to you as my escrow agent it will probably get to you next business day. You then verify whatever and send the escrow to the seller that will take another business day. Where are the weeks coming from?
>And fees. $40 on a $1000 wire is %4. $1000 is a small amount of money to wire.
Get a different bank if you're paying $40 for domestic transfers. Hell even if you're getting charged that much for international transfers.
>Here we are talking about turn around times in under an hour, and fees under %1, even for small transactions.
Except we aren't talking about comparable things since the end result is generally local currency not bitcoin you have to add in the steps and fees to convert it as well.
If I'm going to sell you a TV I don't care if you have the money I want the money before I send you something.
This problem is already solved by having the people pay up front.
A: creates a hash of a contract B: creates a hash of the same contract
A sends 0.1BTC to B (and in the transaction message includes the hash). B receives the money and replies with 0.1BTC and the same hash (and the bitcoin notary keeps a cut). A year later, A claims that the contract said "any member of the board" instead of "most members of the board". So we go to the blockchain and verify what the contract said, and how the two people agreed through a transaction.
What does Bitcoin add? You can't erase the hash of the document to forge any part of it, and you can't erase the signature (transaction from A->B->A).
In this case you would still depend on either a 3rd party escrow person to make the ultimate decision or trust between the two parties.
Bitcoin doesn't actually solve the problem of needing trust it does solve the problem of needing to trust your escrow agent with your money but then again if you don't trust your escrow agent with your money why do you trust them to make a fair decision.
So I'm still confused here. What new kind of contracts could I build using the blockchain? What existing contracts are made easier because of the blockchain?
This way I can authorise payment without entrusting the funds to a third party. If the delivery company doesn't authorise the payment, the money is still immediately available to me.
Sure, it'll get resolved eventually...
This doesn't solve a real problem which is the problem with a lot of things bitcoin tries to solve. They are things that aren't real problems or they are things that are already solved just as well.
This kind of issue almost always requires a third party reputation broken to step in, and today it is one of the reasons I like shopping on Amazon. Complain to them, or file a rating with 1 star, and you certainly get a vendor's attention. Commerce just shouldn't be this difficult. Buying products online should be as easy and as trustworthy as buying products over the counter. I see Bitcoin as that mechanism. We aren't going to all trust Amazon for ever.
Because you're the one making the claim. Making a charge back takes roughly 2 seconds. And if you've bought something online you should have an email receipt you can find just as quickly.
You've spent more time away from your family replying to me telling me how important your time with them is than you would spend making a chargeback.
>This kind of issue almost always requires a third party reputation broken to step in
Right and they already exist so bitcoin doesn't actually improve the process it just reimplements it.
>Buying products online should be as easy and as trustworthy as buying products over the counter. I see Bitcoin as that mechanism.
But bitcoin actually makes it worse for consumers. Credit cards have really strong consumer protection methods. Bitcoin has none unless you toss on 3rd party escrow in which case it is just trying to patch over an inherit problem. Bitcoin is digital cash. The reason cash works well is because almost every time you use it you are in person with the merchant and can verify delivery immediately. Once you remove that it is near worthless.
This is at least 60 seconds. At best. And normally, much longer, especially if you add up the time you spend meticulously filing all receipts.
I don't think it is asking too much that a third party (like FedEx) engage in transaction assurance for Internet purchases. They can get a slice of the pie (they do anyway.)
http://en.wikipedia.org/wiki/Hyperbole
>Meticulously file your receipt for purchase.
I use gmail and type in the domain of the company and I have my receipt. If you can't do that you should upgrade your email provider.
That's not how I thought multisig works. I thought the delivery company would have to authorize giving the money back to you.
If the risk is that well defined, then the cost of escrow seems wasteful if you could just automate the upside/downside in the Blockchain (not necessarily well defined yet). I don't think the package delivery service is even close to an interesting example, so I don't know why anyone would choose it. Packages need to be delivered. Automated contracts can only eliminate insurance middle-men, not physical logistics.
