Should there be Three Laws of Bitcoin?

[bitcool]

With recent events of theft, hacking, combined with the never-ending deflation/inflation debate,  it's very tempting to modify the code solving these issues.

For example:

1. If a large sum of coins is stolen or lost due to malicious attack, the client code can be modified to blacklist these coins (and their subsequent forms), or it can even go further issuing specially generated coins to compensate the victims.

2. Why there can't be a time limit of inactivity, if a coin hasn't been touched in the past 10 or 20 years, can the program just mark it as "lost" in the blockchain and issue a replacement?

3. If a majority of bitcoin users have reached a consensus that the deflationary aspect is bad to the currency, development team will likely be under pressure to modify the coin generation algorithm, e.g. lifting the 21M cap.

I can see a day development team will split on these issues, there will be struggle between "reformist" and "fundamentalist"; how far bitcoin is allowed to evolve and still can be called "bitcoin"? If bitcoin survives in the next few years, this will be the most serious issue determining its long term survival.

I am surprised Satoshi didn't attempt to create something like "Ten Commandments of Bitcoin" or "Three Laws of Robotics" (http://en.wikipedia.org/wiki/Three_Laws_of_Robotics)

Or maybe it does exist somewhere and I am just ignorant?

[1714005579]

1714005579

[1714005579]

1714005579

[1714005579]

1714005579

[1714005579]

1714005579

[shads]

The first rule of Bitcoin is: You don't change the protocol
The second rule of Bitcoin is: You don't change the protocol
The third rule is: If you are new to Bitcoin, you have to think of ways the protocol should be changed...

Which is why I suspect there will probably be a bitcoin2 one day with a migration path to transfer bitcoin1's into it.  Bitcoin is really in beta still.  It's a very neat design with a number of checks and balances built into it but it's really only starting to be tested properly in the last couple of months.  Both the code and rules of the system need to withstand a lot of external factors that it won't face until the economy gets bigger and many which may never be predicted before they happen.  This will inevitably highlight weaknesses in the design that can be rectified by tweaking it.  You can't realistically keep tweaking the current bitcoin, it will undermine confidence if the rules keep changing.  So the best solution is to eventually come up with a bitcoin2 standard and migrate.  Leave the existing bitcoin in place and people can migrate as they see fit.  Market forces will determine whether btc 1 or 2 will prevail and then the loser will die out completely.  As it becomes clear that bitcoin2 is going to become the standard bitcoin1 value will start to drop and motivate people to migrate.

[TiagoTiago]

1 - Can be easily abused to cripple people that got paid honestly

2 - What happens if someone digs up their private keys 20 years from now and tries to spend their savings? I don't think the system should be coded to automaticly steal from people.

3 - It's open source, anyone is free to release clients that will allow for a bigger total number; if people using that variation outnumber people using clients that respect the original design the system will automaticly consider the new rules as the official rules.

[bji]

2. Why there can't be a time limit of inactivity, if a coin hasn't been touched in the past 10 or 20 years, can the program just mark it as "lost" in the blockchain and issue a replacement?

Others may disagree but I think this is a fantastic idea.  Of course it's too late now for bitcoin to include such a feature but I see it working with a simple rule like this:

For transactions over 20 years old, any input referencing that transaction is considered to have successfully claimed the value of the output no matter what the script returns.

In this way, once a transaction is 20 years old, it's anyone's to claim.

If you're nearing the 20 year mark (say at year 15 if you're really paranoid or something), you just re-send the money to a different bitcoin address and you're covered for another 20 years.

No bitcoin would ever be lost; someone would always claim it.  That someone being a miner since they'd obviously put a transaction taking the bitcoin into the next block they are trying to solve as soon as it is valid to do so.

I think it's a great idea and it leads to 0% inflation and 0% deflation in a system like bitcoin since no money is ever lost.

[Mark Oates]

2. Why there can't be a time limit of inactivity, if a coin hasn't been touched in the past 10 or 20 years, can the program just mark it as "lost" in the blockchain and issue a replacement?

Others may disagree but I think this is a fantastic idea.

It's a terrible idea.  Attempting to introduce this into the current protocol (or a newer one) isn't a practical idea.  Mostly because you're attempting to solve a "problem" that shouldn't be solved at the protocol level.  Think layers of an onion.  In this case, we have a solid core.  This is a type of "problem" that can be solved by increasing the precision of the bitcoin divisions.

The reason I put the word problem in quotes is because, as far as I understand, the number of units of coins are theoretically scalable to an indefinite number.  The fear is that we'll run out of coins.  All we need is one coin.

What happens if someone digs up their private keys 20 years from now and tries to spend their savings? I don't think the system should be coded to automaticly steal from people.

