Quote from: JohnDoe on September 20, 2011, 08:03:28 PM

Quote from: jtimon on September 20, 2011, 05:04:54 PM

Investors and debtors will want inflation.
Creditors will want deflation.

You are right, that is undoubtedly their self-interest, but I believe they mostly cancel each other out. Creditors will lend 100 coins and vote for maximum deflation, then debtors will spend/invest the 100 coins and vote for maximum inflation. Debtors get another chance to vote with 100 or more coins when they pay the creditor back but this offset by the higher than average probability that the money sent from creditor to debtor had more confirmations than the money sent from debtor to creditor.

Not sure if all this get balanced.

Quote from: JohnDoe on September 20, 2011, 08:03:28 PM

Quote from: jtimon on September 20, 2011, 05:04:54 PM

Even with demurrage, money holders will want their money to appreciate.
And yes miners will want more subsidy, even if it will be translated in more competition later (that neutralize their increased profits), that won't be instant, the increase in the subsidy would happen first.

I guess you are right. What if the demurrage rate were inversely proportional to the supply growth rate? That way money holders have no incentive to push for deflation because it will translate into increased demurrage and miners have no incentive to push for inflation because it results in less revenue from demurrage.

My initial proposal for demurrage and inflation rate to influence prices was:

Quote from: jtimon on August 05, 2011, 05:30:26 PM

NCC = newly created coins
CDD = Coins Destroyed by Demurrage
With deflation you increase both NCC and CDD, but NCC more than CDD
With inflation you decrease both, but you decrease NCC more than CDD

I guess this is incompatible with demurrage being inversely proportional to the supply growth rate.
But definitely removing miners bad incentive is a priority. If you achieve that, you can just use a decentralized price index.

Quote from: JohnDoe on September 20, 2011, 08:03:28 PM

Quote from: jtimon on September 20, 2011, 05:04:54 PM

About the concrete voting system, I would suggest to define a reference currency and the people vote how higher or lower the chain currency is compared to the reference.

Yeah, I had Terra in mind when I suggested to follow a price index for the recommended/default rate, but I get the vibe you mean create a real parallel blockchain just to set the reference. I don't see the need for that.

EDIT: I read your thread more carefully and it looks you didn't mean real currency that's actually issued, so we are on the same page.

Yes, I don't want to issue or back it. I just want to use a decentralized price index to be able to measure the price (in bitcoins, freicoins or whatever) over time. But for this you need to remove the voting incentives for miners, which they won't have if the reference currency doesn't influence the rewards and issuance of the host chain.
If you want to use the reference coin for more than just contracts and price setting (to make it the target for your dynamic supply chain currency); then the system becomes more fragile. Remember that the time travel attack was based on cheating the timestamps that miners "vote".

Quote from: dacoinminster on July 25, 2011, 04:17:32 PM

Quote from: cunicula on July 23, 2011, 12:27:23 PM

This is a good idea, but how is the infomation about exchange rates fed into the blockchain. You need I thinlk two items of third party data: the current USD exchange rate and the exchange rate between current USD and whatever commodity you are pegging value to. Presumably, humans need to supply this data. How do you incentivize the humans to supply honest data?

If you want a distributed currency pegged to ANYTHING, this is one of the biggest technological problems to solve. However, consider this: there are also attack vectors on bitcoin that involve fraudulent timestamps. Bitcoin uses a distributed timestamp protocol, where nodes reject timestamps that differ significantly from what they think the time is. I believe the same logic can be extended to exchange rates. If somebody lies about the exchange rate, other nodes will reject that block. Consequently, I consider the problem of distributed exchange rates a (mostly) solved problem.