# DAO Split design (prev. known as Ragequit Threshold)
Split can be signaled on proposals. To successfully split, the following conditions must be met:
1. You voted opposite the proposal's outcome, e.g.:
1. You voted against and the proposal passed.
2. You voted for, quorum was reached, and the proposal was defeated.
2. You signaled to split during voting period or during timelock delay.
3. Enough Nouns signaled to split on this proposal to pass the split threshold; Nouns may submit a split signal up until the end of a proposal's queuing period (before execution if successful) or up until a proposal is defeated.
* As of March 2023 20% of total supply seems like the lowest threshold we can safely set.
Assuming these conditions are met, a DAO split can be executed at the end of the proposal lifecycle:
* **Successful proposals**: executing the proposal first executes the split; alternatively anyone can execute the split first, then execute the proposal. The split can be executed at any time after the queuing period, up until the proposal expires.
* **Defeated proposals**: anyone can execute the split at any time after the proposal's defeat, and until defeat time + proposal expiration time (currently 14 days).
## Where Nouns Go
What happens with Nouns that signal a DAO split?
1. Upong signaling these Nouns are transfered from the owner's account to an escrow contract.
2. The escrow contract delegates the escrowed voting power to an account the owner chooses, so they can continue voting until the split is settled.
3. Once the proposal the split was signaled on reaches an end state, Nouns are taken out of the proxy:
1. Successful split: Nouns are sent to the Nouns DAO treasury.
2. No split: owners can withdraw their Nouns.
## Outcome: a new DAO
A successful split creates a new NFT DAO; including a new NFT, auction house, DAO logic, DAO treasury. The fair share of all split Nouns is sent to the new DAO's treasury. For example, if 20% of all Nouns split, then 20% of Nouns DAO treasury assets are sent to the new DAO.
New DAO starting configuration:
* **Token**: each DAO split account is minted the same number of tokens they split with.
* **Art**: new token reuses the existing Nouns on-chain art, and the new DAO can vote to use different art.
* **Founder reward**: new token has no founder rewards; it is an upgradable contract so that new DAO members can adjust it to their needs; new token also supports locking furthur upgrades.
* **Token IDs and art will be the same** as the original Nouns split accounts left behind.
* **Auction House**: a new auction house is deployed paused; new DAO members can vote to configure and start it.
* a new copy of the latest Nouns DAO logic + timelock (treasury) contracts is deployed for new DAO; all logic and configurations are the same unless stated otherwise, e.g. it has the same logic for proposals, voting, DAO split, etc.
* the DAO split threshold is set to a single token, to minimize the risk of a majority attack in the new DAO.
* **Vetoer**: The new DAO doesn't have a set Vetoer; DAO members can vote to set a new Vetoer.
* There will be **a configurable no-proposals period at first**, to mitigate the risk of malicious proposals in the new DAO's early days.
* **UI**: new DAO should be able to launch their own UI, or reuse an existing DAO UI without any coding skills.
* Specific design TBD; it can be a better-packaged version of nouns.wtf, and it can also be Nouns Builder UI automatically supporting auctions and governance for any split-created DAO.
## Outcome: asset split
Nouns DAO can configure which tokens are distributed upon a DAO split, while other tokens remain entirely with Nouns DAO.
The tokens configured to be distributed are sent pro rata, so for example if 20% of Nouns split, then 20% of the configured assets are sent to new DAO.
#### Rounding in favor of Nouns DAO
In the case of NFTs it can get tricky as they don't have decimals. The current solution is to round down in favor for Nouns DAO. For example, say 20% of Nouns split, and Nouns DAO holds 197 tokens of a certain NFT; that comes out to 39.4 NFTs, in which case we would send 39 NFTs to new DAO and the rest will remain with Nouns DAO.
#### Which NFTs will be sent?
We can randomize which IDs get sent to new DAO, using Chainlink VRF or new developments in on-chain randomization that leverage ETH2 RANDAO.
Another approach is to run a lottery of which DAO gets which NFTs. We can use random numbers to determine if a given NFT should stay with the original DAO or go to the new DAO. For example, say 30% of Nouns choose to split, we can generate a random number between 1 and 10; if the number is 1-3, it goes to the new DAO, and if it's 4-10 it goes to the original DAO.
### Why not just burn the Nouns left behind?
TL;DR: it gives the original DAO more options, including the option to burn them.
If the split is due to a malicious proposal (aka a 51% attack), where honest Nouners are escaping to the new DAO, most likely the Nouns left behind in the hands of the attacker won't be very valuable, so it seems fine to leave them in their hands.
If the split leaves honest Nouners in the original DAO, they remain valuable, and we then think it's good to give the DAO more optionality; they can choose to burn these Nouns, or put them on auction again, or grant them to strategic partners, or any other creative idea they may agree upon.
Important to note that these Nouns are excluded from DAO quorum and split threshold calculations; they are counted again if and when they are transferred from the treasury.
### How do we protect inactive voters?
An idea we like is "split delegation", where an account can say they will split if these other accounts or Nouns choose to split.
### What about a penalty?
We think it's best to have zero penalty in order to not leave anything behind for an attacker.
We can include an adjustable penalty parameter which can be changed via DAO proposals should the DAO decide to have it > 0.
### Who can signal to split -- Only Noun holder or also delegates?
By default it would only be the Noun holder. We don't think delegates should have this kind of power.
If needed, for hot wallet kind of convenience, we can design a separate delegation system for this, but that would also add some layer of complexity