Hiring OASIS to referee the EIP process

tags: proposal oasis

The Problem

• Ethereum is growing up fast. The EIP process is starting to become a powerful game.
• The standards process is infamously a political minefield. And for better and worse, very few current players in the EIP process have standards experience.
• Some of our standards are of lackluster quality yielding substantially different implementations.
• If CryptoTwitter is any indication, the Ethereum Magicians wish to proclaim themselves the official governmental body for EIPs. As respectfully as it can be said—when it comes to standards, the Magicians do not have the experience to navigate the trecherous waters that they travel in. The EIP process is too mission-critical to permit such risk.

Proposed Solution

• Several (including me) were considering having Ethereum join the IETF. But members of the IETF management itself thought it was a bad idea to join them, citing IETF would chain us into a slow-moving bureaucratic nightmare. So I went in search of alternatives.

• A battle-hardened ex-Tor colleague suggested OASIS, which operates oasis-open-projects.org. OASIS is a nonprofit standards organization with a pretty solid record. OASIS’s primary benefit over other standards organizations is that they are vastly more flexible and try to help you move forward rather than try to keep you out.

• I propose we establish an Ethereum project under OASIS. OASIS brings experienced yet flexible referees for pre-existing projects and a mature process for creating standards.

  By happenstance, OASIS recently started a framework for FOSS “Open Projects”. Per OASIS regulation, our Open Project would be governed by: (1) the Project Governing Board (PGB) made up of the Sponsoring organizations and appointed experts; (2) a Technical Steering Committee (TSC). The Technical Steering Committee would presumably consist of the Ethereum Magicians’ EIP-ring.

• It remains unclear whether OASIS and Ethereum will be a good match, and we should start small. I propose that we develop the standard for the JSON-RPC specification cited above as an Open Project within OASIS. If it goes well, then we can discuss expanding the number of EIPs developed at OASIS. If that goes well, then we can even consider replacing the entire EIP process.

• OASIS claims to be good at discovering unexpected orgs who would have commercial interests in Ethereum. But it’s unclear how valuable this would be to us.

• Update: The proposed founding PGB members of the Ethereum OASIS Open Project are: Ethereum Foundation, Ethereum Enterprise Alliance, and Consensys. Additionally, Nick Johnson (long-steward of the EIP process) is to be appointed as an individual board member. Candidates for later joining the board are: IBM, Microsoft.

The Upsides

• Foremost, we will have protections in place to guard the EIP process against outside corruption or encumbered by novice hands.
  • OASIS supplies a formal decision-making process that is time-tested and robust to abuse.
  • OASIS supplies independent referees (“Process Managers”).
  • Will inform us of best practices in standards creation, e.g., making specs generated by a grammar. We can decide if we want to incorporate them. This will probably reduce ambiguity in our standards and improve consistency among clients.
• Currently, the lead custodian of the EIP process is Nick Johnson. Outsourcing some of these duties to OASIS will free Nick to spend more time on Ethereum Name Service—a high priority project.
• Ethereum is starting to be seen as one of the responsible adults in the blockchain world. Working with OASIS will further solidify that image.
• OASIS will be more flexible with processes and structure than alternatives like IETF, ISO, or Linux Foundation.
• If we ever want it (and if we are successful, we probably will not least because having them build comepting Ethereum standards instead isn’t a good plan), OASIS-approved standards will more easily garner further approval by larger international standard organizations like ISO/IEC/ITU.

The Downsides

• Doing an OASIS Open Project requires annual dues of >= $25,000.

  • The annual due varies with the size of the organization (for EF it’s 1k, for Consensys 15k). This starting fee covers all core services. We can optionally set higher mandatory dues as gate-keeping. These additional funds become a generic pot of money for the PGB to hire consultants, sponsor events, donate to EFgrants/ETHGlobal, etc.
  • The exact fees from each member remain to be decided. Currently, Consensys is required to pay 15k per year. We probably can’t permit Consensys to be the primary payer, so that means EF will also donate 15k. So that gets us to 30k. We can then put EEA on it for >= 1k. Perhaps 5k would be reasonable.

• It opens the EIP process to some new attacks. Particularly, any organization that is willing to pay the annual Project Sponsor due holds a seat on the Ethereum Project Governing Board (PGB). This risk of “anyone can join” (and thus get a vote), can be mitigated by several factors:

  • Our PGB can require higher annual dues to ensure participation is limited to serious stakeholders. This filters out casual obstructionists unless they have money to donate.
  • Our PGB may grant full PGB seats to appointed experts (in addition to the Project Sponsors). These appointed experts can be used to offset any incoming disruptive Project Sponsors.
  • Our PGB may increase the voting threshold for “critical milestones”. This raises the bar for major changes.
  • OASIS Open Projects have a Code of Conduct that can include pretty much anything we want such as hurdles to reraising already decided issues (OASIS can show clauses that have previously worked). The OASIS Process Managers will then enforce said Code of Conduct.

• Formally incorporating our EIP process under OASIS can be seen as premature. But I still think it’s a good idea because: (1) The Magicians aren’t super familiar with how to do standards. (2) It future-proofs our EIP process against inevitable abuse.

• For every EIP using OASIS, we will have to make some small adjustments to the current EIP process. Fortunately, OASIS’s initial review of the EIP-1 process is that it’s largely okay, and any required changes to be OASIS-compliant would only change the outcome in edge cases.

• The Linux Foundation considers OASIS a competitor to their business model, as such using OASIS could cause mild friction. I’m personally not worried about this. As evidence, IBM is a tier-1 sponsor of OASIS, yet Linux Foundation still closely works with IBM.

The Alternatives

• We could simply leave the Magicians in custody of the EIP process.
• If we weren’t going to use OASIS, but also didn’t want to leave the Magicians to themselves, the most plausible alternative is using the Ethereum Enterprise Alliance. EEA claims to have an experienced standards team, and EEA would presumably be more focused on Ethereum, however, there might be cultural mismatch. One benefit of using EEA is that it might create more cooperation between Ethereum and EEA. If curious about EEA’s work in this area, see the Ethereum Enterprise Alliance spec and instructional video.
  • Update: EEA says that they’d love to help with our standards, but they don’t want to lead it.
    • Proposal Addendum: Give an EEA-person (probably Chaals Nevile) a spot on the Project Governing Board. This will give EEA’s expertise a role to play while still making the EIP largely autonomous.

If we decide to move forward, then we have…

Next Steps

1. The Ethereum Foundation drafts an Open Project Charter using the template provided by OASIS. The charter spells out what success looks like: the project’s purpose, scope, business benefits, as well as naming the initial members of the PGB and the TSC. I propose to make the troubled JSON-RPC spec the first EIP to go through the OASIS process.
2. Ethereum Foundation, with at least one other founding member, submits the charter and OASIS establishes the “Ethereum Open Project”. OASIS can then work with the PGB and TSC to document and fine-tune the day-to-day operating rules as needed to achieve the project’s goals.
3. Ethereum Foundation contributes the JSON-RPC spec to the project. The team works to improve the spec and raise it to the level of minimum OASIS standards.