Gavin Andresen
Legendary
Offline
Activity: 1652
Merit: 2300
Chief Scientist
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 12:33:08 AM |
|
There will be 2.1 quadrillion satoshis. There are approximately 770 trillion pennies in the US M2 money supply. We've got a VERY VERY long way to go before bitcoins are as popular as dollars, and there are much higher priority things to work on right now than adding more divisibility for a problem that is pretty likely to never actually be a problem.
|
How often do you get the chance to work on a potentially world-changing project?
|
|
|
dree12
Legendary
Offline
Activity: 1246
Merit: 1077
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 12:47:43 AM |
|
There will be 2.1 quadrillion satoshis. There are approximately 770 trillion pennies in the US M2 money supply. We've got a VERY VERY long way to go before bitcoins are as popular as dollars, and there are much higher priority things to work on right now than adding more divisibility for a problem that is pretty likely to never actually be a problem. The US is not the world. In fact, it is only 15% of the world, and the number of bitcoins isn't enough for the entire world. Think about the ease of pushing a +4 decimal point update today (I believe Bitcoin uses int64 anyways), compared to pushing it when it is needed, to billions of people. Better now than never.
|
|
|
|
Blazr
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 01:04:44 AM |
|
The US is not the world. In fact, it is only 15% of the world, and the number of bitcoins isn't enough for the entire world.
Think about the ease of pushing a +4 decimal point update today (I believe Bitcoin uses int64 anyways), compared to pushing it when it is needed, to billions of people.
Better now than never.
I don't think Bitcoin will ever be used by billions of people, perhaps in 10-20 years a currency similar to Bitcoin will be used by that many people. We don't need 4 more decimals and probably never will, IMO 8 is too many now as the value of 1 Satoshi is ridiculously small. If we do need to add more its a very small change and shouldn't be much harder even if Bitcoin is being used by billions.
|
|
|
|
dree12
Legendary
Offline
Activity: 1246
Merit: 1077
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 01:13:20 AM |
|
The US is not the world. In fact, it is only 15% of the world, and the number of bitcoins isn't enough for the entire world.
Think about the ease of pushing a +4 decimal point update today (I believe Bitcoin uses int64 anyways), compared to pushing it when it is needed, to billions of people.
Better now than never.
I don't think Bitcoin will ever be used by billions of people, perhaps in a few years a currency similar to Bitcoin will be used by that many people. We don't need 4 more decimals and porbably never will, IMO 8 is too many now as the value of 1 Satoshi is ridiculously small. If we do need to add more its a very small change and shouldn't be much harder even if Bitcoin is being used by billions. Right, sell off your bitcoins now. In a few years, they'll be worthless! Bitcoin was not designed to be used by billions of people; I give you that. But why do we work on scalability, lite nodes, etc., if all we're hoping for is several million users? Even one billion people would put a lot of stress on Bitcoin's decimal system. Assuming ~5% of the coins will be lost (for simplicity), on average, there will be 0.02 BTC per person, or 2000000 satoshi. Some will be richer than others, and we can assume that the lower middle class (the largest demographic) might have a million (1000000) satoshi. This is a ridiculously small amount when every transaction needs a transaction fee of at least one satoshi. Realistically, in this situation, banks will offer a) sub-satoshi amounts, and b) fractional reserve. These advantages make banks extremely attractive, contrary to at least a portion (the decentralization part) of Bitcoin's philosophy.
|
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 01:25:22 AM Last edit: May 31, 2012, 02:27:41 AM by DeathAndTaxes |
|
Even one billion people would put a lot of stress on Bitcoin's decimal system. Assuming ~5% of the coins will be lost (for simplicity), on average, there will be 0.02 BTC per person, or 2000000 satoshi. Some will be richer than others, and we can assume that the lower middle class (the largest demographic) might have a million (1000000) satoshi. This is a ridiculously small amount when every transaction needs a transaction fee of at least one satoshi. Why is that ridiculously small? Even if your example, the tx fee is 1 millionth of the net worth of the median user. Lets use US median net worth (2004 is latest figure I could grab). That is ~$94K per household. Even if we assume the per capita net wealth is half that we are talking about $50K. One one millionth of that is ~$0.05. Oh noes 5 cent tx fees. Surely the ability to send unlimited amounts of money to anywhere on the planet within a matter of seconds isn't worth a "whole 5 cents". I mean just look at ACH fees (~$0.35), credit cards ($0.30 + 3%), ATM fees ($2 to $5+), wire transfers ($10 to $50), western union ($10 to $100). However will Bitcoin compete? DOOM I TELL YOU. The sub 5 cent tx fee will be the end of Bitcoin. FUD aside: when 1 satoshi becomes so large as to be ineffective as a discrete unit (say >$0.10 USD in 2012 dollars) it will not be that difficulty to add 4 decimal places dropping the value to <0.001 cent (USD).
