Topic: If you could restart the block chain... (Read 2470 times)

Yes I've got a prize here waiting for you. Can you guess what it is?

Hint: Answer starts with 'N'.

OK, I understand now where you're coming from, but I don't think it's the most applicable per figure to use.  Bitcoins per capita (measured in bitcoins per person who uses Bitcoin) has gone down by a lot as Bitcoin has become more popular.  By analogy to national fiat currency, let's say there's a country with 100 million people whose population is growing at a rate of 5% per year.  If they're only printing 2% of the country's total monetary supply per year, then while the total amount of outstanding currency is going up, it's actually going down per capita, which will cause real deflation (regardless of how you want to define the term).  Bitcoin is in a similar situation, except the difference between the rate of new currency coming online and the rate of people joining the economy is much, much more uneven.

1. is too easy to game.  Someone would just keep generating wallets and getting lots of coins.

2. not bad

3. the problem with this is it is hard to predict what kind of hashing power will exist in the future.  Perhaps it becomes trivial to get 1 TH/s in the future.   So inflation becomes unpredictable.


I would change it only so that the reward is based on a bell curve with the peak 20 years out.  So initially the reward is small.  Then as adoption increases, it increases.  Then it tapers off after adoption has hit critical mass.

The answer is: NOTHING!, Do i win anything? like a cookie? or btc's to buy a cookie?

This could be implemented immediately without the need to restart the block chain.  It's actually a pretty good idea imho, other than the fact that it's a waste of electricity if done primarily with CPUs.

I don't really see a problem with the early adopters making out handsomely (I'm not one of them--I started mining in January).  I actually stumbled upon Bitcoin in its very early days and wrote it off as too small and going nowhere and decided to use Pecunix instead.  My mistake, but I don't begrudge the people who had faith in it then.  I'm just thrilled that we finally have a viable peer-to-peer cryptocurrency all these years after the idea was floated by Tim May and others on the cypherpunks mailing list.  This is the point we should focus on.

Pooled mining from the client. This would give the immediate impression that you get paid to contribute to the network, no matter how little you get.

If I said price deflation or price inflation, maybe.

But I was talking about actual inflation. Tomorrow there will be more Bitcoins in the economy than today, they are inflating. When you put more air in a balloon, the balloon is inflating.

Actually, bitcoins are presently gaining in value.  That means they are deflating, not inflating.

I would prefer Bitcoins be worth something. Every change you've suggested makes Bitcoins far, far easier to get and I imagine this would have a significant impact on the value. Also, I'm not a big fan of inflation (yes, I realize Bitcoin is currently inflating), so I wouldn't even bother with block chains with (long term) inflationary rules.

Well, you're right.  Bitcoin is not great for microtransactions.  On the other hand, it shouldn't be.  There's one blockchain which, presently, every single client has to download the entirety of.  It's not in anyone's best interest for the blockchain to be spammed up with thousands of microtransactions an hour originating from MMORPG players trading items in-game.  That kind of stuff is what the priority rules were designed for, and why the fees are as "high" as they are.

Fortunately, there's a really easy solution.

Rather than spending bitcoins in-game directly, you deposit bitcoins into your MMORPG account.  So you might open up World of Warcraft, hit a "Deposit Bitcoins" button, and it gives you a one-time receiving address to use.  Then you fire up your Bitcoin client, send over, say, 5 BTC, and WoW credits your account with that much.  Then you go on trading and whatever as normal within the game, with each transaction being recorded into a database somewhere on Blizzard's servers, and say a couple months later you've accumulated 100 BTC that you want to withdraw.  You click "Withdraw", enter your Bitcoin payment address, and boom.

Yes, it requires some trust in Blizzard's systems, but if we're only talking about micropayments, who cares?  We're not talking about moving thousands of savings in BTC into the system, just a little bit of spending money.

The reason people get BTC for mining is because mining is what secures the blockchain.  Assigning rewards for computational power that secures the network is a really elegant solution that works and that cannot be faked (if you find a valid hash, that means you did do the work).  On the other hand, as you point out, it'd be easy to fake how many clients you have open or what IP addresses you can connect to the network on.  Heck, there are sysadmins at large corporations with entire spare B blocks to throw around (that's 254*254 IP addresses I believe).  They could make a few routing rules, and bam, now they look like they're 64,516 people.  And that's with one B block.  And just forget IPv6.  Absolutely forget it.

Besides, why would we reward having clients open?  Having clients open doesn't really help the network all that much.  Yeah, they relay transactions, but that's not exactly a huge burden.  People have spent thousands of dollars on mining rigs because of the rewards for mining, and all of that money has directly translated into the security of the network.  Would people have to go out of their way at all just to run a client?  Nope.  And does it secure the network?  Not really.