BTC advocates should focus on useful ways to simplify existing contracts (while paying very close attention to the profit structure of underwriting).
You don't pay for them to calculate the risk though you pay for them to resolve the issue when there is a dispute. So using bitcoin wouldn't actually lower that cost/allow you to automate away any of that.
>Automated contracts can only eliminate insurance middle-men
The middle-men in most contracts are there for assurance. Bitcoin doesn't provide that even with m-of-n.
With a correctly setup transaction, bitcoins can be held in escrow that require 2 of 3 people to approve moving it. This means that you/seller, you/escrow, or seller/escrow are the only 3 combos to get the bitcoins out. There's no way for the escrow agent to steal the money.
>There's no way for the escrow agent to steal the money.
If you're worried about an escrow agent stealing your money if they had the chance then they really really should not be your escrow agent.
This is as good (and better) than current escrow services as it is decentralised, between the contracting parties only, and can have very fine grained conditional statements or scenarios before a 'contract' is executed.
In the example you gave, the escrow could consists of you, me, and a computer program. You sign to say you've sent it, the computer program I used to open it signs it, and I also sign it when i view it. So in this scenario, you could get paid when 2/3 parties have signed, or have a staggered payment structure of 50% when 2 have signed, then the other 50% when the remaining person has signed. This would be escrow, but escrow based on the quality of a digital good.
Hope that helps :)
https://en.bitcoin.it/wiki/Script
One of the things that didn't make sense to me initially was how the Bitcoin network could "know" that a condition had been fulfilled. For example, paying a dowry requires a marriage to have taken place. Likewise, an escrow contract might require proof of delivery of an item.
How could a contract know that the condition had been met?
Oracles offer a solution. The Orisi project has created a framework for setting up and using Oracles for smart contracts:
https://en.bitcoin.it/wiki/Contracts
Quoting from the top:
A distributed contract is a method of using Bitcoin to form agreements with people via the block chain. Contracts don't make anything possible that was previously impossible, but rather, they allow you to solve common problems in a way that minimizes trust. Minimal trust often makes things more convenient by allowing human judgements to be taken out of the loop, thus allowing complete automation.
Let's take one example, 2-of-3 dispute mediation. Elsewhere in this thread, some people are saying things like "if you don't trust your escrow agent with your money, why do you trust them to make a decision". These are very different kinds of trust. Typically online payment networks bundle them together. You trust your bank and credit card company with your money (more accurately, you don't trust them, but you know ultimately the costs of their failures will be socialised), and then because they are always the middleman you end up trusting them to do dispute mediation as well.
The problem is, they often aren't very good at it. Companies like PayPal tend to see dispute mediation as a cost of business rather than their core competency. Innovation in this space is rare. PayPal's policies, for example, make selling something to someone else who is standing in front of you impossible - you must be able to prove you mailed the item and if you can't, you automatically lose the dispute and the payment will be reversed. This policy is so deeply unintuitive that there's a whole industry of scammers out there who steal things from people by exploiting it. There isn't much scope here for doing anything better: payment networks tend to have a small set of one-size-fits-all policies and you can't go to the market to find a new one.
Now consider the Bitcoin approach. For many transfers there is no possibility of dispute mediation and thus none of the overheads associated with it. This often makes sense: either you know there will be no dispute (e.g. a gift from a family member) or you have other ways to resolve disputes (e.g. salary payments). When you want dispute mediation however you can agree on a mutually satisfactory mediator.
This mediator has very limited power. Contrary to what other people here seem to think, this does matter. For example, the mediator cannot steal the money, thus they make a poor target for hackers. If the mediator is a company there is very little temptation for bad insiders. They are not accepting deposits so all the complicated regulations that are triggered by that act don't apply. There is no possibility of them lending the money in escrow out and going fractional reserve.