|
|
|
|
dree12
Legendary
Offline
Activity: 1246
Merit: 1077
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 02:18:56 AM |
|
Even one billion people would put a lot of stress on Bitcoin's decimal system. Assuming ~5% of the coins will be lost (for simplicity), on average, there will be 0.02 BTC per person, or 2000000 satoshi. Some will be richer than others, and we can assume that the lower middle class (the largest demographic) might have a million (1000000) satoshi. This is a ridiculously small amount when every transaction needs a transaction fee of at least one satoshi. Why is that ridiculously small? The tx fee is 1 millionth of the avg net worth of the middle class user. Lets use US median net worth (2004 is latest figure I could grab). That is ~$94K per household. Even if we assume the per capita net wealth is half that we are talking about $50K. One one millionth of that is ~$0.05. Oh noes 5 cent tx fees. Surely the ability to send unlimited amounts of money to anywhere on the planet within a matter of seconds isn't worth a "whole 5 cents". I mean just look at ACH fees (~$0.35), credit cards ($0.30 + 3%), ATM fees ($2 to $5+), wire transfers ($10 to $50), western union ($10 to $100). However will Bitcoin compete. DOOM I TELL YOU. The sub 5 cent tx fee will be the end of Bitcoin. When 1 satoshi becomes so large as to be ineffective as a discrete unit (say >$0.10 USD in 2012 dollars) it will not be that difficulty to add 4 decimal places dropping the value to <0.001 cent (USD). And I ask: is it easier to push such a change to the 200K bitcoin users today, or to the 5B bitcoin users tommorow?
|
|
|
|
Blazr
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 02:54:25 AM |
|
Even one billion people would put a lot of stress on Bitcoin's decimal system.
1 billion, Lets put that into perspective. There are ~330 million people who use Euro.
|
|
|
|
Blazr
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 03:26:08 AM |
|
And I ask: is it easier to push such a change to the 200K bitcoin users today, or to the 5B bitcoin users tommorow?
Roughly the same. The extra users are only a minor inconvenience. Simply put those who don't update won't be able to spend sub-satoshi amounts, so they'd pretty much have to get the update which would be a 1 line of code change that'd most likely be released months in advance and then auto-activate once we reach a certain block (like how namecoin started merged mining). Also seeing as the world population is slightly under 7 billion, it'd be unlikely any one currency will ever be used by 5 billion people.
|
|
|
|
dree12
Legendary
Offline
Activity: 1246
Merit: 1077
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 03:48:46 AM |
|
And I ask: is it easier to push such a change to the 200K bitcoin users today, or to the 5B bitcoin users tommorow?
Roughly the same. The extra users are only a minor inconvenience. Simply put those who don't update won't be able to spend sub-satoshi amounts, so they'd pretty much have to get the update which would be a 1 line of code change that'd most likely be released months in advance and then auto-activate once we reach a certain block (like how namecoin started merged mining). Also seeing as the world population is slightly under 7 billion, it'd be unlikely any one currency will ever be used by 5 billion people. Why delay though? And I'm not sure where you get slightly under 7 billion from. This US gov't site says otherwise: http://www.census.gov/main/www/popclock.html.
|
|
|
|
Blazr
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 03:59:05 AM |
|
Because its unnecessary to do it now, we already have too many decimals and plenty of other more important things to fix instead. I got that figure here: https://www.google.ie/search?q=world+populationObviously there are many different methods of estimating the population of the world, each giving a slightly different answer.
|
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 03:59:26 AM |
|
Why delay though?
Because there are like a couple million more pressing issues, it serves no point, and it increases the size of every transaction between now and the day when 1 satoshi > 1 penny. I mean say you got your wish and 8 digits were added tomorrow. Then what? Wait? What if Bitcoin is used by some Galactic Imperium? With countless trillions of users. In a milenium or two oh noes we might be out of digits. So to be safe I guess we should add 256 digits now and create exabytes worth of blockchain spam between now and then just to be "safe". Just to put it perspective how asinine the argument "why delay" is. The global money supply (all countries, all people, all forms both physical and electronic) is on the order of $70 trillion. Now if we consider the penny to the discrete unit that means the world has ~7 quadrillion discrete units. Bitcoin has 21 quadrillion.
|
|
|
|
dree12
Legendary
Offline
Activity: 1246
Merit: 1077
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 07:42:28 PM |
|
Why delay though?
