Bitcoin does has fees for transactions. Currently they are negligible, but only for high currency value transactions, and only because miners are subsidised by inflation. If you want to spend under a dollar in BTC, you will probably need to pay a fee which will be quite a high proportion of that dollar.

Yep. That's an important consequence of it being distributed. Another consequence is that the rate of creating of new money is fixed, so no-one can devalue the currency by printing more money.

A third, related consequence is that Bitcoin enables "permissionless innovation". Anyone can build a new service on top of the block-chain, rather like how the WWW is built on top of TCP/IP. That's why the Bitcoin protocol is sometimes compared to the internet. In the long run permissionless innovation may be the most important thing. The block-chain is about tracking ownership, and currency may merely be its first application.

It depends on how cold you want it. I considered Pi, but to do anything with it I'd need to hook it up to keyboard and screen, which wouldn't be convenient for me.

Would there be transaction fees? If so, then isn't that what Bitcoin will turn into once the block reward drops to insignificance?

When you say "private network", do you mean a distributed trustless system, or is it a ledger owned by the company that created it? If the latter, there is no need for the proof-of-work system and they can just trust themselves to update the ledger honestly.

I use an Armory cold wallet, which is encrypted by Armory, then encrypted again by 7Zip, then replicated to various places including DropBox.

We shouldn't rely on personality. For all we know, Satoshi died months ago, and his keys just now fell into the hands of his heirs, who have different personalities and goals. Or maybe a thief broke into his home and stole his paper wallet. Satoshi's past behaviour may not be a good predictor of what a new owner will do in future.

There's an argument that if the alt-coin is inflationary, then the allotment for bitcoin addresses should be in proportion to the number of alt-coins that exist when they are claimed. That's if you don't want to penalise bitcoin owners for waiting (as has been mentioned in some posts). There is also an argument that they should be penalised for waiting.

In practise I'd expect most bitcoin owners to ignore the alt-coin launch, either through shear ignorance (not seeing launch publicity), or else feeling they can't be bothered to track every alt-coin that comes along. If, after a few years, the alt-coin succeeds and comes to their attention, that's when they might claim their new coins. Unless you address this, you will find you are actually rewarding those supremely indifferent bitcoin holders.

You may also find it is a source of instability for the alt-coin. The effective number of alt-coins could suffer large increases at any time, due to previously passive bitcoin owners making alt-coin claims and then dumping them to buy more bitcoin. Even if they don't dump, and hold onto their claimed alt-coins, the currency would surely be very inflationary due to these influxes. Of course, we might hope they would manage their claims so as not to crash the currency they are acquiring, because that would devalue their own new holdings; but people aren't always smart, and anyway they might see the alt-coin as a threat and want to disrupt it. In particular, if the alt-coin uses proof-of-stake, I wonder if this could be a source of 51% attacks.

So for these reasons it might be best to make the proportion of alt-coins per bitcoin reduce over time. Give the biggest rewards to those who support the alt-coin earliest.

I'm not advocating this; just putting the idea out there. One drawback is that it's more complicated. Another is that it diminishes one of the virtues of the spin-off idea, which is that it reduces the threat of the alt-coin to Bitcoin. Spooking Bitcoin investors and venture capitalists by destroying its value could poison the well for all crypto-currencies; a fantastic alt-coin that caused a bitcoin crash could make everyone unwilling to trust any crypto-currency as a store of value in case the same thing happens again. So the way spin-offs preserve bitcoin value is important.

Also, when a new user adopts the alt-coin and becomes active with it, they necessarily increase demand for it, and the coin's value goes up. This deflationary effect may offset the inflationary effect of the new alt-coins generated when they claim, so it the inflation might not be as bad as I thought above. This is only true if their behaviour is broadly the same, and spend/hold the alt-coin in similar proportions to their bitcoin; and it won't be true if they dump and return to bitcoin.

Well, rather than targeting Satoshi personally, you could limit claims to coins from recent transactions. For example, with a hard cut-off you could only give alt-coins to bitcoin users who had been spent or bought in the previous year, in proportion to their spending/buying. These active users are arguably the people who will most benefit the new alt-coin. Or, you could use a softer approach, by awarding more alt-coins to the newer transactions with some more or less complex formula. You don't need to know or care who owns the coins.

(Again, not advocating this, just pointing out some more possibilities if you don't keep a fixed proportional premine.)

I don't really agree, but this is mostly a quibble about terminology and what is meant by a "fork", so it's probably not worth pursuing.

Here I am nit-picking at a rather minor element of the proposal, so this may also not be worth pursuing. (But it's more substantive than terminology.)

