Instant POS

[mrdavis]

So I was thinking about a bitcoin POS system today. Perhaps something similar has been discussed or there is something I am overlooking, but I'll just descibe the method.

Customer with a bicoin client on their portable device goes into a local coffee shop. Their wallet is organized into a number of spending (one time use) addresses, much like I have an assortment of denominations currency in my real wallet. The total for the order comes to 3.5btc. The customer sellects an adress that has 5btc and a QR code is generated of the PRIVATE KEY and a "safe" that the customer has previously set. The merchant scans the QR code and imports the address with the already confirmed balance. The 3.5btc for the order is sent to another adress of the merchants choosing and the 1.5btc change is sent to the address provided in the QR code. A reciept of the transaction is printed for the customer and they leave with their order. Should something happen with the change the customer has a physical location/company to return to or point a finger at with a receipt stating that the transaction should have happened and a blockchain proving if it did or not.

[yogi]

That's not a bad idea.

The only thing I will add. If you have fixed denominations you might not always have the right amount. So you could add functionality for merchant to give change.

[BTCurious]

That's not a bad idea.

The only thing I will add. If you have fixed denominations you might not always have the right amount. So you could add functionality for merchant to give change.

Yes, the OP mentions giving a private key with 5 BTC, and getting 1.5 BTC back. This would indeed work, and would place the burden of trust with the customer, which will improve adoption by businesses.

[MoonShadow]

He did, but forgot to mention that the QR code must also include the bitcoin address for the change.  If a private key, a bitcoin address, and perhaps a small checksum of the first two, can be included on a qr code that fits onto a cell phone screen, this is a doable solution.

Dash7 would still be better, though.

[DeathAndTaxes]

That certainly works.  It does require you to trust the merchant but the "risk" can be minimized.  A smart wallet would always pick the "cash" private keys which result in the smallest change.  The wallet could inform you how much change you are "risking".

i.e. "paying 7.5 BTC expecting 1.5 BTC in change."  You know if you go through w/ that transaction at most you are risking  1.5 BTC.

Still for small face to face transaction a double spend is not a credible threat/risk and this setup is more complex, requires enough "cash" private keys, and still requires "normal" transactions too. 

Honestly I don't think it is necessary.

[BTCurious]

Still for small face to face transaction a double spend is not a credible threat/risk and this setup is more complex, requires enough "cash" private keys, and still requires "normal" transactions too. 

Honestly I don't think it is necessary.

At some point this could be automated, happening behind the screens. The user doesn't necessarily need to know how it works.

[DeathAndTaxes]

Still for small face to face transaction a double spend is not a credible threat/risk and this setup is more complex, requires enough "cash" private keys, and still requires "normal" transactions too. 

Honestly I don't think it is necessary.

At some point this could be automated, happening behind the screens. The user doesn't necessarily need to know how it works.

I guess more simplistically.  This provides additional protection for small face to face transaction.  Those transaction are the ones least likely to be double spent (and almost impossible to accomplish successfully) and thus there doesn't seem to be much value in it.

[btc_artist]

This is a good idea.  The QR code (or some other machine-readable code) could also be printed out on slips of paper (and you could carry around a bunch of them with different amounts of BTC) and held up to a scanner as an alternative to having a smartphone client.

[wareen]

I'd consider making the receipt a digitally signed message from the merchant and displayed as a QR code back to the customer who scans it with his smartphone. Obviously these digital receipts need to meet certain legal requirements - I think most countries already have regulations for that.

Would also work nicely with NFC - then it could be made really transparent for the user. The smartphone app could also instantly check the validity of the receipt by comparing the amounts, check if the signature from the merchant is valid and listen to the blockchain for the transactions.

[SgtSpike]

Still for small face to face transaction a double spend is not a credible threat/risk and this setup is more complex, requires enough "cash" private keys, and still requires "normal" transactions too. 

Honestly I don't think it is necessary.

