I use Bitcoin-Qt + Armory. It's the only approach I've tried, and it seems OK. I picked it because I felt it important not to depend on other people's copy of the block chain. I don't use cold storage yet, but I'm glad it's available. I use DropBox for off-site wallet backups, and also have a copy on a micro-SD card that I carry with me.

Why now?

The current malleability attacks have done Bitcoin a favour, by exposing how serious the problem is while it's still relatively easy to fix. It's caused some bad PR, but the core system holds up and the real damage is limited. If this had happened in a few years, it would have been far worse. If you think the banks did it, then you have to conclude the banks want Bitcoin to succeed.

As soon as we're talking about customers with accounts, we're no longer in the same ballpark as the original post. I don't have an account at every place I buy coffee. I shouldn't need to. I don't have an account at the three supermarkets I use. If they want me to carry their loyalty card(s), that's up for negotiation. This stuff is clearly a layer on top of Bitcoin, and it clearly should be optional, only used where it adds value.

I think the idea is that often when you have an account, you don't have bitcoins any more. You have an IOU for bitcoins. There is a number of coins that you owe to the account, or it owes to you, and these can exist only in the account records, not in the block chain. The off-chain transactions are not subject to the same public scrutiny. For example, some people are concerned that MtGox have been indulging in fractional reserve banking. I don't think they have, but it's hard to prove without access to their internal accounts.

Personally I think FRB will inevitably happen sooner or later with bitcoin, for better or for worse.

You need to separate the "faster because of NFC" from the "faster because of payment processor". There's no reason why NFC can't be used for on-chain transactions. People are only using QR because it's currently more widely supported. Bitcoin isn't tied to QR codes.

The slowness is required as part of the "proof of work", which is at the core of Bitcoin's security design against double-spending. When the mining power increases, the protocol deliberately makes mining harder so it doesn't get any faster.

Other crypto-currencies use other times than 10 minutes. In practice using 5 minutes or even 1 minute doesn't help as much as you'd expect, because it's still too long to wait in retail. It has a downside in that it leads to more resources wasted on discarded blocks.

Neither does cash. Bitcoin behaves rather like cash in this respect. It has advantages over cash for the vendor: it's easier for them to count; they don't have to physically take it to the bank; it's harder for their staff to steal it. (The main reason supermarkets offer cash-back is as a way to get rid of the cash and replace it with something digital, because cash is such a pain for them to handle.)

The main drawback here is that if someone hands you a £10 note, and then changes their mind, you can give them the £10 note back. With bitcoin, you can't do that: you can't send them the output of the transaction they just sent you, because it won't be in the block-chain with enough confirmations. You'll have to give them the output from an older transaction instead. You'll need to keep a pool of old transaction outputs ready, to handle change. Much as you need to keep a pool of notes and coins to provide physical change when customers pay with cash. (Again, it's easier with bitcoin because it's digital; it doesn't take up space and all your tills can share the same pool.)

I'm not saying payment processors won't happen, just that they aren't needed.

Maybe Scotland.

(Scotland is considering breaking away from the UK. We (I am English) have told the Scots that if they do that, they'll have to give up any influence they have over the pound Sterling. So that could leave them looking for a currency. They might go for the Euro (and might have to if they want to be part of the EU), but frankly the Euro hasn't looked very attractive over recent years. It would be fun if they went for a crypto-currency. I suspect they'd carry on using Sterling for a few years regardless and then switch, so Bitcoin would have some time to mature. They may not break away at all; they decide in September.)

I agree, but I think it'll happen. From the customer's point of view, the payment page will have a combobox with a drop-down list of crypto-currencies that are accepted, and it'll be at most a couple of clicks to select the one they want to pay in. Hopefully BTC will be the default, and the one every vendor accepts, but they won't want to turn away customers who prefer some other currency.

Obviously the coffee shop will only accept transactions whose inputs are real, that is, in the block chain. If they aren't in the block chain, in general it can't be sure those input transactions were even submitted to the network.

A problem only arises if the coffee shop wants to let the customer use the change from a prior transaction at the same coffee shop. Which is a bit of a bodge, and unnecessary. The bodge doesn't help in the case where the customer buys coffee from one shop and then tries to use the change to buy a paper from the newsagents next door. To cope with the two-shops case, the customer's wallet needs to make sure it never tries to spend unconfirmed transaction outputs; and that also works in the one-shop case. There's no good reason to treat unconfirmed change differently to other unconfirmed transaction outputs.

It seems like a total non-issue to me, even if malleability isn't fixed.

Although I agree, it seems to me that once you are set up for one crypto-currency, accepting others is relatively easy. I imagine fairly soon we'll get payment processors offering a wide variety of currencies to vendors.

Um, no it doesn't? Malleability merely renames a transaction. The transaction itself is essentially unchanged. It still has the same inputs, outputs and amounts.  Renaming a transaction does not enable double-spending, nor does it allow bitcoin funds to be re-directed. It doesn't matter if the renamed transaction gets accepted by the network instead of the original, because they are the same. They just have different names. It's only a problem for exchanges that use the wrong name to lookup the transaction in the block chain. Coffee shops don't have to make that mistake.

Explain to me how you would use transaction malleability to defraud a coffee shop that allows zero confirmations.

If anything, the coffee shop is less affected, because the customer is not waiting around to ask for (or be asked for) a refund if the transaction is renamed, accepted, then not recognised under its new name. By the time transaction mutability comes into play, they'll be 10-minutes gone.