Do you mean from the largest market cap, or to the largest market cap? The latter makes more sense to me. For fairness I'd prefer the network based on the most recent known good block-chain, and for greed I'd prefer the network that gave my coins the largest real-world value. If there were two candidate alt-coins, and one launched just before I bought a house and one after, then I might prefer the older one because it would give me more coins.

I do think a block-chain that is years old is probably not recent enough.

Well, merely "lasting" doesn't mean that, but if it is the most successful alt-coin then presumably it would be.

I'm not yet sure I do want aethereum to be a viable crypto-currency. I'm still thinking about it.

That's basically what happens when Bitcoin forks, isn't it? After the fork there are two block-chains, and people who owned coins before the fork now own coins in both block-chains. The alt-coin is a more radical fork than the Bitcoin community would normally consider, but conceptually it's still a fork of Bitcoin.

That's mainly helpful if there isn't much growth in the user base between the fork and the attack. If æthereum launches, and, say, 2 years goes by, during which Bitcoin sees a 10-fold increase in users, and then the attack happens, 90% of Bitcoin users will be holding no addresses that they can turn into æthereum coins.

(Numbers picked for simplicity. I expect and hope to see exponential growth in the Bitcoin user base over the next several years. Even if it's 4x in 5 years rather than 10x in 2 years, the benefit of what you say here will fade.)

Most of the practical benefits are for people who don't have credit cards. Which usually means people in countries other than America. We can also hope that the retailers will pass some of their benefits on, in the form of cheaper prices.

I think a lot of westerners support Bitcoin for ideological reasons. Some see it as an investment and/or a hedge against inflation. Those two combine: they don't want to have to trust the government to inflate its way out of debt. Some think it gives them privacy.

Fees are necessary to discourage spam and denial of service attacks.

If it weren't for that, we could eliminate fees by making the currency inflationary forever. Miners could be rewarded at a rate of, say, 1% of minted coins per year. In effect, everyone would be paying them to validate transactions through the devaluation of their held coins. Some economists prefer this model because they think the inflation encourages spending which encourages economic activity and wealth-creation. (Others dislike it because it encourages consumption, and because of edge effects where-by you can't set interest rates to below zero.)

Alternatively, the actual costs of mining could be reduced by switching to a proof-of-stake system. Then we wouldn't need ridiculously powerful computers to mine on, and ridiculous amounts of electricity to run them. You'd still need some way to discourage denial of service attacks, though.

It is indeed fixed in 0.91-beta. (At least, it let me send to two addresses; I didn't try three.)

Do you want updates on what other bugs are or aren't fixed in this version?

Although it can't actually cause loss of coins. It can only cause problems for local clients, and I think the pastebin over-states the effect.

My Win8.1 PC reported a virus detected in the Bitcoin database today. I just marked it as "Allowed" and then told it to ignore the Bitcoin directory there-after. If I'd picked the default action of "Quarantine" instead, I imagine I'd have been able to unquarantine it later, or else just download the block again. As it is, not only am I fine, but I can re-broadcast the block to anyone else who needs it. As long as one person has a copy, we're fine. The crypto means the block can't be forged. No coins are going to be lost.

I have wondered about this. How hard would it be for companies that are backing Bitcoin to switch horses? Let's say Google invents a new proof-of-stake currency and Amazon start accepting it - and it's technically excellent.

I suspect that at a programming level, a lot of the code for wallets etc could be reused for the new currency. Especially if the new currency was designed to make it easy, and the wallets were written in a modular way. Address generation etc could be the same. Surely people could switch in 3-6 months if they had to?

That's true. Arguably it would not be in the best interests of someone promoting a new currency, to crash the bitcoin economy. It would undermine the whole concept of crypto-currencies. If my hypothetical Google/Amazon conspiracy happened, they'd probably include a way to convert bitcoin into the new currency.

True. Although I currently think proof-of-stake surely has to win in the long run, and a forward-looking technological company might decide that the sooner we adopt it, the better. The longer we wait, the harder to switch, and the more energy (of various sorts) spent/wasted on Bitcoin.

I suppose I think that if Bitcoin survives the next 2 years, it'll be so entrenched it'll probably survive the next 50.

Surely thumb drives and paper wallets could be copied?

Eventually we may have battery-powered secure hardware wallets that can exchange transactions via NFC or Bluetooth. Without a network there'd be no way to confirm them, so you'd have to trust the buyer wasn't double-spending, but standard hardware wallets wouldn't generate double-spends.

There are if you use OP_RETURN. Everyone can verify from the block-chain that they became unspendable.

