I just don't understand your concepts well enough to say, "let me propose an encoin like system, except..." Had I done that, I would have obviously been talking out my ass. I have no idea why you made most of your design decisions. Like say, why you changed 10 minutes consensus periods to X hours. How each miner is benchmarked against kWh. Or how effort based new coin allocation attempts to affect monetary policy.
All right, I'm going to at least try to convey my logic on how these systems converge into a stable price point in one post.
• Network Trust Block difficulty will be based around an average network being able to find one block every six hours. 6 hours x 50 people x 200Wh = 60kWh or 6 ENC.
200Wh is the BIG MAGIC NUMBER of encoin. The 6 hours and 50 people are variables. If the average network had 100 people, the average network would have to find one block every 3 hours to equal 6 ENC. If the avg network was 25 people, it would be 12 hours. The hash algorithm difficulty would be based on this in the same way as bitcoin's except for like so:
200Wh * NumTotalUsers = total kWh ASSUMED to be used by the network and thus the goal for the average amount of coins awarded.
Hash/s / NumTotalUsers = AvgUserHashRate
AvgUserHashRate == 200Wh
In reality it is more complicated than this since not every user will stick around for the whole block and so on. I guess "AvgNumTotalUsers" might be a better variable name.
Now, if everybody is using exactly 200Wh and has the exact same hashrate and has the exact same cost for electricity, we are living in a perfect world. Obviously all kinds of things will affect this, and the user will have to determine in their own case whether or not it is profitable to mint coins.
BUT--because I mentioned there are voteable payout structures within each TrustNet that can help smooth out variations in efficiency. For example, set payout structures based on finish for best hash. 1st place hash always gets X%, 2nd place always gets X%, and so on (this is not how it works in pools and it simply can't be done this way). This will keep the race for the newest and best hardware down to a minimum, and people with inefficient hardware can siphon off a bit on the people with more efficient hardware (I'm talking in terms of Mhash/W or Mhash/J, please don't misinterpret this). But there will be a minimum to qualify, and a requirement to be present until the end of the block, and so on to keep abuse down. These are voteable properties (trust modules--another terrible name I know) so they can be adjusted down the road in the face of new exploits or whatever. Kinks will have to be worked out.
The point is to award electricity, not efficiency as I stated before. Encoin will require less of a hardware investment because of this which means it can potentially draw more people in. More people == more popular == currency is valued. And hardware investment uses a god awful large chunk of any ROI you might see for a very long time. So the less hardware investment the better. We're looking for sunk costs here.
Now with the whole TrustNet system, it is almost like a game. Gain levels (trust), get more coins, be able to join respected networks, and so on. Who knows, I may come up with a shitload more ideas (aliases for trustnets and users is one, cheaper global messages could be another, all kinds of things to help make it a community). If the network doesn't like them, they don't vote for that module. If eventually people do and the vote is >50%, then it becomes part of the network. It's so much more flexible, although it may be hard to see at this point. Nothing that trustnets can vote on will be able to change the value of the currency though. Clients will not accept that unless there is some other designer who comes along and decides to fork it. The potential for that can't be denied, it is the nature of open source.
But *anyone* can always become trusted and get the benefits. They just have to stick around awhile. And for every person you can convince to stick around, the more secure the network is. This does not mean the more hash power there is because of the cool-down and the ENC benefit. That stuff has to be worked on in a little more but let me give a quick example:
Say the ROI on a typical minted coin is 33%. An average coin costs 10kWh, say the average price is 0.15/kWh, so $1.50+33% = $2.00 sell price. Market "crashes" to $1.60. TrustNets go into cool down mode and only mint 1/10th the coins at 1/10th the electricity for 1/8th of the award--I'm thinking of allowing 1/2, 1/10, and 1/20 but we'll see. Now 0.75 ENC is made for the cost of 0.6 ENC and people who need to make money to see any use to minting get a 25% ROI inherent, but on a very small amount. If the economy is routinely stable, people can run Encoin as essentially a background process that uses very little of their GPU. The computer can still be (almost) fully used. Yet the network remains just as secure as before. It avoids proof-of-work being the end-all be-all form of consensus.
And all kinds of neat things can be done when you can rely on consensus in lieu of proof-of-work. When you can know how many people are out there and what the network looks like. When you can vote on little details of how the network operates. When you can chat with your fellow TrustNet members. When you don't have to worry about the network becoming vulnerable if the hash rate drops. LOL when you don't need a 8gb/s connection to use it, or download a 1GB block chain before you can even see your first transaction. I have a vision, and that vision is Encoin (but possibly by a different name hahah).
You gave constants for coin creation that seem unrelated to the transaction fees. I don't understand why you chose those constants and why (as both are part of monetary policy decisions) they are unrelated. If they are place holders to be calculated later that's OK. I just genuinely don't understand yet.
