Summary

  • RIPE-861 recommends that volunteers send nominations directly to working-group chairs whose terms are not ending, rather than post them publicly, partly to encourage less extrovert candidates.
  • The twelve active working-group selection pages do not expose one common handoff. Some promise to announce every candidate after private intake, some require public self-nomination, and several state a call or deadline without specifying the submission channel on the checked page.
  • Privacy and accountability can coexist if each selection produces a nomination-handoff receipt: counts and state changes can be public while identities and correspondence remain private until the person consents to candidacy.

The first message may now be private. The authority it eventually creates is not.

RIPE-861, published in final form on 12 June 2026, recommends that volunteer nominations go directly to those working-group chairs whose terms are not ending, not publicly to the mailing list. The document gives an inclusion reason: the route is likely to encourage more nominations, especially from people who are less extrovert. It is a sensible diagnosis of a real participation cost. Asking publicly for a position in front of an established technical community is not socially neutral.

The recommendation also changes the evidence path. A public self-nomination is visible at the moment it occurs. A private expression of interest becomes visible only when somebody later turns it into a candidate. The crucial accountability question is therefore not whether the first email should be published. It is whether the community can verify that every eligible person who consented to stand crossed the same boundary into the public candidate list.

No checked source shows that a volunteer was omitted, favoured or exposed. This is not an allegation about a selection. It is a design problem created by a useful privacy choice.

A best practice, not a central election code

The wording matters. RIPE-861 says working groups follow a number of best practices. Alongside direct, non-public nominations, it recommends terms of no more than three years, consideration of a maximum number of terms and a process that permits online participation. It does not replace every local procedure with one ballot, one term or one appeal route.

Its predecessor, RIPE-692, placed responsibility on each working group to develop, maintain and implement its own chair selection and removal procedure. The official document diff makes the provenance visible: the common private-intake recommendation belongs to the 2026 update, not to an old uniform rule.

That decentralisation is not an accident to be eliminated. The 2018 RIPE Accountability Task Force report said that having each working group maintain its own replacement procedure was consistent with a bottom-up approach. Its Recommendation 8 also said inconsistencies should be aligned to provide more clarity. Alignment for clarity is not the same as identical voting machinery. A shared evidence vocabulary can coexist with different ways of choosing.

The consultation trail is itself a useful model. RIPE records the first draft on 23 December 2025, the second draft on 27 March 2026, the consensus declaration on 29 May and final publication. Readers can tell which text acquired authority and when. A chair selection needs a smaller version of the same chain.

Twelve pages, several handoffs

At the research cutoff, RIPE's active-working-groups index exposed twelve working groups with chair-selection pages. Reading them together does not produce twelve failures. It produces a map of different state transitions.

Working group Intake stated on the checked page Public handoff and decision stated on the page
Address Policy Candidate introduces themself on the mailing list Session consensus, with remote and physical participation; physical-only STV ballot if consensus fails
Connect Mailing list or direct to chairs All candidates announced; list discussion; chairs declare decision
Cooperation Mailing list or direct to chairs All candidates announced; list discussion; chairs declare decision
Database Direct to chairs All candidates announced; list discussion; non-subject chairs judge; RIPE Chair appeal
DNS Invitation for applications; channel not stated on the checked page Mailing-list consensus; subject excluded from judgement; RIPE Chair appeal
Internet of Things Channel not stated on the checked page Working-group consensus judged by remaining chair; RIPE Chair resolves uncovered cases
IPv6 List call and closing date; submission channel not stated on the checked page Meeting selection; physical participants; ballot if unclear; candidate may be remote
MAT Interest made known through mailing list List discussion; chairs declare consensus
Open Source Direct, off-list contact with remaining chairs Simultaneous candidate announcement, feedback, preferred candidate, final WG approval and RIPE Chair Team confirmation
Routing Nominations accepted until the day before; channel not stated on the checked page Consensus of those present at the meeting
Security List call and closing date; submission channel not stated on the checked page Meeting selection; physical participants; ballot if unclear; candidate may be remote
RIPE NCC Services Two-week interest period; channel not stated on the checked page Two-week discussion; remaining chairs call consensus; normal appeal process

Some procedures already contain the join that RIPE-861 needs. Connect and Cooperation allow interest by two routes and then tell the chairs to ensure that all candidates are announced. Database moves direct interest into an all-candidate announcement and excludes a co-chair who is the subject from the consensus judgement. Open Source is more explicit still: volunteers contact remaining chairs off-list, all candidates are announced simultaneously, feedback is assessed, the working group gives final approval, and the RIPE Chair Team confirms that the process was followed.

