PR #96 introduces an RFC process whose acceptance rule depends on "maintainers": at least two must approve, none may request changes. For now we are treating the people listed in .github/CODEOWNERS as the maintainer set, which is pragmatic but informal:
- There is no public list of who counts as a maintainer.
- There is no documented process for adding or removing maintainers.
- The CODEOWNERS file is uniform across paths today, so it does not reflect actual ownership specialization.
- External contributors have no way to know who can vote on an RFC, who decides FCP, or what the path to becoming a maintainer looks like.
Proposed follow-up work
- Add
MAINTAINERS.md listing maintainers, their areas, and contact info.
- Add
GOVERNANCE.md covering: how decisions get made, how new maintainers are added, how maintainers can step down or be removed, and the relationship between maintainer status and CODEOWNERS entries.
- Once
MAINTAINERS.md exists, update rfcs/README.md to point at it as the source of truth for "maintainer," replacing the current "use CODEOWNERS as the maintainer set" stopgap.
Not urgent. The current rule works for a small AWS-internal set of maintainers, but it will not scale once external contributors start submitting RFCs or want a path to become maintainers themselves.
References
- rust-lang/rfcs has a model worth borrowing from.
- ASF projects (Cassandra) have a more formal voting structure that may be overkill at this stage.
PR #96 introduces an RFC process whose acceptance rule depends on "maintainers": at least two must approve, none may request changes. For now we are treating the people listed in
.github/CODEOWNERSas the maintainer set, which is pragmatic but informal:Proposed follow-up work
MAINTAINERS.mdlisting maintainers, their areas, and contact info.GOVERNANCE.mdcovering: how decisions get made, how new maintainers are added, how maintainers can step down or be removed, and the relationship between maintainer status and CODEOWNERS entries.MAINTAINERS.mdexists, updaterfcs/README.mdto point at it as the source of truth for "maintainer," replacing the current "use CODEOWNERS as the maintainer set" stopgap.Not urgent. The current rule works for a small AWS-internal set of maintainers, but it will not scale once external contributors start submitting RFCs or want a path to become maintainers themselves.
References