Because there are like a couple million more pressing issues, it serves no point, and it increases the size of every transaction between now and the day when 1 satoshi > 1 penny. I mean say you got your wish and 8 digits were added tomorrow. Then what? Wait? What if Bitcoin is used by some Galactic Imperium? With countless trillions of users. In a milenium or two oh noes we might be out of digits. So to be safe I guess we should add 256 digits now and create exabytes worth of blockchain spam between now and then just to be "safe". Just to put it perspective how asinine the argument "why delay" is. The global money supply (all countries, all people, all forms both physical and electronic) is on the order of $70 trillion. Now if we consider the penny to the discrete unit that means the world has ~7 quadrillion discrete units. Bitcoin has 21 quadrillion.Emphasis added. I'll give it to you; we all make mistakes, and it may as well have just been a typo. But rather than argue about this pointlessly, there's another argument that relates to the entire point of this thread: lost bitcoins. The bitcoin transaction protocol is not non-extensible; it will require a chain fork, but it is in fact very extensible. It is possible to avoid blockchain spam by using a different transaction parameter to turn on "extended bitcoin precision", while keeping all applications compatible even today. Said chain fork is easier done now than when Bitcoin is used bysome Galactic Imperium.
|
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 07:49:31 PM |
|
There are many things needed TODAY. * Implementing nTimeLock transaction * Implementing Sequence * Implementing mult-sig redemption (via multiple wallets) * Implementing blockchain pruning * Implementing standardized export/import formats * Developing double spend prevention systems * Developing mechanisms to better protect hotwallets * Developing 0-confirm guarantees (for truly instant purchases) * Expanding support for bitcoin libraries (bitcoinj is a nice start but lets see something equally comprehensive for php, python, c#, C++, C, etc) * Expanding the number of clients (centralization of developer is a risk to a p2p network) * Developing dedicated backend "clients" (think service/daemon connected to enterprise DBMS or did you think Amazon was going to use bitcoind) * Developing user friendly clients (I mean really user friendly. Like my Grandmother can split a check w/ friends using Bitcoins friendly). * Improvement in blockchain download / new node bootstrap process/speed. * Developing 51% attack detection/mitigation/recovery systems. * (about 10,000 other things)
"Worrying" (trolling) about not enough discrete units if Bitcoin becomes millions of times more popular in some future decade is just pointless.
|
|
|
|
dree12
Legendary
Offline
Activity: 1246
Merit: 1077
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
May 31, 2012, 08:24:51 PM |
|
There are many things needed TODAY. * Implementing nTimeLock transaction * Implementing Sequence * Implementing mult-sig redemption (via multiple wallets) * Implementing blockchain pruning * Implementing standardized export/import formats * Developing double spend prevention systems * Developing mechanisms to better protect hotwallets * Developing 0-confirm guarantees (for truly instant purchases) * Expanding support for bitcoin libraries (bitcoinj is a nice start but lets see something equally comprehensive for php, python, c#, C++, C, etc) * Expanding the number of clients (centralization of developer is a risk to a p2p network) * Developing dedicated backend "clients" (think service/daemon connected to enterprise DBMS or did you think Amazon was going to use bitcoind) * Developing user friendly clients (I mean really user friendly. Like my Grandmother can split a check w/ friends using Bitcoins friendly). * Improvement in blockchain download / new node bootstrap process/speed. * Developing 51% attack detection/mitigation/recovery systems. * (about 10,000 other things)
"Worrying" (trolling) about not enough discrete units if Bitcoin becomes millions of times more popular in some future decade is just pointless.
Excellent, then I assume the next chain fork will include more decimal points. Or is that going to be put off the next (inevitably failing) chain fork?
|
|
|
|
ripper234
Legendary
Offline
Activity: 1358
Merit: 1003
Ron Gross
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 01, 2012, 12:47:40 AM |
|
There are many things needed TODAY. * Implementing nTimeLock transaction * Implementing Sequence * Implementing mult-sig redemption (via multiple wallets) * Implementing blockchain pruning * Implementing standardized export/import formats * Developing double spend prevention systems * Developing mechanisms to better protect hotwallets * Developing 0-confirm guarantees (for truly instant purchases) * Expanding support for bitcoin libraries (bitcoinj is a nice start but lets see something equally comprehensive for php, python, c#, C++, C, etc) * Expanding the number of clients (centralization of developer is a risk to a p2p network) * Developing dedicated backend "clients" (think service/daemon connected to enterprise DBMS or did you think Amazon was going to use bitcoind) * Developing user friendly clients (I mean really user friendly. Like my Grandmother can split a check w/ friends using Bitcoins friendly). * Improvement in blockchain download / new node bootstrap process/speed. * Developing 51% attack detection/mitigation/recovery systems. * (about 10,000 other things)
"Worrying" (trolling) about not enough discrete units if Bitcoin becomes millions of times more popular in some future decade is just pointless.
Excellent, then I assume the next chain fork will include more decimal points. Or is that going to be put off the next (inevitably failing) chain fork? Not chain fork, but rather "potential chain fork". BIP16/17 was a potential chain fork that wasn't a real fork, because it was well managed. The process will get easier, not harder, as we practice more and more potential forks along the years, for reasons far more important than decimal points. All those 1B of users that you mentioned won't be running full clients anyway. Even today there's no reason to run a full client today if you just want to pass some BTC around. Give up the thread ... please, "including more decimal points" won't happen in the near-medium future.
|
|
|
|
|