1. Propose a terrible idea and ask people to tell you why it's terrible
2. Be told why it's terrible
3. Get pissed
4. ? ? ?
5. Profit
2. Be told why it's terrible
3. Get pissed
4. ? ? ?
5. Profit
pretty much this
Topic: Proposed solution to "lost coins" (Read 4214 times)
pretty much this
I will be lazy instead of learning to read the code. Can the block size increase dynamically, or is there a hard-wired cap?
The block size is 1MB. However, it used to be 32MB.
You could search for max block size or something to get the discussion.
Transactions are the real limit for the network. Every transaction has to be verified by someone.
If you keep the blocks at 1MB or lower, then everyone can easily verify and mine new blocks.
A larger block size may shift the balance to those with faster internet connections.
However, a larger block size keeps transactions fees low, which is one of the main draws of bitcoin.
The average blocks are around 0.5MB at the moment, so not quite critical yet.
The security model of bitcoin assumes network bandwidth is cheap, but hashing is expensive.
I will be lazy instead of learning to read the code. Can the block size increase dynamically, or is there a hard-wired cap?
blockchain space doesn't relate to coins?
There is 1 MB of data per block at the moment. The space requirements are determined by the number of transactions.
Lost coins don't add any extra data to the block chain.
They have to be available to be retrieved though. However, really old coins could be stored on some disk somewhere.
blockchain space doesn't relate to coins? I guess I am running up against the limits of my technical knowledge of btc. i assumed that you need at least one / 100,000,000 coins to transmit any information over the btc network. if only a small number of these dust units are accounted for in the world, say 50 years hence, then there will have to be a lot of workarounds accreted to core bitcoin to make it work for more than the most valuable transactions. but then, i suppose we should expect that these problems will be overcome. anyway, i am probably not describing this correctly b/c i don't understand the software fully. my point is that the bandwidth over the network, however that is defined in the code, tends to shrink over time (but admit that this is probably solvable, and i don't know enough about the technology to contribute to that discussion)
But actually what happens is you approach an asymptote of peak oil, and alternative energy sources replace it in the market.
That is what it means to say we won't run out. The price goes up, so people who really need it (or more to the point will pay). There is still always oil available to buy, but it is expensive.
Quote
Or with bitcoin, the cost of using available blockchain space becomes very high and alt chains become more attractive.
Block chain space has nothing to do with the number of lost coins.
Joel, I don't think the problem is solved. Rather we're arguing over whether it's a problem. As far as I am aware, the only solution to lost coins is that BTC are divisible to high precision. The argument being that even if there are only 10 BTC left on wallets for which people remember the key, then there are actually 10 hundred million units available. And btc-using apps will adapt to this. But I find this argument lacking, much as the economic argument that oil is not scarce (ie, there will always be some oil b/c as the price of oil responds to scarcity, market, and extraction difficulty, there is always enough oil to meet demand). But actually what happens is you approach an asymptote of peak oil, and alternative energy sources replace it in the market. Or with bitcoin, the cost of using available blockchain space becomes very high and alt chains become more attractive.
Bitcoins are infinitely divisible. Just update protocol to store value in a 128 bit integer instead of the current 64 bit integer. That would satisfy the decimal place need for a loooong time. When you bring back lost coins, you dilute the aggregate value of everyone's bitcoins, similar to what happens when the Fed prints USD.
I see this come up a lot, Bitcoins are not infinitely divisible.Quote from: Bitcoin Specs
Each bitcoin is subdivided into 100 million smaller units called satoshis, defined by eight decimal places.
0.00000001 is the smallest amount you can spend. If you remove the decimal point it becomes easier to visualize with whole numbers.So if we changed the protocol to use a 128-bit integer for the Tx value as opposed to the current 64-bit integer, you are saying we couldn't use the extra bits for more decimal places? Why not? Yes, we would also have to update the "scaling factor," but we coud keep 1 BTC = 100,000,000 satoshis, and 1 satoshi = 100,000,000 or so integer units of the 128-bit integer.
Think of it like this: 1 BTC represented as a binary 64-bit integer (how it is now):
Code:
0000000000000000000000000000000000000101111101011110000100000000
1 BTC represented as a binary 128-bit integer:
Code:
00000000000000000000000000000000000000000000000000000000000000000000000000100011100001101111001001101111110000010000000000000000
21,000,000 BTC represented as a binary 128-bit integer:
Code:
00000000000000000000000000000000000000000000000000101100011110000001111101110000100011000101000010011111010000000000000000000000
So you see, with a 128-bit integer and the scale factor 1 satoshi = 100,000,000 128-bit integer units, we can add more decimal places. There are still many unused bits on the left, so we could use a larger scale factor for even more decimal places.