Well that wouldn't work (for the obvious reason that you point out).  This is actually a really tricky issue, and the option Bitcoin chose was the best out of bad alternatives.  Bitcoin needs to be handed out on a basis that is impossible to fake; you can't fake the hashes, but you can equally make as many new receiving addresses as you'd like.  Plus there is a rational reason why we assign BTC to miners, because their computational power is providing security for the block chain.  People generating lots of random addresses wouldn't be providing anyone anything.  If we restarted the blockchain today, at least, there'd be much more interest in it from the very beginning, so the wealth disparity would be a lot less.  I think that's the best we could do.


There's no reason in principle this couldn't be done.  It would certainly counter the claims that Bitcoin has a built-in deflationary spiral.


So basically you would eliminate the difficulty increases.  Thus you would get insane long term inflation roughly on the scale of the size of the growth of the network (around 5% per day right now) times the growth in hashing power due to better hardware (if we use Moore's Law, say it doubles every two years).  Thus, no one would ever use Bitcoin, because it would be losing value due to inflation at obscene levels; really, you could only do worse by using the Zimbabwean dollar.

Plus, how would you assign bitcoins on a basis proportional to hashing power?  Can you come up with an algorithm that would actually manage to do this in a distributed manner?  Keep in mind you'd need to keep the number of blocks generated relatively steady at one per ten minutes on average (the entire system falls apart if you're generating a block, say, every second; most of your time mining would be wasted on blocks that were already invalid, you just didn't know it yet).  So the only way to do it would be to keep the number of blocks being generated steady, but scale the BTC reward per block appropriately.  I guess you could build into the 2000 block window readjustment algorithm that the block generation reward would be multiplied by the same factor as the difficulty level.  So, for instance, this last increase in difficulty from 244K to 434K would've also increased the block generation reward for each block by 78%, up to 89 BTC per block.  I don't like this though.

I am finding that bitcoin seems to be turning out to be not so great for micro-payments.

At first blush the idea of having eight decimals sounds great because it brings to mind conversion rates such as those of World of Warcraft gold (less than a dollar for several million WoW gold), and sounds as if it could be useful if players of http://galaxies.mygamesonline.org/ thought a mere 100,000 or less units (intended to represent on the order of a ton or more per unit probably) of metal or crystal or deuterium should similarly sell for less than a dollar.

For individual character scale items that probably should cost less than a ton of crystals, being able to send someone just one of that very last decimal-place sounds useful.

World of Warcraft gold is awkward not only because of rules they might have against trading it but also because it is hard to hold as a speculator who has no interest in playing the game nor in paying a monthly player-fee to maintain a character in the game to hold the gold.

Thus I figured something like bitcoin would be great. Various games could be built with the "satoshi" unit of currency in mind - that last decimal place of the eight decimal places.

In Crossfire RPG there is already so much implicit inflation that they don't even bother with coppers they use silver as the lowest denomination coin, ten of those to a gold, five of those to a platinum, 100 platinum to a jade and 100 jade to an amberium. So an amberium is 500,000 silver. It seems quite likely that an amberium might well sell for less than a dollar, let alone less than a bitcoin.

The new default fee is what, 50,000 satoshis? How useful is one satoshi if you have to pay 50,000 satoshis to send it to someone?

So I think I need a design that will specifically favour using it as dirt cheap, even throwaway, money.

Deliberately encourage it to be considered worthless by anyone other than gamers and speculators-speculating-that-gamers-might-be-mad. (Mad as in holy moly who the heck would fork out real money for a virtual "magic sword" for gosh sakes?!?!)

Some speculators might also speculate that it could be a sleeper, sneaking up on the world by appearing to be maybe worth even less than world of warcraft gold until some day when all 21 million are in circulation and the fact that this distinguishes it from the normal run of the mill MUDflation-currencies (that are created out of nothing by falling from the pockets of goblins that, it turns out, are themselves created from nothing, possibly not even needing to be bred in vats using limited resources by wizards who might, being wizards, maybe themselves have been created from nothing in the beginning when the universe was created and might even, being wizards, be able to eventually even defeat death) really starts to "hit home".

These coins would not only provide an item that players, speculators, market-makers etc need not even be aware of the game (or of any specific game) in order to use, but that also is a glaring exception to the MUDflationary creation of goods resources items currency and so on in so many games. Thus from an in-game perspective it seems reasonable they should be so valuable that one might even be able to buy tons of mass of materials with them. But depending on how addicted one might or might not be to any specific game they might nonetheless seem worthless to anyone who happens not to know of some market where they could sell them for something they do consider quite valuable, maybe for example some coin or other that is of notable use in some game that they *are* addicted to. (Possibly even with the term game here including "the game known by some as 'real life'"...)

I have all along thought that it would probably turn out to be better to have a separate currency intended and possibly even designed for use in/with games rather than to try to impose upon bitcoins requirements that would make them more useful in games (such as being able to readily and standardly use that last decimal for small purchases without outrageous-by-proportion transaction fees even maybe?)