I don't follow. Why don't you charge the full £100 and then using it to buy £100 of BTC, instead of accepting £80 worth of BTC? Offering a discount for BTC means you end up with less BTC.

As an aside, your freedom to offer a discount for BTC is why Gresham's Law doesn't apply to bitcoin. There may come a time when such discounts are used to encourage bitcoin hoarders to spend. That will lead to more deflation, while increasing the velocity of money.

It wasn't really a design decision. Malleability is naturally allowed unless they take steps to prevent it, which they presumably didn't think was necessary at the time. Which it isn't, really. Only one transaction makes it into the block chain, and that has a single, unambiguous hash. Even now, they have taken steps to tighten it up, but it's not a priority for devs compared to more fundamental issues (like scalability).

That's not actually a Bitcoin double-spend, though. The second spend involves different bitcoins to the first. There are no double-spent coins in the block chain.

When you think you hold bitcoin in MtGox, you actually hold an IOU for the bitcoin, and you might be able to double-spend that IOU. That's a matter internal to MtGox. It's not a problem for the wider Bitcoin community, any more than MtGox getting hacked would be.

It's not true. Both versions of the transaction will have the same inputs, outputs and amounts; they are two different ways of expressing the same transaction, and only one will be accepted by the network, so there is no double-spend. No-one should care which version of the transaction gets accepted. (MtGox did care, and that's their mistake.)

Paper wallets are no different to other Bitcoin addresses in this. Each transaction has to take all the money at the address, but it can then transfer "change" back to the same address. So you can effectively withdraw a partial quantity.

(With digital wallets it's common to generate a new address for the change, but that's because some people think doing so improves anonymity, which they care about. It's not necessary.)

Why not?

If bitcoin is seen as the better currency, vendors will want to be paid in it, so they'll offer lower prices to people who pay in bitcoin. (Their freedom to do that is why Gresham's Law doesn't apply.) Then employees will want (some of) their salary to be paid in bitcoin, because it will have more spending power and be less subject to inflation. This applies to government employees also. The government with then have a need for large amounts of bitcoin to pay its employees and suppliers, and it will make sense for them to accept bitcoin in taxes as a way of getting it. The same logic applies to them as to anyone else.

Do you imagine a future in which the only role for dollars is paying taxes, and only government employees get paid in dollars? It so, I guess it will suck to be a government employee.

Where do they say they will fix the issue in two days? They've just said they'll give a public statement then. They have two days ascertain what has happened, decide what to do about it and how long it will take, and then draft a statement explaining this.

At the moment I tend to believe it's the technical issue described in the other thread. That is, they'd been issuing slightly invalid transactions, that were rejected by the network but that other nodes were cleaning up and re-issuing as valid but with different hashes. This meant the transactions were getting into the block-chain but MtGox didn't recognise them. This led to them double-spending the same coins, leading to more failures. It's going to take a while to sort out. They need to figure out which transactions they thought had failed, actually succeeded, and which genuinely failed due to double-spending, and they need to relate that to their internal accounts. It should be possible, though, and when they manage it, everyone should end up with the coins they own.

Isn't the real power in the hands of the miners? If a coin requires 51% of miners to accept an update, then we have a de facto voting on monetary policy, which could be an improvement over centralised control.

If I don't want people to know I'm buying porn, then I don't want to use a coin that has "sex" in its name to pay for it.

That seigniorage revenue is not free. It comes from devaluing the money held by the citizens. They could achieve a similar result, more openly and honestly, with taxation.

The difficulty of mining should adjust more smoothly than that. I doubt we'll see a sudden drop in difficulty due to a lot of miners giving up. New miners will stop joining the system because it won't be profitable enough to invest in new hardware. Old miners will tend to keep going, because they've already paid for their hardware. The ones who give up, because they can't even afford their electricity bills, will be the ones who have the least efficient hardware, so they will tend to be the ones who were making least contribution to the network, so their quitting won't have much impact. And of course, as each one quits, the remainder become more profitable.

Also, as transaction volume increases, the transaction fees per block will increase. This is for two reasons. First, there will be more transactions in each block. Second, there will be more users competing for the space in each block, so there will be more incentive for users to offer higher fees so their own transactions get processed quicker. So as long as Bitcoin is successful, there should be enough reward for mining to keep enough miners at it to discourage 51% attacks.

(Personally, I think governments and large companies will eventually start mining themselves, not for block rewards but as a kind of public good. America will want to be sure Russia can't mount a 51% attack. Microsoft will want to be sure Google can't mount one. One day, they will all be dependant on Bitcoin remaining healthy.)

I'm not seeing a problem specific to undeveloped countries. The difficulty of mining will self-adjust until the money raised from mining roughly equals the costs of mining. Then mining pays for itself, but isn't a profit centre. So the undeveloped country isn't losing out by letting developed countries do it.

Are you worried that developed countries will stop mining bitcoin? Surely Bitcoin has that covered, with transactions fees? Undeveloped countries can use bitcoin without worrying about maintaining the currency. (They will have to pay the transaction fees, but so does everyone else, and competition between miners should keep the fees to a minimal level.)

It does mean their government can't use inflation to transfer value from users of their currency to themselves. However, that's a dodgy thing to be doing. Arguably they will benefit more from having a stable, trusted, non-inflationary currency. It's dishonest governments that lose out.

That kind of reasoning is why some people think bitcoins are still massively undervalued. It suggests each bitcoin will be worth around 5 million dollars.