To me, this really nails it on the head.  Any of the time-out proposals would introduce a footnote into the system.  Your savings will be GONE if you forget this little long-term clause.  Exceptions like this don't belong in the core protocol.

[ascent]

This is a type of "problem" that can be solved by increasing the precision of the bitcoin divisions.

I don't believe this is correct. The problem isn't solved by increasing bitcoin divisions. That isn't what the problem is about. The problem is that the ratio of saved coins to lost coins becomes less knowable over time, while at the same time, the ratio of coins known to be in circulation to the coins that are currently not in circulation (lost or saved) becomes smaller and smaller, which leads to a system that becomes, at the extreme, so uncertain that it is unusable.

See the thread on wallet heartbeats.

[FreeMoney]

I think OP should have to move his coins every 20 minutes or I get them.

[TiagoTiago]

Btw, what if someone gets rich, and then lives the rest of their lifes spending a small fraction of their fortunes per month; i would be seriously pissed if after 20 years suddenly my fortune got claimed back to the ether leaving me pennyless...

[ascent]

Btw, what if someone gets rich, and then lives the rest of their lifes spending a small fraction of their fortunes per month; i would be seriously pissed if after 20 years suddenly my fortune got claimed back to the ether leaving me pennyless...

And why would that happen? If you are capable of being pissed about the loss, then you're aware of your holdings, and thus clearly in a position to lift your finger every 19 years or so.

[bji]

It's a terrible idea.  Attempting to introduce this into the current protocol (or a newer one) isn't a practical idea.

It's as practical as any other idea; perhaps even more so, because we'd have 20 years for all bitcoin peers to update themselves to this agreed protocol change before it would even take effect, plenty of time for conformance testing, planning, etc.

Mostly because you're attempting to solve a "problem" that shouldn't be solved at the protocol level.  Think layers of an onion.  In this case, we have a solid core.  This is a type of "problem" that can be solved by increasing the precision of the bitcoin divisions.

Yeah now that I think about it more, perhaps there is no value in guaranteeing that no bitcoins are ever lost.  Because practically speaking, there is no difference between a bitcoin that cannot be spent because it's lost, and a bitcoin that will not be spent because the owner wants to hoard it.  Still I wonder if perfect knowledge of exactly how many bitcoins are circulat-able is valuable in any way; if it is, then my proposal has merit.

What happens if someone digs up their private keys 20 years from now and tries to spend their savings? I don't think the system should be coded to automaticly steal from people.

They already lost their money if they waited 20 years.  I mean come on, it's not that hard to 'refresh' your bitcoins at least once every 20 years.  If you can't do it for some reason, then you've lost your money, sorry.  I would bet that orders of magnitude more people are going to lose their money because they lose their private key somewhere along the way than would ever lose their money due to this 20 year rule, and this kind of loss is already built into the bitcoin system.  But you're willing to accept that I guess because, even though it's many orders of magnitude worse than the 20 year rule, it's already built into the bitcoin system so I guess it's not a problem to you?

To me, this really nails it on the head.  Any of the time-out proposals would introduce a footnote into the system.  Your savings will be GONE if you forget this little long-term clause.  Exceptions like this don't belong in the core protocol.

If it were a built-in rule then it would be very clear to everyone and I doubt that anyone would be caught out in the cold by it.  Seriously.  Someone else pointed out that many banks have a policy already that if you don't use your account in 7 years, the bank reclaims the money.  Or something like that.  People can handle rules like this.

It's not that I'm saying it's super important to get it into the bitcoin protocol; I just think it's an interesting idea and there's something satisfying about a system that can guarantee 0% inflation and 0% deflation.

[ascent]

Yeah now that I think about it more, perhaps there is no value in guaranteeing that no bitcoins are ever lost.  Because practically speaking, there is no difference between a bitcoin that cannot be spent because it's lost, and a bitcoin that will not be spent because the owner wants to hoard it.  Still I wonder if perfect knowledge of exactly how many bitcoins are circulat-able is valuable in any way; if it is, then my proposal has merit.

See this post: http://forum.bitcoin.org/index.php?topic=20866.msg262100#msg262100

There is a difference, and it becomes more pronounced over time.

[bji]

It's not that I'm saying it's super important to get it into the bitcoin protocol; I just think it's an interesting idea and there's something satisfying about a system that can guarantee 0% inflation and 0% deflation.

Just to be more precise here: while there would be 0% inflation and deflation over long time spans, this doesn't mean that there wouldn't still be uncertainty about the bitcoin supply.  Some coins would still be lost and although they would be claimed no more than 20 years later, there would always be some unspendable coins out there, and this amount would fluctuate.

[MoonShadow]

With recent events of theft, hacking, combined with the never-ending deflation/inflation debate,  it's very tempting to modify the code solving these issues. For example

1. If a large sum of coins is stolen or lost due to malicious attack, the client code can be modified to blacklist these coins (and their subsequent forms), or it can even go further issuing specially generated coins to compensate the victims.

