Summary

  • A July 2026 working-group draft describes principles, not final APNIC SIG Guidelines wording, across twelve linked subjects including terms, charters, eligibility, removal, dissolution, working groups, conduct and appeals.
  • Some principles are already specific—the term proposal would cap service at four terms or eight years across all SIGs—but recorded questions show that retrospective service, vacancies, existing charters and appeal scope still need transition rules.
  • Before any consensus call, APNIC should publish one operative redline and a twelve-row decision table showing the old rule, proposed rule, effective date, affected roles, evidence threshold, decision-maker, notice, response, remedy, appeal and review path.

Consensus needs an object

The strongest sentence in the 13 July draft is also its boundary. The working group says it is taking a principles-first approach and is not yet presenting final Guidelines wording or detailed implementation. Feedback would be used to prepare a consensus-call draft ahead of a planned call at APNIC 62.

That is a sensible drafting sequence. It keeps the group from polishing clauses before it agrees on the problem. But the sequence cannot end at principles. Support clearer removal rules is an orientation; it is not a rule. A participant cannot know whether to support it until the record says who may begin removal, what evidence is sufficient, who decides, what notice is required and what happens while an appeal is pending.

The decision object therefore has to be a text, not a theme. It should show each change against the current Guidelines and reveal the dependencies among changes. Otherwise a room may express support for twelve attractive goals while different participants imagine twelve different implementations.

The twelve items are a dependency graph

The draft groups its recommendations under term limits, charter governance, formation and charter endorsement, dissolution, chair responsibilities, an initial co-chair vacancy, eligibility, chair and co-chair removal, working-group governance, Birds of a Feather alignment, application of the Code of Conduct and appeals.

That list is useful because it exposes the surface area. It is dangerous only if treated as twelve isolated checkboxes. Eligibility determines who can stand; vacancy rules determine who can temporarily exercise the role; term limits determine when continuity must end. Charter endorsement defines a SIG's purpose; review and dissolution determine what happens when that purpose has expired. Removal rules depend on conduct, evidence and a remedy path. Appeals depend on knowing which decisions are appealable and whether the first decision remains effective during review.

A final package may still ask for one consensus. If so, it must make those links visible. A person who supports term limits but objects to retrospective counting needs a way to identify that disagreement. So does someone who supports dissolution for inactivity but rejects an undefined threshold controlled by an unnamed actor.

A concrete term cap still has edge cases

The term-limit proposal is unusually exact: four terms, or eight years, in chair and co-chair roles across all SIGs. Service of less than one year, including as acting chair, would not count. For continuity, chairs already at the four-term maximum could serve one additional term through 2027.

Specificity helps, yet the recorded response of 16 July identifies the decisions still inside those sentences. Does service before the new rule count retrospectively? Does an initial one-year co-chair appointment count? How do existing SIG charters move into the new common framework? The answers change who is eligible immediately and who gets a transition.

Those are not drafting trivia. A transition clause distributes office. It can favour incumbents, accelerate turnover or create an unexpected vacancy. The public text should therefore include dated examples: a chair with four completed terms, a co-chair with an initial one-year term, an acting chair who served eleven months, and a person who moved among SIGs. Readers should be able to reproduce the eligibility result from the rule.

Dissolution, removal and appeal form one chain

The draft sensibly asks for objective and clear criteria, notice, an opportunity to respond or remediate, written reasons, community involvement and appeal rights around dissolution and leadership removal. These protections matter most when they are connected.

An objective criterion is incomplete without a measurement period and evidence owner. Inactivity, for example, might mean no meeting, no mailing-list discussion, no work product or no progress against a charter. Those are different tests. Notice without a response window is ceremonial. A response without a named decision-maker has no destination. An appeal without scope, timing, standard of review or interim effect may arrive after the office or SIG has already disappeared.

The same July response asks whether a formal appeal is required for every Guidelines decision or only for dissolution and removal, and whether broader appeal references should be removed. That is exactly the kind of fork a consensus call must resolve. It should not be left to later custom, when the first dispute will reward whichever actor controls the ambiguity.

Charters define the unit being governed

The draft also reaches formation, charter endorsement, periodic review, working groups and BoFs. Here the comparison with RFC 2418 is informative but not controlling. The IETF document shows how a charter can state objectives, administrative details, milestones and deliverables; it also makes rechartering, termination, leadership and dispute review explicit. APNIC is not required to copy that model. The comparison simply demonstrates that a charter exists is weaker than a charter names the work, boundary, outputs, review date and termination condition.

This matters for institutional legitimacy. A SIG can acquire a durable name and leadership structure even when its work has changed. A BoF can test an idea without becoming a permanent mandate. A working group can deliver a bounded task without inheriting the authority of the entire SIG. Exact charters keep these containers from silently expanding.

Publish the decision table before the call

The June working-group notice set an ambitious sequence: a first draft at the end of June, final recommendations in early August and community input before the conference, with no new feedback during the consensus process. If that sequence is retained, the pre-call text carries extra weight. Participants need enough time to find conflicts before the moment when feedback closes.

The minimum useful package is one redline plus a twelve-row table. Each row should state: current clause; proposed clause; reason; effective date; affected offices or groups; transition cases; initiating actor; evidence and criteria; decision-maker; notice and response; remedy; appeal; review date; and dependencies on other rows. An accompanying change log should say which July questions were accepted, rejected or deferred and why.

This is not a demand for unanimity or procedural bulk. It is the price of knowing what rough consensus concerns. RFC 2418's account of rough consensus warns against equating a decision with message volume or a simple vote. Heng Lu's Multi-Stakeholder Mirage supplies the complementary limit: participation is evidence, expertise, warning and objection, not authority over absent parties. Exact text lets objections attach to consequences rather than personalities.

The reviewed external record does not establish a final consensus-call draft, adoption, an effective date or implementation. It establishes a serious diagnosis and several unresolved forks. APNIC can turn that work into a legitimate decision by making the object inspectable before asking the community to stand behind it.

Sources

  • Public mailing-list archive, July principles-first SIG Guidelines recommendations
  • Public mailing-list archive, response questions on transition, eligibility and appeals
  • Public mailing-list archive, working-group establishment and schedule
  • IETF, RFC 2418: IETF Working Group Guidelines and Procedures
  • Heng Lu, The Multi-Stakeholder Mirage