Summary

  • AFRINIC’s Article 9.1 restriction in 2022 applied to the people choosing the candidate slate, not to Northern or Western candidates, voters or seats.
  • The Board appointed Bill Sangiwa, Noah Maina and Ali Hadji Mmadi as community members of NomCom and Dr Abdalla Omari as its Board representative on 19 March 2022.
  • AFRINIC’s Board index calls the appointment Resolution 202203.682, while the signed minutes give the same wording and appointees under Resolution 202203.681; the conflict cannot responsibly be edited away.
  • The rule created distance between selectors and their home-region contests, yet it also removed resident Northern and Western voices from a small, voting gate with power to convert nominations into candidates.
  • The public record shows the appointment, election timetable, screening outline and final slate, but not the volunteer denominator, the Board’s selector criteria, every nominee or individualized screening reasons.
  • The appropriate remedy is not an accusation of misconduct or a claim of public authority. It is a privacy-aware audit receipt that makes a private registry’s delegated gate traceable from volunteer call to slate, ballot and result.

The exclusion was at the gate

The most important word in the 2022 arrangement is selector. AFRINIC was preparing to fill Seat 1 for Northern Africa, Seat 2 for Western Africa and Seat 7, which was region-independent. Its 1 April call for nominations made residence in Northern Africa a condition for the first seat and residence in Western Africa a condition for the second. A qualified Northern or Western resident could therefore seek the relevant Board seat. The same resident could not join the Nomination Committee for the cycle in which that regional seat was open.

That is not a semantic nicety. Saying simply that “North and West were excluded” can invert the design. Candidates from the two regions were not merely permitted; regional residence was part of eligibility for Seats 1 and 2. The restriction operated one stage earlier, on membership of the committee that assessed nominations and determined which of them would become candidates on the public slate. It removed a class of possible selectors while preserving the same class’s access to candidacy.

AFRINIC’s published election process gives the restriction an intelligible conflict rationale. A person domiciled in a region whose Board seat was open for renewal could not serve on NomCom. A member of NomCom also could not stand in an election for which that committee was finding candidates. Taken together, those rules placed distance between a home-region contest and the people screening it. A Northern resident did not have to choose between influencing the Northern slate as a selector and entering that contest. The rule assigned the gate to people outside the affected regional domiciles and left those inside free to stand.

The cost was equally precise. The committee did not merely arrange a meeting or verify a list after others had made the substantive choices. It could prescribe criteria, assess nominations, request more information, conduct interviews and use Election Committee verification. It could qualify or disqualify nominees. In the process AFRINIC described, nominations were inputs; candidates were the people who emerged on NomCom’s final slate. The Board approved election guidelines, but NomCom finalized the candidates.

North and West therefore remained present in potential candidacy while being absent, by domicile, from the panel exercising the pre-ballot conversion power.

This is the narrow institutional problem worth examining. It is not a claim that the named appointees were partial, unqualified or beholden to anybody. The available record supplies no sound basis for such judgments. Nor is it a claim that a resident selector would necessarily have produced a better slate. Local proximity can contribute knowledge, but it can also create conflicts and network pressures. The design question is whether a geographic firewall, implemented through a small committee appointed entirely by the Board, had enough other safeguards to make the gate inspectable.

That question matters because exclusion and protection can be two descriptions of the same mechanism, depending on which part of the process one observes. Viewed from the candidacy side, Article 9.1 protected the option to stand and separated potential candidates from selector service. Viewed from the committee side, it meant that neither regional contest had a resident at the table. A serious account has to hold both truths at once. The rule cannot be criticized honestly by pretending candidates were barred, and it cannot be defended completely by pointing only to conflict avoidance.

A four-person appointment with two official numbers

On 19 March 2022, AFRINIC’s Board met online in a special meeting to appoint the Nomination Committee. The act named Bill Sangiwa, Noah Maina and Ali Hadji Mmadi as community representatives and Dr Abdalla Omari as the Board representative. The composition followed the published four-person structure: three community volunteers and one member drawn from the Board.

