Thanks for the clear proposal and the nudge to provide feedback!
-
The normalization model seems solid to me; variations on this have been floated in other forum posts, but this captures the key features the community seems to agree on, including the need to start ramping down emissions and the inefficiency of rewarding non-stable borrows. I happen to disagree with the last point because it breaks COMP’s initial promise of “market participation yields governance rights”, but the community consensus on this seems clearly in favor.
-
The decay schedule itself looks good and is well-reasoned (disclaimer: I haven’t checked your formulas for accuracy but would do so before voting on-chain). My concern here is with the implementation as a series of governance proposals at 28-day intervals. Even though the proposals can be expected to be mostly copy-pastes of one another, running them monthly seems like an avoidable security risk, on top of not being an efficient use of community members’ gas to confirm or contest each monthly decay. Given the 10-year time horizon, wouldn’t quarterly or even twice-a-year adjustments be adequate? Finally, it’s tempting to wonder whether the decay schedule could be implemented by a change to the comptroller that uses epochs (based on timestamps or block numbers) to determine the correct compSpeeds; this would reduce the number of proposals to vote on at the cost of increased risk of smart contract bugs, so I can understand hesitancy around such an alternative.
-
For my critique on the compensation plan, you can pretty much take item (2) from my feedback on GFXLabs’ proposal and replace each instance of USDC with DAI (while acknowledging that this proposal differs in that it seeks split compensation in DAI and COMP). I definitely want to see this ultimately funded, as with GFXLabs’ proposal. I also don’t want to see these two awesome proposals unintentionally drain the protocol’s stablecoin reserves or meaningfully limit our ability to recruit and retain other developers as the protocol’s user and dev communites evolve. It would help me get on board if you could share a back-of-the-envelope calculation showing what fraction of expected annualized DAI revenues coming into the protocol will be spent on this proposal at the requested $3M DAI / $1M COMP annual rate. Personally I would want to see this figure come out to less than 20% of annual DAI revenue but would support up to 50% of annual DAI revenue. For reference, by my math, the GFXLabs request comes out to about 100% of annualized USDC revenue, which makes me question whether it is a sustainable commitment for the protocol to make.