Really all they can do is resolve a dispute badly, but there's no profit in that. So by preventing the mediator from accessing the money, all kinds of possible failure modes are avoided. In turn this means you can access a far wider set of possible mediators. For example, if you are trying to buy an antique violin over the internet PayPal's mediation policies are less than ideal:
http://www.theguardian.com/world/2012/jan/04/paypal-buyer-de...
With Bitcoin you could ask an actual expert in antique violins to act as the mediator. They do not need to be a computer security expert, they don't need to worry about becoming a target for physical theft or extortion, they don't need to worry they might accidentally lose the keys and destroy the money. There's tons of things they don't need to worry about. They could (if it existed) just use a friendly desktop app to pick a winner in case of a dispute. That means they can spend all their time inventing precise and well designed ways to resolve any dispute, fully leveraging their domain knowledge.
That's just one example of how redistributing and carefully controlling trust relationships can result in better outcomes.
Except there is profit in that if one side has paid them to resolve the dispute badly. Which is exactly back to where you are in the current system so you haven't actually solved or improved on anything you've just changed it.
Systems that enforce lock-in of mediators will be pushed out of the market when more open systems come online.
Also, both sides put up some collateral, which both forfeit in the case of an unresolved dispute. A scammer will still pay some value into the multi-party transaction, so operations will not last long if many disputes go unresolved.
All the bitcoin escrow solutions are just a variation on "I want to use this 3rd party to ensure trust but I don't trust them enough to actually do so". It doesn't work like that though so you end up with "I want to use this 3rd party to ensure trust but I don't trust them enough to actually do so but I will end up having to".
>Also, both sides put up some collateral, which both forfeit in the case of an unresolved dispute. A scammer will still pay some value into the multi-party transaction, so operations will not last long if many disputes go unresolved.
So if I am scammed I not only lose out on the scam but I am further punished by the protocol to lose this collateral as well?
Well, that's the Bitcoin philosophy in a nutshell: they solve the problem of being screwed by third-parties by arranging the system so that when you inevitably do get screwed, you will have to sagely acknowledge that it was your own fault the whole time.
The math is perfect after all, which means the system is perfect, so by reductio ad absurdum if you get a crazy result (e.g. being screwed by a mediator) it must be due to a crazy axiom on your part (e.g. your decision to use a shitty mediator).
Now, doesn't that make you feel better? Welcome to the Jungle; you lost your money fair and square, and you have only yourself to blame.
Unless they're governments /s
Yes, you still need some trust that the mediator is not corrupt and actually knows how to resolve disputes. You obviously can't pick someone completely at random. But, finding someone who is not corrupt and has specific domain knowledge is still easier than finding someone who is not corrupt, has specific domain knowledge AND is able to safely handle your money in escrow, which is what you'd require otherwise.
Trust isn't a black and white thing. It's a very complex and subtle topic. People can be trusted in some ways and not others.
Still, let's take a look at how we can address your concern in ways that don't involve funny attempts to arrange incentives via coin deposits and things (I was never keen on such protocols myself).
The simplest and most obvious way to avoid a corrupt mediator being paid off is to have a mediator who agrees to enforce some (mostly) predictable rules, instead of it being purely a matter of opinion. For example a mediator who offers to mediate mail order sales could publish the evidence they will use to resolve the dispute, and the act of making a payment involving them would involve effectively signing a contract with them (obviously you try and set up the software and contracts such that it's pretty standardised and all done with one or two mouse clicks). Now the mediator has to produce not just a decision but also the rationale for how they arrived at that decision. If they can't do that, the mediator will quickly gain a reputation as being unreliable and to be avoided.
This is the same approach as used by the courts. People take disputes in front of judges all the time, in the expectation that the judge is not corrupt. One reason they can reasonably expect this is that judges aren't supposed to just pull an opinion out of their ass, they are supposed to show how they are applying and interpreting the law with references to other cases where there is ambiguity, or at least an understandable rationale for the decision when it genuinely strays into new territory.
