If I was gonna do it, then I probably would set it up with the following parameters.
Same 21 million just like Bitcoin. There is nothing wrong with this number. But instead of 50 coins for the first 210000 blocks, I would make the reward try to mimic an objective measure of the community size with a formula.
Here is an example formula (BTC per block) for that: B * (ln(difficulty)2 4) where B is a percentage of coins not yet distributed (1 for 21 million, 0.5 for 10.5 million, 0 when they're all distributed).
This would result in 50 coins being given out per block once the difficulty reached 2.3 million. Current difficulty of about 70000 would pay about 30 coins. If I fired up just my own GPU's (eight 5970's), the difficulty would be about 700 and the payout would be about 9.50 per block.
Then I would correlate the maximum block size with the difficulty.
so, not inflata-coin, just a different block bounty distribution over time, i see.
the idea of indexing it to the difficulty is actually quite an interesting one, i think.
as long as you keep the chain separate from the main block chain (so that the mainchain clients don't have to deal with the extra traffic from your fork that would be treated as invalid), i see nothing wrong with you giving it a try to see what happens.
i still foresee some difficulty in attracting miners away from the main bitcoin chain and into the new one. early adopters need to be able to expect a higher compensation than the later adopters, in order to be convinced to put their effort in, at a time when the risk is highest. on this metric, main bitcoin still has you beat. but compared to inflata-coin, it's a lot better in this respect
if, while you are at it, you also put in some other improvements (or things that some people would perceive as improvements), e.g., more decimal places of divisibility, etc., it may have some future.
