Hifi Improvement Proposal (HIP) 1

Summary (TL;DR) The first HIP introduces several technical improvements to enhance the user experience, software updates to stay up to date with the package dependencies Hifi v1 relies on, and a new license to protect us from unauthorized forks. These changes will be deployed on Polygon Mainnet.

Background Hifi v1 has been running smoothly on mainnet for a few months now, and this has given us the necessary time to slightly change the architecture of the protocol, and iterate on the user experience.

This proposal is the result of these iterations.

Problem There aren’t any real problems with the current version of the protocol, but there are quite a few things we could iterate on and improve. A low-severity bug was recently found in the protocol, a contract which is currently upgradeable doesn’t need to be, and the user experience could be improved through the vault logic as well as the transactions when interacting with the protocol.

The community has also mentioned at several occasions that the fully open-source nature of the codebase wasn’t necessarily appreciated due to the possibility of another project copying our code.

Solution What follows are the key elements of our proposal.

Make the Fintroller non-upgradable When we developed the first version of the protocol, we were unsure whether the Fintroller would end up holding users’ funds, along with the BalanceSheet. With flexibility in mind, we made it upgradeable. But it is now clear that the Fintroller will not hold any funds - it’s just a permissions controller. If there’s a bug in it, we can redeploy a new version and reset the permissions to the original values.

Most importantly, it’s much easier to maintain the Fintroller as a non-upgradeable contract because it contains many structs, and structs cannot be edited when upgrading.

Fix low-severity bug in Repay Borrow event A low-severity bug was recently found by one of our engineers in the Hifi v1 protocol and this release provides a fix for it. The bug is low-severity and doesn’t impact users’ funds. The above linked GitHub issue gives more context, but the bug is simply that two function parameters were in the wrong order.

Rearchitect underlying backed hTokens This logic has been removed from the balance sheet and placed in the hToken implementation. Depositing underlying to the hToken contract provides a better UX for end users, and removing underlying-backed vaults solves an issue related to using volatile cryptos like WBTC as “underlying” for hTokens.

Update the Software License We propose to change the current license (LGPL) to the Business Source License 1.1 (BUSL-1.1 1) which restricts unauthorized commercialization of our codebase until June 1 2024 after which the license will be changed to the GNU General Public License v3.0 (GPL) or a later version of it.

This will prevent unauthorized forks of the Hifi protocol and allow the Hifi DAO to benefit from the strong foundations the core team has built in the past couple of years.

Upgrade to Solidity v0.8.12 This new version of Solidity improves the JavaScript / Wasm binary and fixes several bugs. It also offers a partial migration to Typescript.

Upgrade to latest versions of PRBMath The leading advanced fixed-point math library which we created has had several small updates in the past few months.

Add support for EIP-2612: Signature token approvals This is a major upgrade for the user experience of our protocol. Users can now authorize the spending of their tokens in the same transaction as the actual spending of their tokens.

This is a first step towards an eventual deployment on the Ethereum mainnet where the gas fees are significantly higher and where this will really help users save money and make the general transaction flow smoother.

Conclusion We think this proposal is the right step forward after releasing Hifi v1 on polygon mainnet a few months ago. It greatly improves the user experience, especially for new users. It protects the codebase from unauthorized forks and fixes certain issues regarding the protocol architecture.