As far as this end of the technology is concerned I am not sure how that all works out...

Can you explain:
1 How yours solves the issue of inflation?
2 How yours can calculate a meaningful price system?
3 Outside of a commodities market, how does yours have a meaningful value for the BTC?

Alright that sounds interesting. I don't know how fast can propagating blocks go, but let's assume for starters that it takes 10 seconds to propagate a block across the network. There are 8,640 deci-seconds in a day, which means that on average, the system will be able to support something like 2,880 daily blocks (third of 8,640) at a maximum to reasonably avoid collisions. In other words, if we froze the difficulty of generating 1 block so a modern computer would need to work for 1 day on average to create it, it wouldn't be good enough because we hope for much more than 2,880 users to join soon.

The solution would be to make generating a block extremely difficult right from the start. Say, a modern dual-core computer would need to work continuously for a month to generate 1 block regardless of how much CPU power is in the network. With the 10-seconds assumption, this will give us room now to accommodate about 86,400 (which is 2,880 x 30) connected machines without considerable problems. You might complain today that a month of waiting is too much, but assuming the current model succeeds, it will soon need much more than a month to see a block coming under it. Additionally, coin production will be done via specialized businesses and dedicated computers left to work only for that purpose who'll sell to you and me, so a month wouldn't bother them and it would create a killer proof-of-work. And as NLS always fears, a botnet can and will stop all of us from producing under the current model. A botnet will not be able to do so under the suggested model. It will only make its owner rich, which happens anyway.

"What if we listened to your stupid idea, and soon enough we got more than 86,400 connected users?" you may ask. First, please don't resort to name-calling. Second, remember that as time progresses, internet speeds improve faster than computer speeds. So by the time we get 86k PCs connected 24/7 (which may take a couple of years for example), we'll have a propagation time of only 5 seconds. Two more years and it will be 3 seconds, naturally allowing the system to accommodate more newcomers (including a couple of botnets ) without destroying the economic system via perpetual deflation. The network however wouldn't grow too large because Westerns would prefer to buy coins from people who live in cheap-electricity countries to save on their bills, keep reading...

I think that request was for NLS but I'll answer them for my model too

Under my model (and indeed, the current one) the generating costs are generally tied to the price of energy. Those may go up and down all the time, but aren't affected by how much people are in the network and don't double every four years either.

Prices will be determined in my model (and also in the current model) by how much cost (i.e. electricity mainly) was spent to generate a block of say, 10 BTCs (or whatever).

My favorite question . If we assume that running a high-end computer in a Western country continuously for a month costs $10, then under my model, a ฿10 block will more or less have an agreed-upon value of $10 ($1=฿1). As more people from countries with cheap electricity (eg. Venezuela) join to exploit that difference and compete with each other in selling to Westerns, price will more or less stabilize just a tad higher than the average cost of electricity in these countries. This will also serve as a way to limit the network so it wouldn't grow too large for propagation (because why run your PC for a month in the US and pay a $10 bill when you can purchase that block for $4 from an Indian dude while avoiding the melting of your machine -and nervous system- via all that heat and noise?). The network will be filled with generating Mexicans and Zimbabweans (and botnets) because they're most efficient at electricity cost. It's a shame because botnets may eventually become our main providers, but that's better than having them ruin the whole system by stopping participants from producing under the current model, right?

Now that Bitcoins would rarely increase in value (because people will continuously find a place where electricity is 3% cheaper), nobody will be hoarding his coins. People will gladly use them as a means of exchange and even spend them ASAP before they lose another 3%, causing the ฿ economy to flourish. Additionally, even 20 years from now, anyone will be able to generate a coin by running his machine for a few weeks, granting newcomers who don't want to purchase from botnets for ethical reasons or from Mexicans for security reasons a chance to join at any point in the future.

I wholly appreciate NLS's model, which is exactly like mine with the difference of freezing the generation at, say, 2,880 daily blocks. Under his model, if one computer is connected, it creates 2880 blocks/day. If 10,000 computers are connected, they create something like 1 block/4 days each. The problem with this is that the cost of generating a block would continue to deflate for the foreseeable future until the number of participating nodes stabilizes, which could never happen as long as the price keeps increasing because of the increasing cost due to more people joining the network (due to more people using the internet and hearing about Bitcoin, for example), inevitably having us stuck in the perpetual-deflation scenario. Additionally, if a couple of large botnets join the network (which will happen sooner or later, to be honest), the average user would then have noway to generate his own blocks and would be forced to buy from them instead. Some people afraid of government witch-hunting those who purchase BTCs might be afraid to use conventional methods like paypal and would thus have almost noway of acquiring coins (both because botnets/crowds make generation difficult and because 16 years have passed effectively making it impossible). Finally, if those botnets/supercomputers belong to a government, they would not sell their 2,879 daily blocks in order to hinder Bitcoins's usability. Ego aside, that's why I prefer my model of user-base-dependent-block-generation over NLS's of perpetually-fixed-block-generation.

Satoshi, please do consider changing the current model to one of those two (or something new). We're not trying to win an argument here; we're trying to find the best way to make this project successful. This becomes more difficult with time. If we're going to change anything before the word catches on, now is the time to do it.