Why? They aren't dangerous, like drugs or guns or alcohol or tobacco. They will probably be auctioned off, as if they were the proceeds of crime, eg like when a drug-dealer's car gets seized.

Yes. There are three main ways to get rid of unwanted coins:

• Send them in a transaction that has a spending script that always returns false. Technically, consists only of OP_RETURN. This is provable and does exactly what you want.
• Send them in a transaction with zero outputs. This doesn't destroy them, but it distributes them in a more or less random way, to the miner who mines that block. It's better - less deflationary - for the community. We would want to be sure they weren't broadcast in a way that favoured a miner who was connected to the public servant charged with getting rid of them.
• Send them to a nonsense address for which it is unlikely that anyone knows the private key. This isn't quite provable, although if they use an address like 1crimedoesntpaycrimedoesntpayguys and it turns out someone has the private key, they probably deserve the money.

Yes.

It's value as an investment is predicated on it being useful in every day life. However, in practice today, there isn't much you can do with it that you can't do at least as easily with a credit card. For me it's interesting for intellectual and philosophical reasons, and I expect/hope it will eventually be more practical.

Success of a rival crypto-currency. If someone like Google or Amazon decided to support Litecoin and not Bitcoin, for example. That seems unlikely today. Bitcoin has all the momentum; for good or ill it has all the visibility and PR, and it's where the venture capital money seems to be going. However, that could change very quickly.

They can regulate it but not stop it, and generally they have no valid reason to stop it.

Read a lot.

Short answer: yes, stay away. Long answer: don't invest more than you can afford to lose. Bitcoin is high risk. There's a real chance it will crash suddenly and in 6 months will be worth a fraction of its current value. It's not a safe and easy route to make enough to pay off your debt.

Perhaps. Today probably not, but if it takes a few years to pay off your debts, the situation hopefully will have changed by then.

It depends on how you manage your private keys. Most thefts have happened because people let websites manage the keys, and the websites got hacked. Some thefts have happened because people stored their keys themselves, and had bad security - not using encrypted wallets, or having a compromised machine with a keylogger that took their password. Good security is possible, but takes knowledge and effort.

It would be a lot of work, and time. If/when you succeed and have a successful site, chances are eBay will start taking bitcoin. You lose your USP and your site has no role any more, so it dies. It'd be better to invest the effort into an idea with long-term prospects.

People who say eBay will never allow bitcoin are being naive, in my view. It's just another currency for them. If anything, their buyer-wins adjudication policy is even more important as a layer over Bitcoin because raw Bitcoin transactions are not reversible. If you want an eBay-like that doesn't have that policy, that favours the seller, then it might have long-term prospects but it's not actually specific to Bitcoin. (And it's not clear that the business will succeed, because you arguably need buyers more than sellers and buyers are less likely to use a site they don't trust because it favours sellers.)

I did try Ubuntu 10.04-32bit (I forget whether I got it from your link or elsewhere). It would not install on my device. My SSD was too small. It's a touch under 4GB. The device is an Asus EEE PC701. (It also has a small screen, 800×480.) Part of what made the attempt frustrating was a lack of documentation about what the hardware requirements actually are, but as far as I can tell most versions of Ubuntu with a GUI need at least 5GB and 800x600.

Having done this recently, one practical hurdle I found was getting a Linux distro that would run on the offline device, and then getting Armory to run on that distro. I gave up on the first step after a few failed attempts, partly because I figured that for the second step, I'd need to rebuild Armory from source, and it seemed like too much hassle, especially if the rebuild was to happen on the offline computer.

(Getting a clean WinXP isn't easy now either, since Microsoft have long stopped selling it and I wouldn't trust a third party distro not to come with keylogger pre-installed.)

That does sound like a good approach. It should be simplest for the user (provided the cut-down distro would run on almost anything) and wouldn't involve extra hardware or faffing about getting the machines to see/hear each other.

One drawback for me is that I'd presumably lose the ability to use my password manager. Currently I have the same password database on both online and offline machines, so a shared master password. I believe a password manager is crucial for practical password-based security, and I'd rather not have to learn a password just for Bitcoin. So being able to run certain 3rd party software on the offline machine can be helpful.

Using a smartphone intermediary could be more convenient for QR codes. It'd be easier to hold the smartphone so its camera can see the online computer's screen than to hold a laptop, and easier to move between devices if they are on different desks.

Even better might be to use the smartphone as the offline computer. Old smartphones are almost as common as old laptops. Take the SIM out and switch off networking, and use either QR codes, sound or a micro-SD card for I/O. Alternatively combine this with the previous approach, and have a custom build of Android that just does Armory. Either way the small form-factor would be a bonus.