On the surface this sounds like a good idea.  But then how would it work in terms of the bitcoin creation schedule?  What would happen to the 21 million coins around 2030?  If you gave out more coins when the difficulty was higher that would just encourage the difficulty to go higher.  If you gave out fewer when the difficulty were lower that would encourage the difficulty to go lower.  The whole point of the mining schedule is to build up a nice stable base of miners to support the network.  The point of the creation schedule is to give people confidence that bitcoins won't be debased by a central bank and that their money will retain value.

This is not relevant because you can also vary the number of coins per block to achieve the same effect.

Yes, keeping to the coin generation schedule is a big issue. I don't see why the 50..25..12.5 coins per block commitments are important, however. What matters is that there must be approximately x million bitcoins at the end of 2011, approximately y million bitcoins at the end of 2012, and a maximum of 21 million bitcoins in the infinite future. You can approximate these rules without holding to the nitty-gritty details of the short-run generation schedule. I think these nitty-gritty details could be adjusted to dampen short-run swings in the exchange rate.

Some points:

1) volatility will die down naturally if merchant adoption succeeds on a large scale. Thus the fact that volatility cannot be dampened eternally is not worrisome.

2) The biggest problem is short run volatility- say booms and busts within a 1-2 month time frame.
 
3). My suggestion is two adjustment schedules for the number of coins per block. A short run adjustmenent schedule will cause the number of coins per block to rise when difficulty growth exceeds long run trends and fall when difficulty growth falls short of long run trends. The definition of short-run here could be say 1 week. A long run schedule will try to keep the long-run average number of coins per block set at 50 or 25 etc. with the definition of long-run here being say 6 months.

4) This suggestion would make the hashrate more volatile and harm security. Accordingly, you cannot allow the number of coins per block to drop to zero or shoot up to infinity, since the hashrate volatility associated with this will make the system insecure. I think you could allow the number of coins per block to drop and fall within say a 20 coin to 125 coin band without causing unacceptable damage to security.

Implementing this should enable some short run exchange rate volatility to be passed on to miners. I think that miners can bear volatility better than merchants.

Fixed exchange rate systems are susceptible to speculative attacks and currency crises. If the speculator's cryptocurrency warchest is larger than the central bank's fiat warchest (merchants are the central bank here), the speculator can break the peg and cause the cryptocurrency's value to plummet. By buying up options to sell the cryptocurrency at the peg value beforehand, the speculator can profit by simply exercising the options.

I doubt the merchants will keep big enough war chests to fend off these attacks.  

Here is a famous example to make things more concrete:
http://en.wikipedia.org/wiki/Black_Wednesday

In short, I don't think the fixed exchange rate systems make sense.  The goal of my system is not to prevent bitcoin from changing in value, but to establish some roadblocks to slow down the process of appreciation and depreciation.

This same topic already was created recently, in the right forums:

http://forum.bitcoin.org/index.php?topic=28296.0

What formula would you propose to achieve this? To be implemented, it would have to be strictly mathematically defined (and every node should come up with the same answer, so it cannot be dependent on external information).

Yeah, I thinking of a function in which current difficulty and a series of past difficulty levels are the right hand-side arguments. Every node should agree on these right hand-side arguments. A specific formula that works may take some time to come up with. I may screw it up on the first try. I will think about it for a few days and post an example then. In the meantime I welcome any suggestions from others on a working formula.

This is why some of the groups creating or thinking of creating their own currencies have been thinking of issuing half their coins up front.

Basically unless the issuer is the speculator, the speculator's warchest of the cryptocurrency can only exceed the fiat warchest of the issuer if the issuer fails to retain as reserves the fiat that the speculator, or who-ever the speculator obtained the cryptocurrency from, paid for the cryptocurrency.

How can someone possibly obtain more dollars worth of what you sell than you have dollars that you sold it for?

Only if you cash out instead of acting as a responsible backer holding what you sold it for as reserves with which to "back" it.

Such considerations have led to the current work that is being done on implementing "mining licenses".

EDIT: Consider PayPal, issuing database entries of PayPal Dollars (PPUSD). If they fail to have enough actual dollars to "buy back" those PPUSD with, they are acting as a "fractional reserve bank", which might be beyond their mandate/license (or might not?)

-MarkM-

Yeah but what the heck, the trust issues in currency are already huge.

How many wheelbarrows full of paper does it take to buy a loaf of bread?

Adding all the audits and "legal requirements" doesn't sound that great an idea when I think about the purported agency of MtGox's big recent problem (an auditor) and the agencies of the wheelbarrows of money to buy a loaf of bread scenarios (the lawmakers).

Nice thing, purportedly, about free markets is anyone and their dog can put a currency of their own on the market and the market will decide whether the dog's currency or the anyone's is worth anything.

In the Galactic Milieu, the Hackers, being possibly about as mythical as the preposterous mythical planet known as Earth, are reckoned pretty amazing so their currency (Bitcoins) is reckoned pretty valuable. The Martians are not reckoned as advanced but are thought to be mighty in war so theirs too is reckoned maybe ok, maybe especially if you think "sending in the Marines" might some day be necessary. The Brits and Canucks are held in reasonable esteem among Democratic nations so theirs too is widely traded. And so on.

A big question is whether the major holders of bitcoin who picked up huge quantities early on did so in order to be able to build reserves of fiat to uphold bitcoin's value or are just ponzi players cashing out and leaving bitcoin in the dust.

-MarkM-