EIP Improvement Process #129
Transcript
Call summary
Summary
EIPIP #129 debated a proposed strict rule for changing an EIP's author list, agreed that authors must not self-allocate EIP numbers, and worked through why the Glamsterdam meta EIP is stuck in Draft. Editors decided to document number allocation as an editor responsibility (in contributing.md, the template, and the EIPW message), leaned against a hard author-removal rule in favor of an objection period, and agreed to raise the meta EIP's out-of-status references and an undocumented mascot addition at ACD. The EIP working-group Rust rewrite is at risk of being dropped without more editor buy-in.
Decisions
- •Authors must not self-allocate EIP numbers; the editor or associate editor assigns them. Documented in contributing.md, the PR template, and the EIPW message.
- •No final decision on the strict author-change rule (CFI 409); editors lean against a hard removal rule and are exploring an objection/waiting period plus a 'disputed' label. Kept open for editor input (deadline Aug 15).
- •Glamsterdam meta EIP-7773: editors will follow the process and not take over the meta; the mascot addition and out-of-status referenced EIPs (8066, 8254) block moving it to Review, and Pooja will raise this at ACD.
- •Re-enabled the bot check for Final EIPs referencing non-final EIPs, to be verified on the ERC repository.
- •EIP-numbering automation is dropped from the agenda until someone takes ownership after the protocol-support-team transition.
Highlights
- •Strict author-change rule (CFI 409): Yam proposed authors may only append to the author list, not remove others, to stop one author from removing the rest and taking the EIP number hostage. Sam noted the current all-equal model is objective and that a removed author can just fork under a new number. No final decision; leaning toward an objection/waiting period plus a 'disputed' label.
- •Authors must not self-allocate EIP numbers; number assignment is an editor/associate-editor responsibility. Triggered by EIP-8363 self-allocating a number already taken (8361).
- •Glamsterdam meta EIP-7773 is blocked from Review because referenced EIPs are not in Review, an undocumented 'mascot' PR (EIP-8066) was merged, and EIP-8254 was added directly as DFI without being PFI first. Editors will not take over the meta; Pooja to raise it at ACD.
- •An agents.md file (read by Copilot/Codex/ChatGPT, not Claude) may be added instructing LLMs not to allocate EIP numbers, since new proposal text is increasingly AI-written.
- •The bot check flagging Final EIPs that reference non-final EIPs was re-enabled; needs verifying on the ERC repo.
- •The EIP working-group Rust rendering rewrite may be dropped unless more editors help review it; EIP-numbering automation is stalled with no owner and was dropped from the agenda.
- •An EIP Hub (4-day space: process basics, protocol/governance, application standards, enterprise) is planned for Devcon India, seeking programming partners.
Action Items
- •Sam - Fix the Unicode-standards PR merge conflict (or force merge); add the no-self-allocation note to the template and EIPW message; verify the re-enabled reference-status bot check
- •Yam - Open a draft agents.md PR (do not allocate EIP numbers) for the EIP and ERC repos; help review the working-group Rust rewrite
- •Pooja - Raise the Glamsterdam meta issues (undocumented mascot, out-of-status EIPs 8066/8254) at ACD; continue the strict-author-change discussion on Discord
- •Editors - Approve the contributing.md PR (two approvals) and the associate-editor PR; give input on CFI 409 by Aug 15
Key decisions
Authors must not self-allocate EIP numbers; the editor or associate editor assigns the number. This is documented in contributing.md, the PR template, and the EIPW message.
EIP-8363's author self-allocated 8361, a number already assigned to another EIP; protecting number allocation avoids ecosystem confusion.No final decision on the strict author-change rule (CFI 409); editors lean against a hard author-removal rule and are exploring an objection/waiting period with a 'disputed' label, kept open for editor input until the Aug 15 deadline.
What is really being protected is ownership of the EIP number; a removed author can otherwise fork under a new number.Editors will follow the process rather than take over the Glamsterdam meta EIP-7773; an undocumented mascot PR and out-of-status referenced EIPs (8066, 8254) block moving it to Review, and this will be raised at ACD.
The ACD (fork) process and the EIP editorial process are independent; adding EIPs in the wrong status stalls the meta.Re-enabled the bot check that flags Final EIPs referencing non-final EIPs; its behavior on the ERC repository still needs verifying.
EIP-numbering automation is dropped from the agenda until someone takes ownership following the protocol-support-team transition.
EIPs discussed
AI Disclaimer: Some content or metadata on EIPsInsight may be AI-inferred or automatically compiled. If you find any discrepancy, please contact us at dev@avarch.org.