By starting out with a multiple-galaxies setting as one of the first settings considered, I do have the option of claiming that obviously propagating your transactions across many galaxies probably will cost at least a much as a beer and quite likely more than a beer. So even within indvidual games it might make sense to have multiple blockchains, one (or more: not only the original Hacker-nation "bitcoins" but also Martian Botcoin, United Kingdom Britcoin and maybe even CZech Bitcoin) propagating intergalactically but at least one "local to one solar system or a small group of relatively close to one-another solar systems or even just to one world" blockchain useable for buying beer and pretzels without incurring intergalactic transaction-propagation fees.

There is also a feedback loop to consider, because Earth is considered mythical by most places in those galaxies so instead of thinking Earth currencies are vastly more valuable than galactic currencies the tendency is to assume any talk of Earth is probably by roleplaying gamers playing games involving the mythical/fictional planet earth and that earth currencies are actually roleplaying game currencies.

Characters in the game are maybe about as likely to value earth currencies as earthlings are to value game currencies, thus opening vast possibilities for arbitrage. You might not even be willing to pay one dollar per million game-goldcoins for yourself yet be able to discover players willing to pay significantly more than a dollar for some item that some shop somewhere routinely sells for a lot less than a million game-goldcoins. You might be able to buy one game-currency dirt cheap from players met in some game that does not directly honour that particular game-currency or does not value it highly and sell it to players met in some other game where that specific game-currency is the main unit of value used in trading valuable metals or crystals or even fuel for fusion powerplants.

How can one prevent a currency almost identical to bitcoin zooming up in earthling-currency exchange-value or is it, as some seem to be claiming, sufficient merely to not be the original bitcoin? Maybe no changes whatsoever are needed in the rules in order to achieve this goal? Would a new blockchain using the same rules as the original one be well able to maintain an exchange rate as low as or even lower than testnet's even though unlike testnet it is intended as an actual store of units of account not as a doomed to be erased record of units intended to be erased?

I do not favour creating far far more than 21 million coins as the game economies might well in total be smaller than the entire planet earth economy that bitcoins are intended to serve so using 21 million should seem like more of an oversupply of coins, or less of an undersupply of coins, than bitcoins seem relative to their target service-volume.

So in summary I am not yet sure exactly what needs to be changed, maybe nothing if those theorists are correct who claim that a second blockchain based on the same rules as the original blockchain would automagically find itself enjoying a far lower valuation than the original.

Since the original question is about restarting, I guess what I would do differently would be to make an extra *persistent* (unlike testnet's non-persistence) blockchain from the start, maybe by having a -game switch as well as a -testnet switch. (Maybe with use of both switches meaning to use the game testnet not the main testnet?)

If the game and non-game blockchains were thusly both created from the outset though, neither would be "the original", so we are back to the question of what might need to be different in the rules in order to lean the non-game blockchain toward real world valuability and the game blockchain toward being dismissed as a mere roleplaying game currency by short-sighted speculators who fail to see the value in games and/or game currencies?

-MarkM-

What stops me from making 1 million wallets?

Hint: Answer starts with 'N'.

I think the idea of hashing is not that great of an idea.  Many probably don't even have the client running.  I think you should randomly get BTC if you have a client open.  Then you will have people having 1000s of clients open, but I think you need to have the client open to protect the network and that is where the real work is.  Sending block parts to other people.   1 client per ip.

Mmmm, I would like such a change as well.

There was a thread like this before.

The ONLY thing I would change is to make retargets more frequent; say, every 1008 blocks, so that the network can adapt more quickly to large changes in hashing power.

What rules would you make differently?

A couple of rulesets I've thought of:

1.  Each new wallet gets a predetermined number of coins (say, 50).  Mining is paid for by transaction fees only, so coins are only added to the network through new people joining.  This would be to fight against the early adopter billionaire issue (if you want to call it an issue).  Downside is that I can't think of any good way to prevent someone from creating wallet after wallet and dumping it all to a single address.

2.  Similar ruleset to what is currently in place, only there is no decrease in the 50 BTC bounty for each block found.  This bounty would continue indefinitely, to help balance against the loss of bitcoins through lost wallets.

3.  Bitcoins are given according to hashing power provided.  This would make the currency highly inflationary for the earlier periods of adoption, and inflationary to some extent virtually forever.  Say, for every GH/s of mining capacity a miner has, they are paid 1 BTC per day.

I am curious to hear why you would or would not rather have one of these rulesets.  Or suggest your own.  Note that I am not saying that any of these should be adopted, or endorsing them in any other way, I am just curious to see what other people have to say about them.  I think the current setup is pretty dang good, except for the early adopter problem.  It's too exponential of an issue to allow universal acceptance of bitcoins as a currency, IMO.

Regardless, thoughts?