Other procedures start in public. Address Policy asks volunteers to introduce themselves as candidates on the mailing list. MAT asks interested parties to use the list. Those pages do not need the same private-to-public conversion at intake, although they still need a durable result and process version.

Several pages describe notice, deadlines or the decision forum without stating where an application is sent. That sentence must be read narrowly. It means the channel is not in the checked page. It does not prove that no channel exists, that a current call lacks instructions or that working-group practice is defective. A public call can supply operational detail that a standing procedure omits.

Interest is not candidacy

The handoff becomes clearer when the nouns are separated.

An expression of interest can be tentative. A person may ask about workload, eligibility or travel before deciding. A nomination can be made by the person or by someone else. Eligibility applies the published requirements. Consent determines whether the person's name may enter the candidate slate. Public candidacy allows scrutiny and support. Support informs a consensus judgement or ballot. The decision produces an appointee. Confirmation or appeal may then test whether the procedure was followed.

Collapsing those states creates two opposite errors. Publishing every private inquiry defeats the inclusion purpose. Treating every inquiry as a hidden candidate makes a complete slate impossible to attest. The record needs to show transitions without turning uncertainty into exposure.

A simple state model is enough:

interest received → consent checked → eligibility checked → public candidate → discussion → decision → confirmation or appeal

A withdrawal can occur before or after consent. An eligibility problem can be corrected before the deadline. A late expression can be handled under a published rule. Each branch should have a reason class and timestamp. The public does not need the private facts behind every branch; it does need assurance that the rule was applied to the full intake set.

The receipt protects both the quiet volunteer and the electorate

The smallest durable object is a nomination-handoff receipt. It is not a dossier and not a public inbox. It records the working group, procedure URL and version or hash; call time and deadline; permitted channels; authorised recipients; and whether a recipient is also a candidate or decision-maker.

It then records privacy-safe counts: interests received, people who consented, withdrawals, late submissions, eligibility outcomes and compact reason classes. At publication, it records the candidate-list timestamp and an integrity assertion that every eligible, consenting candidate received through the declared channels was included.

The decision side records the discussion forum, open and close times, online or physical participation, method of assessing support, recusal, result, confirmation and appeal state. Term dates and later corrections complete the chain.

The public surface might say: five expressions received; four people consented; one withdrew before disclosure; all four were eligible; four candidates were published together at time T; the receiving chair was not a candidate; two other chairs judged consensus; no appeal was filed. The restricted record can preserve the minimum correspondence needed for correction under a retention rule. Before consent, no name needs to appear publicly.

This receipt is different from the evidence proposed for private policy drafting. A policy thread must return changed text, reasons and alternatives to public review. A nomination handoff must prove completeness, consent and eligibility before public evaluation begins. Both problems cross a public-private boundary, but they carry different objects and different privacy risks.

Who controls each transition

The candidate controls consent and can withdraw. Continuing chairs may control intake and publication of the slate. The working group supplies discussion, support or consensus. Serving chairs often judge that consensus, sometimes with an explicit exclusion for the person concerned. RIPE NCC staff count specified ballots. The RIPE Chair or RIPE Chair Team appears in some procedures as an appeal body, final decision-maker, overseer or confirmer.

Those roles should not be blurred into a claim that RIPE NCC selects every working-group chair. The RIPE community and the RIPE NCC are institutionally distinct. The secretariat hosts pages, lists, meetings and records; local communities retain the selection authority described in their procedures.

Role separation also matters inside one chair team. The person who receives a private message may later publish the candidate list and judge support. That combination is not automatically improper. It does make a recusal and completeness field useful, especially if the recipient becomes a candidate or has to decide whether an interest satisfied requirements.

What the evidence does not show

The twelve pages are current published rules, not a dataset of completed elections. They do not tell us how many people considered standing, whether private intake increased diversity, whether every call used the standing text or whether an appeal occurred. They do not support a league table of working groups.

RIPE-861's phrase about less extrovert candidates is a stated expectation, not a measured outcome in the checked sources. The benefit should be tested through privacy-safe aggregate trends over several cycles, not through demands that volunteers disclose personality or motivation.

Nor does clarity require one electoral system. Address Policy's fallback STV ballot, mailing-list consensus in DNS, support assessment in Open Source and meeting consensus in Routing represent different local choices. A common receipt can reveal which choice applied without deciding that one is universally correct.

The verifiable question is narrower: when the first approach is private, can the community reconstruct the authorised intake channels, consent and eligibility transitions, complete public slate, recusal, decision and review state? If the answer is yes, inclusion has not displaced accountability. It has changed the record accountability requires.

Sources