At some point this could be automated, happening behind the screens. The user doesn't necessarily need to know how it works.

I guess more simplistically.  This provides additional protection for small face to face transaction.  Those transaction are the ones least likely to be double spent (and almost impossible to accomplish successfully) and thus there doesn't seem to be much value in it.

+1, I agree.

[BTCurious]

Still for small face to face transaction a double spend is not a credible threat/risk and this setup is more complex, requires enough "cash" private keys, and still requires "normal" transactions too. 

Honestly I don't think it is necessary.

At some point this could be automated, happening behind the screens. The user doesn't necessarily need to know how it works.

I guess more simplistically.  This provides additional protection for small face to face transaction.  Those transaction are the ones least likely to be double spent (and almost impossible to accomplish successfully) and thus there doesn't seem to be much value in it.

I concede, you have a point.

[becoin]

So I was thinking about a bitcoin POS system today. Perhaps something similar has been discussed or there is something I am overlooking, but I'll just descibe the method.

https://bitcointalk.org/index.php?topic=28565.0

[DeathAndTaxes]

Still for small face to face transaction a double spend is not a credible threat/risk and this setup is more complex, requires enough "cash" private keys, and still requires "normal" transactions too. 

Honestly I don't think it is necessary.

At some point this could be automated, happening behind the screens. The user doesn't necessarily need to know how it works.

I guess more simplistically.  This provides additional protection for small face to face transaction.  Those transaction are the ones least likely to be double spent (and almost impossible to accomplish successfully) and thus there doesn't seem to be much value in it.

I concede, you have a point.

As someone else pointed out it might be more useful for those without a smartphone.  Think of them as print on demand bitcoin "bills". 

You have your wallet create the "cash" private keys (want to make sure they are segregated so you never use them again), load some currency into them and then generate the QR code along w/ human readable note.

You print these out and put them in your wallet.  Hand one to a merchant, he scans it, verifies the balance and sends you the "change".  You still need to trust the change is "good" but then again that always exists w/ cash. 

[amincd]

How is this more safe from a double-spend? Couldn't the buyer pull off a Finney attack by finding a block, including in it a transaction that transfers the bitcoins on the private key QR code to another address, and not broadcasting the block until he's left the store?

[MoonShadow]

How is this more safe from a double-spend? Couldn't the buyer pull off a Finney attack by finding a block, including in it a transaction that transfers the bitcoins on the private key QR code to another address, and not broadcasting the block until he's left the store?

It likely doesn't change the risks of a Finney attack in any meaningful way, either good or bad.  Such an attack requires 1) a large minority percentage of the hashing power 2) excellent timing and 3) luck.  It's also very easy for the party that has been defrauded to know what happened, even if the transaction is a fairly anonymous in-person transaction.  It's extremely unlikely that such an attack would even be attempted for anything except a highly portable and high value item, and there are not many that fit that bill that an anonymous buyer can just walk into a store and buy with cash and walk out.  Certainly not without  being on video.  There is literally no single item in a Wal-Mart store that would qualify.  Maybe a gold coin dealer, but those guys don't even assume that the cash is genuine.  I know of no case of anything that you could buy for cash and carry that would remotely justify the resources required to even attempt such an attack.  Although a finney attack can theoretically reduce the percentage of network hashing power to perform a double spend to less than a majority, the risks of such an attack are IMHO remote.  Anyone selling such a high value item is just going to insist upon confirmations, and you will get used to the idea.  We wait a half hour or more to buy a new car while they check our credit, why would it be different with bitcoin?

[amincd]

In that case a normal bitcoin transaction is just as instant as this one. The only possibility that a bitcoin transaction will reverse after it's been registered by the payee is that the payer is going to double spend through a Finney or >50% attack, which are impractical.

I do see some cases where this method would be useful though. For one, if you had no connection to the net or no wallet installed on your phone, you could still spend bitcoins.