Developer Call Notes
Jared – Gateway Update – (Starts at 00:00)
Jared from Compound Labs provided an update on Gateway development. After running into issues with the current Gateway implementation on Substrate, the team weighed switching to other chains before ultimately deciding to stick with substrate and re-architect the implementation from the ground up. The team decoupled the application code from Substrate and separated the validation of blocks from Oracle capabilities in order to increase security and permit the use of hard code keys for Oracles. This also permits the team to not need to rely on off chain workers in Substrate. The new architecture greatly decreases what the runtime is doing, splits the responsibility to the Oracles, and still permits the optionality to operate as a parachain. Code progress can be monitored in the Compound GitHub and in the Compound Discord.
Getty – New Assets – (Starts at 10:10)
Getty reviewed Proposal 56, which added collateral factors for four assets: MKR, SUSHI, YFI, and LINK. He detailed how the new CFs represent conservative thinking based on three main factors: exchange liquidity, exchange market availability, and market use cases. While the CFs remain relatively low for now, these factors can always be revisited further down the road and eventually moved higher. More information can be found here.
TylerEther – COMP speeds – (Starts at 13:50)
TylerEther discussed a few bugs related to the compSpeed parameter. He mentions that two separate but related bugs were affecting particular markets. The first bug arises when compSpeeds are set to 0 for new assets and then an attempt is made to increase those speeds. In this instance, the compSpeed would not update as expected. The second bug takes place when comp speeds were changed in between hypothetical blocks [A], [B], and [C]. The speed dictated on block [C] would retroactively accrue COMP from block [B], which represents the last market activity regardless of how much time has passed. Neither bug threatens fund safety or protocol stability but should be patched nevertheless. Finally, TylerEther mentions a thought for the future where the community could set separate COMP speeds for borrowers and suppliers. More information regarding these two bugs and future improvements can be found here.
Whitelisting Test – Arr00 – (Starts at 22:36)
Arr00 delivered a quick update on the previous discussion surrounding a community whitelist for individual addresses to be able to propose improvements without the 65k COMP required by Governor Bravo in the status quo. The code is working properly on the Kovan Testnet and fork testing has been completed. The next steps are a small community bug bounty and drafting a proposal. community multisig appears to be operating properly as well. Arr00 will continue to provide updates and move forward with the proposal.