victorges
commented
1 year ago We've decided to go with #611 instead given the reward() call requirement is not that big of an issue, especially since by avoiding the "lump sum" minting to the treasury we will also make sure that our effective inflation rate does not change with this upgrade.
e.g.: on an extreme case, if no transcoder calls reward, we get no effective inflation regardless of what the protocol inflation is set to. If we had a lump sum mint to the treasury on every round start, we would have an effective inflation of at least that on every round, which changes the expectations on how the inflation is actually going to work.
This is another simpler implementation for the treasury rewards calculation, mixing #611 and #612. The caveat is that we will have a bit more rounding errors on this approach, since there are more perc calculations. The losses might be relatively trivial tho (worst case 1 fractional unit of LPT for each transcoder, so up to 100/1**18 LPT per round).