add eip-zodiac

[auryn-macmillan]

This PR adds an EIP for Zodiac, a standard for composable DAO tooling.

[eth-bot]

All tests passed; auto-merging...

(pass) eip-5005.md

classification
updateEIP

• passed!

[wschwab]

Very excited to see this! I've made some suggestions cleaning up the header, and made a couple of other points.

[lightclient]

I get the sense that this shouldn't be an EIP. Can you elaborate on the benefit of standardizing this ERC would be vs. developing the philosophy as a standalone thing? Also, you will need to remove the external links to Zodiac. EIPs should be able to stand on their own.

[auryn-macmillan]

Can you elaborate on the benefit of standardizing this ERC would be vs. developing the philosophy as a standalone thing?

ERCs are a Schelling point in the Ethereum ecosystem. If you're building something that benefits from composibility (most things in web3), then it makes sense to build it in such a way that it conforms to whatever most relevant standards are. So the motivation for standardizing this as an ERC really just to help solidify it as a credibly neutral standard in the Ethereum ecosystem. It feels much more likely to be widely used, and thus beneficial to the ecosystem, as an ERC than as a standalone standard.

Also, you will need to remove the external links to Zodiac. EIPs should be able to stand on their own.

No worries, will remove those links.

[MicahZoltu]

The litmus test I usually use for deciding whether something should be a standard or not is whether or not there is a many-to-many relationship between components of a system. Twitter's API doesn't need to be standardized because there is a single provider and many clients (one to many). Browsers on the other hand need standards because there are many web pages and many browsers (many to many).

What are the components on each side of a particular communication channel in the proposed system?

[auryn-macmillan]

The litmus test I usually use for deciding whether something should be a standard or not is whether or not there is a many-to-many relationship between components of a system.

This seems like a reasonable litmus test.

What are the components on each side of a particular communication channel in the proposed system?

IAvatar.sol is the core interface. It was prototyped on the Gnosis Safe's ModuleManager.sol, simply because that was the natural choice being the most widely used contract wallet. That said, it was a conscious decision to label it Avatar, rather than something explicitly Gnosis Safe related, because we wanted to make it explicit that this is an interface we fully expect other programmable accounts to leverage.

On the other side of the interface, modules can be any contract that can call execTransactionFromModule() or execTransactionFromModuleReturnData() on an Avatar. This could be just about any of the existing DAO frameworks or any of a number of modules that have been written to comply with the standard already.

The intent is definitely for it to be a many-to-many relationship between components, where each component should be mostly agnostic the specific implementation of the other components.

[SamWilsn]

Suggested change

	title: Zodiac
	title: Zodiac Avatar Accounts and Extensions

I feel like the core contribution of this EIP is Avatar accounts, and I think "extensions" covers modifiers, modules, and guards well enough?

[auryn-macmillan]

I'd prefer not to introduce additional terminology that isn't used elsewhere.

If we were to add anything to the title, I'd be inclined to go with something like "Zodiac Avatar Accounts and Mods".

Thoughts?

[SamWilsn]

Hm, how about just Zodiac Avatar Accounts?

[auryn-macmillan]

That seems reasonable as well.
One thing I have been conscious of lately is that this pattern is useful beyond just DAOs. So I wonder if it would be worthwhile changing the rest of the language to be less opinionated about how it is used / what it is used for?

[auryn-macmillan]

Added this PR to generalize the language. This feels like the right direction to me.
https://github.com/auryn-macmillan/EIP-Zodiac/pull/1

[SamWilsn]

I think this is good to merge as a draft. I think the new title and generalization PR is a good idea, but that can get merged after.

If the bot doesn't merge this, try opening and closing the PR.