Another way to reduce the trust needed in mediators is to use machines instead of people. This is the "oracle" concept described in the wiki page. The idea here is that a machine uses trusted computing to perform a remote attestation of the software it's running. For example using Intel TXT / SGX, or AMD SKINIT. This lets you know how a dispute will be mediated with the precision of computer code. It can be a good approach if a contract or dispute is precise and measurable enough to be decided by programmatically, although again, these techniques are quite advanced and nobody is using them today.
Except you're not since no one tries to be their own bank they have their bank to safely handle the money for them. As an escrow agent I have Alice send money to me and I choose to send it to Bob or back to Alice but at no point in that am I typically holding actual cash. They are funds in my bank account which I trust to manage funds.
>For example a mediator who offers to mediate mail order sales could publish the evidence they will use to resolve the dispute, and the act of making a payment involving them would involve effectively signing a contract with them (obviously you try and set up the software and contracts such that it's pretty standardised and all done with one or two mouse clicks). Now the mediator has to produce not just a decision but also the rationale for how they arrived at that decision. If they can't do that, the mediator will quickly gain a reputation as being unreliable and to be avoided.
This doesn't remove the problem of a malicious 3rd party though. I could say "I arrived at the decision by inspecting the box from Alice and seeing the product before sending it to Bob". I have my process, I inspect the box and make sure the products in it match the order. But I am still capable of lying to you. Bitcoin still hasn't solved that problem and hasn't actually improved on the existing trust protocol.
>these techniques are quite advanced and nobody is using them today.
There is also an extremely limited range of contracts which could be solved by this type of system and none of them that I can think of that people do today.
We're talking about Bitcoin here :)
I'm not saying you are a libertarian, but this is an argument that libertarians use a lot. I don't buy it. Look at the companies, organizations, and people out there today who deserve bad reputations, but don't have them. Ask most random buyers what they think of PayPal, even, and they won't realize there is anything wrong with it. Hell, a surprising chunk of sellers don't even have a problem with PayPal.
Unfortunately 2 years later, we're not much further in those developments. Bitcoin has severe limitations, also because it has now a legacy to support. New projects can start with a fresh design. So far we've only seen very few new approaches. While many Bitcoiners realize the imperfections of the system, very few actually think there could be something better.
PayPal started as a security company, back when they were above the bike shop on University Avenue in Palo Alto. Their original business was selling crypto tokens. They wanted to store money on PDAs.
Then they discovered that business model was going nowhere, and went to a more centralized system for sending money. The big problem remained security. Most of the effort goes into preventing theft and fraud, not merely sending the money.
http://bitcoinmagazine.com/10468/daos-scary-part-1-self-enfo...
I've always wondered, would it be possible to make the difficulty setting dynamically determined based on some econometric derived from the bitcoin chain? (such as if you see the economies velocity slowing down, perhaps you lower the threshold) or the opposite.
If it was, that may be one way to stabilize the prices, at least of the currency itself.
This would make for different amounts of bitcoins being released rather than the (fairly) predictable current schedule.
This is kind of the point. A predictable release of new bitcoins into the economy has no basis in good economics.
But, the bitcoin people have philosophical reasons to avoid that.
[1] https://blog.ethereum.org/2014/11/11/search-stable-cryptocur...
Disclaimer: I own some BitShares. I won't answer more questions here to keep things on topic, so email those to me.
http://chimera.labs.oreilly.com/books/1234000001802/index.ht...
It is available to read online for free.
Seems like this will be a very useful thing with more people starting to use bitcoin!
The current plan seems to be to start increasing the block size limit by 50% per year in order to allow higher transaction volume. Block size limit has always been a contentious issue.
Just to clarify for anyone new to bitcoin... blocks aren't "issued" every 10 minutes exactly. On average there is a new block found every 10 minutes (or so).
There's also: https://bitcoin.org/en/developer-reference - though I'm not quite sure how the two relate.