Coin generation has to be based on kWh, it can't change because then 1 ENC != 10kWh. So the only way to constrict the economy when in need is by transaction fees and allowing coin generation to securely reduce in output. They can not be tied together. Transaction fees
could be variable, and perhaps there could be two fees, one for stable/expanding economy, and one for contracting economy. But this would have to be voted on by the TrustNets, the software can't decide on its own, it is against how I think this should work. That could piss off people who don't make coins. So I would really like to have a transaction fee set in stone.
What I'm trying to understand is to what value/function does this system converge: If a new ASIC is 100 times more efficient? If a big new vendor joins the network? If the price of electricity spikes? If in any situation, the value of coins tends either to zero or infinity, I know the algorithm is not stable.
If some new thing that only 1 or a few people had was way more efficient, then they would siphon some money off of the new coins being made. It's hard to calculate how much, but the bigger the network, the less it will take from each coin (and I have even more ideas to lessen the impact of this type of thing, but it is for a much more advanced version of the proposal). I don't know what you mean by a new big vendor joining. I don't see how the price of electricity could "spike" especially across the entire world at once. It should be a gradual process that will gradually increase the value of encoins vs fiat. The value of coins tends to 10kWh, whatever 10kWh is.
To combat increases in hardware efficiency over time, you propose rotating algorithms. Rotating doesn't seem to work, because if someone is willing to build an ASIC they might as well keep it around for when rotation comes back. 4 algorithms, 4 ASICs. Run when appropriate.
No, I proposed changing algorithms. I did think of this exact same scenario, believe it or not.
Beyond that, your concept seems to require competing engineers. Who decides if existing hardware is too efficient? Who gets to choose the next proof-of-work algorithm? How do you make everyone else adopt those new rules? That seems like a different kind of consensus building completely unrelated to your TrustNets.
If by "too efficient" you mean in the future where everything uses less electricity, I have an idea that will add in some deflation to the economy slowly, over a very long period. And it is simple and regulated and would not cause any drastic consequences like satoshi's boneheaded award halve. This is what I was referring to when I said I didn't want to give something away yet. This deflation can still be counteracted by increased minting and so on if necessary. The proof of work algorithm will be changed randomly, every X primary blocks, by using a hash of the previous block. When the developers (or anyone with clout) releases a new module with a new, futuristic hash algorithm, it will be added to the pool of algorithms already in use by a TrustNets vote. The vote is not a one time thing. When people upgrade their client, they will have an option to vote yes or no to this new potential change to the network. Then the client will automatically cast their vote during every primary block cool-down phase. Once >50% agrees, then it is added as a valid module.
I'm genuinely interested in understanding how your system meets its monetary policy goals. To me, all the other bits are optimizations to previously solved problems. They are interesting in their own right. I'm just interested in monetary policy first.
Let the people control whether or not coins are minted. Coins are always minted for approximately the same cost, so it is not like they are voting them into existence from nowhere like fiat. They always have an inherent value of 10kWh. Transaction fees are a mild deflationary measure to counter over-minting and low demand without requiring an expansion of the economy, even though stable economies
tend to always expand as population increases.
I'm going to give some reasons why I chose 200Wh over 100 or 300.
100Wh disadvantage:
* Electric value enters the economy at a slower pace. People would have to adjust their GPUs to make sure they are using only approximately 100Wh, half or whatever it comes out to be. (I plan on having a helpful calculator in the program.)
* With a higher difficulty (time * computation, I'm referring to time here) of adding value to the economy, extreme deflation is a lot more likely if the network booms.
* Psychological. A month of effort nets you 7.2 coins if I keep the 10kWh figure. I could make it 5kWh I suppose.
* Potential for abuse. Hacked clients *will* come out that divert half of your GPUs resources to one node and half to another so that you can make double the profit. This isn't a
huge problem, but by restricting it to 100Wh you are allowing easy access for some people to have this advantage over others. I suppose, in the end it really doesn't matter but it would waste network resources just to allow people who want to go full blast to go full blast.
100Wh advantage:
* FPGAs and what have you have less of an impact. While they're still going to be more efficient, they are not siphoning off as much of the economic value.
* Future cpu/gpu electricity usage going down could have less of an impact because instead of going at half the normal rate they could go full blast or whatever ends up being the equivalent.
* Since deflation is more likely in a boom, it would encourage more people to be part of a TrustNet (blah blah whatever you want to call them).
* There would be less "extra coins" in inflationary periods. e.g. the inflation won't be as big because people are not putting in as many coins before they figure out it's happening. So it could bounce back quicker.
300Wh disadvantage:
* A lot of people may not be using 300 watts worth, so they are getting a bonus even though they are not using as much electricity (assuming an efficient card).
* FPGAs could have a much larger impact if they become any significant portion of the economy.
* lot of extra coins in inflationary periods, may take a very long time to settle back to equilibrium.
* lot of extra coins in deflationary periods, that won't last long and the demand will probably all be grabbed up by anyone that was there before the boom.
300Wh advantage:
* basically the opposite of everything in the 100Wh disadvantage category.