No.  This would introduce an obvious abuse vector.  He who could influence the blacklisting of coins, or their unlisting, to his own advantage would eventually do so.  Don't bother trying to claim that you are above corruption.  No one is.  Everyone who runs for congress for the first time does so because he believes this fallacy.

2. Why there can't be a time limit of inactivity, if a coin hasn't been touched in the past 10 or 20 years, can the program just mark it as "lost" in the blockchain and issue a replacement?

Again, abuse vector.  How is the time limit decided?  Can this mechanism be spoofed?  I suppose one could use the block number, but why bother?  The future expected deflation due to lost coins is bound to be so slow as to be unimportant.

3. If a majority of bitcoin users have reached a consensus that the deflationary aspect is bad to the currency, development team will likely be under pressure to modify the coin generation algorithm, e.g. lifting the 21M cap.

Absolutely not.  Get out.  And go start your own currency.

[Desu]

Not a smart one.
Say an average joe gets talked into investing into BTC. Encrypts it on a drive and puts it in a lock box, Saving it for 20 years to retire, buy a new car or house, or even their childs college/car.(Considering they believe the price of BTC's will raise to a certain amount at this point in there life.) They go back after all this time pop it in a computer(Of the future, ooooo ahhhh! lol) And boom, Bitcoin says there data is to old. They depended on that as a back up plan and now it's absolutely worthless. Whose to take from all of us today, and throw it out in 20 years?
JS.

[ascent]

Do you know what happens to bank accounts after seven years of inactivity?

[kjj]

Do you know what happens to bank accounts after seven years of inactivity?

Free toaster?

[bji]

Not a smart one.
Say an average joe gets talked into investing into BTC. Encrypts it on a drive and puts it in a lock box, Saving it for 20 years to retire, buy a new car or house, or even their childs college/car.(Considering they believe the price of BTC's will raise to a certain amount at this point in there life.) They go back after all this time pop it in a computer(Of the future, ooooo ahhhh! lol) And boom, Bitcoin says there data is to old. They depended on that as a back up plan and now it's absolutely worthless. Whose to take from all of us today, and throw it out in 20 years?
JS.

The 20 year rule would be common knowledge.  With commonly disseminated knowledge about the rule and an easy way to avoid the 20 year rule, I don't see any problem with expecting people to use the system properly.

Do you also think it's a problem that if you delete your private key file, you lose all of your bitcoins?  What about the average joe that saves for 20 years and then accidentally deletes or loses their private key?  Does the fact that this can happen mean that bitcoin itself should not exist, just like your assertion that the fact that a 20 year rule should not exist for the same reason (i.e. someone know knows, or should know the rules, still makes a mistake and loses their money)?

[bji]

Do you know what happens to bank accounts after seven years of inactivity?

Yes. And, of course, we want Bitcoin to emulate the modern banking system when it grows up to be 'real money'.

You missed the point completely.  Was that on purpose?

The point was not to say that bitcoin should be like a banking system.  It was to show that the kinds of problems that people are proposing with this system ("but people who ignore common knowledge can get screwed!") are already accepted by people in entrenched systems.  If such rules are not already too onerous for people to deal with, why would they be too onerous when implemented in bitcoin?

[Mark Oates]

Do you also think it's a problem that if you delete your private key file, you lose all of your bitcoins?

No, it doesn't belong in the bare-bones protocol.

At the kernel level of an OS there is no "undelete."

[billyjoeallen]

With recent events of theft, hacking, combined with the never-ending deflation/inflation debate,  it's very tempting to modify the code solving these issues. For example

1. If a large sum of coins is stolen or lost due to malicious attack, the client code can be modified to blacklist these coins (and their subsequent forms), or it can even go further issuing specially generated coins to compensate the victims.

No.  This would introduce an obvious abuse vector.  He who could influence the blacklisting of coins, or their unlisting, to his own advantage would eventually do so.  Don't bother trying to claim that you are above corruption.  No one is.  Everyone who runs for congress for the first time does so because he believes this fallacy.

2. Why there can't be a time limit of inactivity, if a coin hasn't been touched in the past 10 or 20 years, can the program just mark it as "lost" in the blockchain and issue a replacement?

Again, abuse vector.  How is the time limit decided?  Can this mechanism be spoofed?  I suppose one could use the block number, but why bother?  The future expected deflation due to lost coins is bound to be so slow as to be unimportant.

3. If a majority of bitcoin users have reached a consensus that the deflationary aspect is bad to the currency, development team will likely be under pressure to modify the coin generation algorithm, e.g. lifting the 21M cap.

Absolutely not.  Get out.  And go start your own currency.

+1