I think there's a general issue with "can" as it is used both in the present and future time. At the moment (i.e. right NOW), you "can't", but given a move to 128-bit integers, then you can.
BTW - I liked your explanation there. Nice and clear.
There is a real issue that gradual waste of bitcoins over many years will eventually, by laws of mathematics, result in there being almost no bitcoins remaining on wallets where people can recall the keys.
That's a solved problem. The real issue is the costs of an unpredictable monetary supply. What's Bitcoin's primary advantage over fiat? The supply is predictable. Lost coins will erode that advantage. 
I see no need to recover lost coins. If they are lost, then the only meaningful recovery (restoration to their owner) is impossible. There is a real issue that gradual waste of bitcoins over many years will eventually, by laws of mathematics, result in there being almost no bitcoins remaining on wallets where people can recall the keys.
What "law of mathematics" would that be exactly? 
The only solution to that is to change the software to allow new coin creation at a rate that equals or exceeds such loss. However, it is impossible to know what that rate is. At the end of the day, I think the real solution is one provided by the market -- if over some years bitcoins fall into the cracks between the couch cushions until there are too few left for any real use, hard forks and alts will take up the slack in payment processing bandwidth. There is no need to take any action with respect to bitcoin software today, as this will happen whether we like it or not.
No that is not the only solution. Recycling lost coins is also a solution which addresses the problem you mention. The point of recycling is to maintain the intended 21 mill coin supply.
I see no need to recover lost coins. If they are lost, then the only meaningful recovery (restoration to their owner) is impossible. There is a real issue that gradual waste of bitcoins over many years will eventually, by laws of mathematics, result in there being almost no bitcoins remaining on wallets where people can recall the keys. The only solution to that is to change the software to allow new coin creation at a rate that equals or exceeds such loss. However, it is impossible to know what that rate is. At the end of the day, I think the real solution is one provided by the market -- if over some years bitcoins fall into the cracks between the couch cushions until there are too few left for any real use, hard forks and alts will take up the slack in payment processing bandwidth. There is no need to take any action with respect to bitcoin software today, as this will happen whether we like it or not.
Let me make sure I understand the premise of this proposal.
In order to avoid the potential for inflation caused by someone suddenly discovering a large stash of old coins and spending them all at once, you're going to wait for some predetermined amount of time and then suddenly give large stashes of old coins to the miners so they can spend them?
No. In order to avoid the costs associated with the risk of someone suddenly releasing a large stash of old coins, you'd wait a predetermined amount of time and then add those coins to a pool that would be released to miners at a known rate.In order to avoid the potential for inflation caused by someone suddenly discovering a large stash of old coins and spending them all at once, you're going to wait for some predetermined amount of time and then suddenly give large stashes of old coins to the miners so they can spend them?
This would have two benefits:
1) It would avoid "bursty" mining. When the block reward is zero, people might not mine until sufficient transaction fees build up to make it worth it (because idle hardware draws less power). This could make the block generation rate much less predictable.
2) It would avoid costs associated with a less predictable monetary supply. Anything that could be harmed by a change in monetary supply would have to spend money to hedge this risk and some productive activity that otherwise might take place will be discouraged because of those increased costs.
There are certainly good reasons to criticize this idea, but please understand it first. (And, of course, it should definitely *not* be added to Bitcoin now. Upsetting people's expectations by fundamentally changing the rules now is the last thing Bitcoin needs.)
Actually, leaving it alone is the option that changes the rules the least. Putting a lifetime on coins is a huge change, regardless of what happens when they expire.
It seems possible to me that miners might just decide to delete old parts of the block chain. If miners tend not to accept old coins, then other miners might decide not to risk including them.
Worryingly, this is actually a soft fork, since you are not adding a new feature to the network. You are making the network more strict.
But I'm not intending to address the distinction between locking the coins and recycling the coins. I actually think it doesn't matter.
They are distinct, and it does matter. One transfers wealth from the economy at large to miners and one does not. You might want to argue that it's reasonable to make the transfer, but it's easy to point out scenarios where it would be harmful. ... So I just suggest that if you want to talk about the uncertainty issue you do it in terms of the pure form, not one which is also mixed up with a wealth transfer. 
Let me make sure I understand the premise of this proposal.
In order to avoid the potential for inflation caused by someone suddenly discovering a large stash of old coins and spending them all at once, you're going to wait for some predetermined amount of time and then suddenly give large stashes of old coins to the miners so they can spend them?
In order to avoid the potential for inflation caused by someone suddenly discovering a large stash of old coins and spending them all at once, you're going to wait for some predetermined amount of time and then suddenly give large stashes of old coins to the miners so they can spend them?
Nope, read the post I made above. They obviously wouldn't all be given away to miners in one block lol...
