- Deterministic: The content of a proposal is objective actionable items, not the subjective human-readable text that comes along with it. Subjectivity threatens guarantees that are paramount for a DAO's governance.
- Understandable: The idea is that anyone should be able to quickly know what a proposal does (most probably through a user interface). This principle is paramount to lower the risk surface of malicious proposals camouflaging as complex ones.
- Secure and simple: The
Expensetype only works with ERC20 transfers. ERC20 allowances, with its resulting complex expenditure operations need to be done from other entity (like a multisig). The rationale is twofold:- The DAO shouldn't be voting on everything. It should vote on major decisions with deterministic outcomes. Other entities can handle richer payment dynamics (by tranche, by stream).
- Complex expenditure operations resulting from contract interactions are harder to understand for users, resulting in some attack surface.
- Automated: By making proposals machine-readable, we can automate the whole governance flow, since proposals adhere to the same governance spec no matter which tool they go through.
These are some possible improvements. At this stage, they would likely just add complexity with no upside.
- Specifying which are the contracts that the DAO governs. This could be a good way to establish which are the official contracts that the DAO commits to govern and safeguard.
- Allowing spending ERC721.