My exchange rate will not always rise without ever falling. The only reason why it is rising steadily right now is because I calculate it using a large average to reduce volatility. If I used a smaller average, you would regularly see much higher highs and very frequent lows.

Neither the current model nor my model will continue to deflate forever. Once generating bitcoins becomes exceedingly difficult, most people will not try to generate bitcoins.

I know that my model is similar to the current model. What I don't like about the current model is that people generating bitcoins at any given time can generate twice as many as others after four years. That is what I want fixed.

It's Suggester's model that strives for price stability from the beginning.

Democracy Dollars is now the name for my theoretical currency which is based on Bitcoin but with a few modifications. $1.00 DMD would be awarded per generated block instead of ฿50.00 BTC. The amount of dollars awarded for generating a block would not decrease every 4 years like with Bitcoin. Two decimal places would still be displayed because dollars usually have two decimal places. Ideally users would have the option to display and use hidden decimal places. The client would be called Citizen Bank.

BTW New liberty, what is the site you do your exchanges at...
I would like to look.....

I think you're mixing the terms "inflation" with "deflation". We currently have a sharp deflation (i.e the currency's value is increasing). I predict it will continue to occur.

Yes. If he's willing to pay less, less people will produce BTCs causing their production to be easier until price equates with cost. If he's willing to pay more, more people will produce BTCs causing their production cost to increase until it equates with cost. If you don't understand this very basic economic principle then you should've read the previous posts. And under my model, a similar equation is in effect but is not changed as more people join the network and isn't doubled every 4 years.

No. And I'm tired of having to explain why a deflating currency is a bad idea. Please refer to previous posts.

It does. Because as long as energy costs don't increase by, say, a constant 10% annually, coins won't.

Oh God...
Did you forget that 99.99% of goods purchasable via coins are also purchasable via cash? You're not making any sense.

Refer to my previous posts, please. I thoroughly explained why this won't happen.

Because it greatly reduces the wish to spend now.

Many will leave the system via computer failures, many will leave the system via sabotage attempts, and there would be noway to bring them back.

That's exactly what a "monetary deflation" is.

I'm mysteriously resisting the urge to give a really mean response here.

Same here. The current model doesn't grant that.

They go hand-in-hand. With constant deflation there would very little market trade if any.

If he earns $ via exchanging them, he wouldn't have a reason to stop generating even more. Nominal prices now does indicate that more people are producing, whether botnets or otherwise.



You forgot the most important question: How many DMD's do you suggest being produced per day on average? You do want it to be a fixed amount, no?
I did a quick and dirty research and it seems that the lowest electricity cost worldwide is about 1/4 of the US average. In an attempt to shoot for ฿1 = $1 to make calculations easier, I think making my block worth about ฿3 (for a month of work) should be close enough.

So you want the rate to be constant @ 144 blocks/day forever? Why not increase it to, say, 1/minute (1440/day) so the system could accommodate more nodes without much delay? With 10,000 nodes at the current difficulty, you need to wait an average of 70 days for a coin to appear. If your dollar became an international currency, you might have to wait for a whole year or more for a coin to appear!

I imagine that my idea will make downloading the existing blocks very difficult after a couple years or so though because of their size (1M blocks?!), isn't there a way to compress them or something?

I could just name mine Electric Bitcoins (EBC's, yummy!), because they're directly proportional to how much electricity you've put into them regardless of how many people are in the network.

My model should be fairly simple to implement. We'd just freeze the amount of effort needed to find a new solution so the odds of finding it would need, on average, a month of continuous work for a modern 2.0 Ghz core-2-duo for example.

Satoshi and other pioneers: I understand your reluctance to change the current system: you guys don't want to sacrifice all the effort you've put into the project and start from scratch with no coins. But all we need to do for Bitcoin 0.3 is remove a couple of zero's for an average month's production to equal 3 new coins for instance. Perhaps every 4,000 Bitcoins can equal 1 NBC (new bitcoin) or something like that?

At 1 block/month on average, with only 30 computers in the network it would take about 1 day to create a new block. 1/2 a day with 60 computers etc.

I admit that we need to correctly estimate the minimum amount of time to broadcast a new block across the network though. And is it possible to use timestamps for blocks so even if the difference was 0.1 second they could be re-organized on our machines? How about something like programming lines with 10-figure differences reserved in case of double-block-creation before broadcasting is complete, is it possible?

That's what I'm talking about. If we have something like "spare blocks" designed to be occupied by double-entries as in programming lines (they go 10, 20, 30 not 1,2,3 to make room for later changes), and if we have timestamps which say when was each block exactly generated down to the 100th second, maybe we can solve that problem altogether.

If worst comes to worst, and the number exceeds the proposed 86,400 before we achieve enough technological progress to significantly reduce the latency time, then some blocks will inevitably be lost until this progress is achieved. So, say, if 100,000 nodes are simultaneously generating coins, maybe 10% of the new blocks will be lost which should be an incentive for some of them to leave the network (or, ironically enough, perpetually raise the price they charge to accommodate for their losses!).

This scenario, however, is extremely unlikely. If each node is connected to 10 exclusive others, it should take only 5 "hops" to reach the whole 100,000 connected network. How much time would it take for each hop at most?