This isn't really a big deal, but it doesn't give a predictive income for miners. If the early Satoshi coins are recycled during one year, it will be an enormous boost for miners and then a gradual drop. There is no reason to have such an irratic payback curve. People would invest in rigs like crazy that year and throw the rigs in the garbage the next year.
Fixing the transaction issue can be done by either increasing the transaction costs, or as a general tax, called demurrage. This depends on whether people think it is ok to have "freeloaders" own bitcoins without paying for the maintainance of the network, or not.
IMO it would be predictive income. In 100 years we would know exactly how many coins will be recycled and there are many many different ways the coins can be recycled to address these issues. One method would be to base the number of coins recycled per block as a % of the transaction fees included since that will be the only block reward in 100 years. So for example if it was 10% and there are 10BTC in transaction fees included in a block then an 1 additional BTC would be recycled. That would make the increased block reward very predictable and stable. That's just one of many different ideas...Fixing the transaction issue can be done by either increasing the transaction costs, or as a general tax, called demurrage. This depends on whether people think it is ok to have "freeloaders" own bitcoins without paying for the maintainance of the network, or not.
X should be minimum 75 years. That's 1 life 
I agree, I think 100 years is good since it should be > 1 life 
Let me make sure I understand the premise of this proposal.
In order to avoid the potential for inflation caused by someone suddenly discovering a large stash of old coins and spending them all at once, you're going to wait for some predetermined amount of time and then suddenly give large stashes of old coins to the miners so they can spend them?
In order to avoid the potential for inflation caused by someone suddenly discovering a large stash of old coins and spending them all at once, you're going to wait for some predetermined amount of time and then suddenly give large stashes of old coins to the miners so they can spend them?
Not sure if this is the best place to post this but I was thinking about all of the lost coins this morning and had an idea that may or may not suck.
What if extremely stale coins got recycled? For example, if an address has not be touched in X years then Y BTC from it gets added to the next block reward. Not sure what a good age and ammount/% would be but perhaps 10 years. Would be trivial to move your coins every few years to keep them "fresh" and wallets could automate this and have warnings as well. With a system like this, no coins would ever be permanently lost and would add some reward to miners down the road. OK now tell me why this is a terrible idea...
What if extremely stale coins got recycled? For example, if an address has not be touched in X years then Y BTC from it gets added to the next block reward. Not sure what a good age and ammount/% would be but perhaps 10 years. Would be trivial to move your coins every few years to keep them "fresh" and wallets could automate this and have warnings as well. With a system like this, no coins would ever be permanently lost and would add some reward to miners down the road. OK now tell me why this is a terrible idea...
X should be minimum 75 years. That's 1 life
To me it seems that just declaring them unspendable is the option that changes incentives/rules the least. Giving the coins to miners changes the rules.
Imagine that this situation happened when 99% of the coins were mined and suddenly 10% of coins were going to be given to miners, giving them 10x as much as before this decision.
Doesn't make much sense to me.
Edit: The change that does the least change in incentives is to announce that there will be a window of spendability that reaches X years back at all times, where X > the oldest coin. Then no coin is invalidated now, but a rule that keep uncertainty bounded has been introduced.
Why exactly do you feel it's so wrong to change the incentives in this fashion? If recycling was added now with a 100 year shelf-life on coins, there would be PLENTY of time to account for the added mining incentive coming way down the road. I know some are concerned that transactions fees won't be enough incentive to secure the network down the road and this would add to the incentive without increasing the intended supply of coins. Again, I no longer think this is needed but like to play devil's advocate...Imagine that this situation happened when 99% of the coins were mined and suddenly 10% of coins were going to be given to miners, giving them 10x as much as before this decision.
Doesn't make much sense to me.
Edit: The change that does the least change in incentives is to announce that there will be a window of spendability that reaches X years back at all times, where X > the oldest coin. Then no coin is invalidated now, but a rule that keep uncertainty bounded has been introduced.
This isn't really a big deal, but it doesn't give a predictive income for miners. If the early Satoshi coins are recycled during one year, it will be an enormous boost for miners and then a gradual drop. There is no reason to have such an irratic payback curve. People would invest in rigs like crazy that year and throw the rigs in the garbage the next year.
Fixing the transaction issue can be done by either increasing the transaction costs, or as a general tax, called demurrage. This depends on whether people think it is ok to have "freeloaders" own bitcoins without paying for the maintainance of the network, or not.
To me it seems that just declaring them unspendable is the option that changes incentives/rules the least. Giving the coins to miners changes the rules.
Actually, leaving it alone is the option that changes the rules the least. Putting a lifetime on coins is a huge change, regardless of what happens when they expire.
Jump to: