Can someone develop this?
3 coins isn't really pocket change. I'd be willing to use a trusted forum member as an escrow for the transaction.
Bump
It sounds like what you are proposing, in essence or in effect, is that in some far future year when the 50 coins per block that GRouPcoin constantly mints happens to work out to be 2% of the number of coins generated up until that moment (up until that block, in practice), the time will have arrived for GRouPcoin to adjust its rate of coin-minting if, by that far future time, your theories as to what would be the perfect mining curve still seem to be as valid as they seem today and the benefits are actually worthwhile enough to actually bother to change the number of coins per block.
Thus I suggest that the coin you want is already up and running, has been for years, and still has years to go before it will need a tiny tweak to make its generation rate stick at 2% (or whatever percent has been determined, by that future point in time, to actually be the exact precise perfect percent rather than just some guess made by you years in advance of the actual coming into effect of such a rate change).
So we might as well just keep chugging along using GRouPcoin and simply add to our over the coming years concerns the concern that possibly it might prove beneficial, at some future date, to cause the generation rate to stick at some percentage of the total coins minted, so that over those coming years we can make observations and track statistics and so on that will enable us to know, come that far future, what exact percentage is in fact ideal. (If it does turn out to be 2%, how many decimals of accuracy is that? 2.0%? 2.00%? 2.000%? To what number of decimals has it been determined that "zeroes all the way down" is in fact the ideal?)
Basically though, there are years yet to work out the exact best percentage, in the meantime the 50 coins per block has been chugging happily along and will continue until the exact details of the later stage are worked out / negotiated.
Please note also that your concept of "annually" can be a whole can of worms, because nodes are not synchronised in time, their synchronisation is by block number. So we also will not even know how many blocks any given year will actually consist of until we know how much up and down see-saw of hashing power will be happening during that year...
(For example the advent of ASICs is likely to make a "blocks counted year" much shorter than a "calendar year" this year for bitcoin...)
-MarkM-