How do we address editors being overworked with a better governance method and what does it look like?

There’s been some recent discussions on the Ethereum Cat Herders discord server about what is a more scalable method for governing and enforcing the EIP process. Right now there’s two majors concerns that seem like they need to be addressed:

  1. Editors are having to help juggle and steward many different EIPs through the process
  2. The quality of EIPs is largely dependent upon the author because EIP editors can’t be experts in all different topics that are discussed

For this reason, it seems like we’re encountering some boundaries of the current model of governance where we can efficiently ship standards that lead to rough consensus and running, interoperable code. I’d like to start the discussion of what might a new process look like to scale governance of the EIP process.

5 Likes
  • Auto-merge EIPs as draft once they pass validation, including issuing an EIP number.
  • Split off ERCs.
  • Sell unused EIP #s to fund EIP editors
2 Likes

@kdenhartog, thank you for bringing this up, I hope that we can get some helpful comments and ideas going in this thread.

I don’t want to weigh in too strongly before understanding the problem better, and also understanding how other communities (IETF, W3C, etc.) deal with stewarding proposals.

What do you think are the key causes of these challenges?

The editors do seem to lack the operational support and ability to recruit help vs. other groups who are a part of the larger process. And it also seems as if there has not been enough editors to handle the workload, particularly as there are EIPs and ERCs being submitted from so many different topics / technical areas.

For me the key challenges I think editors face are having to wear multiple hats as domain expert for specific classes of EIPs, acting as enforcement of the process, acting as pseudo-governance of the process, and also actual editorial review (e.g. missed spelled a work etc).

To me each of these tasks are big enough these days that they could be separated out at this point. Especially for volunteers who often time have secondary work outside of this.

Even just splitting off the need to be a domain expert and editorial review from the enforcement/governance of process would be helpful IMO. This could be done via the usage of a process to form working groups (WG) is what I’m thinking and as a part of this “chairs” and “editors” of the WG can exist to help steward specific proposals through the EIP process.

At face value this sounds simple, but there’s likely a lot of things that would still need to be addressed if there’s buy in to go in this direction.

For example,

  1. how do we know if there’s enough interest to establish a WG?
  2. How do we go about selecting people to run these WG?
  3. What expertise do they need to do so?
  4. What do we need to do to bring the appropriate people into these WGs to better establish consensus?

Additionally, I think it would be useful to consider how this might affect this other thread @xinbenlv has proposed here: Discussion of Criteria for advancing EIP status: A Straw-man Proposal - #8 by kdenhartog

Just some thoughts to keep the discussion moving here

3 Likes

Why not help govern the process by making two lanes of proposals?

People with successful EIP
Lane one: EIPs that have had previous EIP Authors vouching for the new author or they themselves are an author of the EIP in waiting.

People without successful EIP
Lane Two: pending EIP authors must wait a min. of x amount of days before acceptance such that their is sufficient time for the res publica to be informed about this. Additionally have them submit to some mandatory event or process. Eg attending one ACD (not necessarily to speak, etc) or even better having them implement some portion of their proposal (can be a toy implementation etc).

The idea being that the process itself filters out less serious proposals by imposing a min. fixed cost of time on their part that is commensurate with the scope of their EIP.

I think having previous author’s involvement for providing a reference will bring additional benefits as well.

Edit: Also this would mean making a list of EIP Process Status, i.e.:

Draft / Pre Qualify
Pending Consideration,
Accepted but not Implemented,
Implemented,
Obsolete

These would be editor status, a sort of EIP meta categories

2 Likes

@xinbenlv and I were talking about this discussion yesterday. We were curious if anyone has considered how to move the EIP process into an area directorate at the IETF? This seems like it would allow us to adopt their governance and processes so that the community can remain focused on publishing, editing, and standardizing specs.

Thoughts?

1 Like

I would be in favor of that idea if anybody has the requisite connections. My only concern would be that the EIPs would have to be renumbered if they are to be converted to RFCs.

It was a great chat with @kdenhartog. We shared a lot of thoughts and views. Given EIP has earned a lot adoption and respect while the latest activities on drafting and reviewing EIPs seems encounter a few challenges, it’s good to start discussion about how we could borrow the wisdoms from various mature standard body. I think we can try to learn from IETF/W3C and a few others. Whether to “move into” is an open questions and I don’t have a good answer yet. I recognize the difference between Web3 space vs traditional web space, and the role of EIP for Ethereum rather than IETF/W3C for general web. Also the legitimacy of a standard body mainly comes from the adopting stakeholders. And the main stakeholders (browsers, email client, websites, etc) of IETF/W3C and other standard body are quite different from EIP’s adopting stakeholders (Web3 wallet, Smart Contract etc). Therefore we probably want to sort out a better way to identify stakeholders and invite them.

Let’s chat a bit more.

@kdenhartog and I will start some Zoom / Discord voice chat soon. And if anyone on this thread of FEM are interested in participating, please let us know by commenting or DM us in FEM. We are happy to add you to schedule together.

Since my previous comment seems to be too radical, may I suggest a modest one: have you tried paying them?

Additionally if you trace the lineage of the EIP process, it originates from Python/Gentoo — they are narrow in scope wrt the project itself. IETFs only interesting aspect is its trust structure for its own incorporation, which isn’t relevant considering EIPs are in the public domain (I think this is potentially more worth reconsidering).

This has in fact been considered, and would certainly attract more editors. The question that remains is: Where the funds would come from?

I didn’t think this was radical enough actually. Sorry I didn’t respond to this originally. I did spend a bit of time thinking about it. One of the reasons I wasn’t a fan of that suggestion was it creates a larger amount of bureaucracy for authors without really helping take the harder more contentious decisions off the editors plate in my opinion.

My concern is that the path we’re taking is creating a committee of benevolent dictators unintentionally that require consensus to make decisions. I see that as the primary problem that we need to be splitting up here is that all final decision making power shouldn’t be centralized to a small core group of editors. Instead, it’s my opinion that the decisions should be split out to prevent issues from arising as the work is scaled and limit controversy centralizing around a small number of important decision makers made.

If you look at other standards processes there’s a reason the governance processes always seem so complex. It’s because the complexity emerges from trying to establish checks and balances to each roles power within the process.

This has always been the case. I agree that this could use fixing.

I don’t care about being controversial. I either convince everyone and it becomes the consensus, or am outvoted and the rules changed.

My actual concern is that a small group of decision-makers, such as we have, simply doesn’t have enough bandwidth to keep up.

Hi everyone!
I want to ask this question - how do successful self-governing communities work?
Why don’t we take some time to study in history, the best self-governing examples, and take it from there?
Also - are there any editors who are open to talking about this? Would be great to hear from the editors themselves what’s going on and what their thoughts on making the process more decentralized.
It looks like there is a proposal for how to “become” an EIP editor- but looks like you just have to make a PR and then get approved by the list of editors.
But how do you get approved by the editors? LOL

I am an editor, and also one of the authors of the EIP you linked :slight_smile:

The process is as simple as the EIP describes. If enough editors say yes, then the PR gets merged and you are an editor. There are technically no requirements to become an editor (which is why I am still annoyed that @xinbenlv hasn’t been allowed to become an editor yet).

1 Like

Thanks for the kind words @Pandapip1 . I am here to contribute to open standards, in any way our community needs. Appreciate you for your support.