The official record contains a documentary problem that should survive every summary of the appointment. AFRINIC’s public Board index labels the NomCom act Resolution 202203.682. The signed minutes of the 19 March meeting reproduce the same operative wording and the same four appointees but label the act Resolution 202203.681. The index, meanwhile, assigns 202203.681 to a different decision concerning legal costs and a non-disclosure agreement. These are not rival descriptions that can safely be merged into a tidy citation. They are conflicting identifiers attached to an otherwise recognizable act.

The responsible way to identify it is by the stable facts: the date, the wording, the purpose and the four appointees. Both numbers should be disclosed. Choosing 202203.682 because it appears on the index would suppress what the signed minutes say. Choosing 202203.681 because it appears in the signed record would leave readers unable to reconcile the index. Inventing a corrected or composite number would replace evidence with editorial repair. The inconsistency may be clerical, but the public sources inspected here do not establish how it arose or which record AFRINIC intended to be definitive.

The minutes add useful procedural detail. They record Benjamin Eshun as proposer and Habib Youssef as seconder. They record seven votes in favour, none against and one abstention. That roll call is evidence of the Board’s decision on the appointment. It does not reveal how the underlying volunteer pool was assessed, what criteria separated the appointed people from other volunteers or what discussions preceded the motion. A recorded vote is one receipt in a chain, not the whole chain.

On 28 March, AFRINIC publicly announced the four members. That announcement confirmed the composition but did not recount the regional exclusion or the Board vote. A reader seeing only the announcement would know who had been appointed and could see the broad NomCom mandate. The reader would not learn from that page why Northern and Western residents were ineligible, how the appointment was adopted or that the two official Board records disagreed about the resolution number.

This difference between a public announcement and the underlying meeting instruments is instructive. Governance often becomes less legible as it moves from an internal decision to a reader-facing notice. Names and roles survive; conditions, alternatives and voting detail fall away. Nothing about that compression proves improper conduct. It does, however, show why an auditable gate requires documents to connect cleanly. A member should be able to move from announcement to appointment rule, from rule to volunteer criteria, and from appointment to vote without having to infer which identifier belongs to the act.

From appointment to slate

The published chronology shows how quickly the selector rule became operational. The Board acted on 19 March. AFRINIC announced the committee on 28 March. On 1 April, the call for nominations opened the three seats. It set a deadline of 23:59 UTC on 18 April and anticipated that the election would conclude at the annual general members’ meeting on 3 June. The final candidate slate was published on 10 May, with electronic voting planned to begin on 19 May.

Those dates describe an intended decision chain, not proof of a completed election. The source record within this article’s scope does not establish that the vote occurred, that any candidate assumed a seat, that Board capacity was restored or that later governance difficulties were cured. The election was planned in a setting where filling Board seats was expected to restore or preserve institutional capacity. Expectation is not outcome. The analysis stops at the published slate and planned voting timetable.

The nomination call described eligibility and experience criteria. For the two regional seats, residence tied the nominee to the relevant region. Seat 7 could draw nominees from any AFRINIC subregion. Once nominations arrived, the committee’s own report says it worked through online meetings and email, qualified or disqualified nominees under the criteria, sought additional information and relied on Election Committee checks. That account confirms that screening involved judgment and verification rather than the automatic reproduction of every submitted name.

On 10 May, the published slate contained one Northern candidate, two Western candidates and two region-independent candidates. Those five names are the visible output of the gate. They are not the denominator. The record examined here does not disclose how many nominations were received in total, who every unsuccessful nominee was, how many submissions failed threshold requirements, or why any particular person did not appear. Five final candidates do not imply five original nominations, and the distribution across the three seats does not establish the distribution of the unseen pool.

The same caution applies to the pool from which NomCom itself was chosen. AFRINIC’s record says several community members volunteered, but it does not publish the total number, the volunteers’ domiciles, their individual assessments or the reasons the Board chose the three community representatives it did. Article 9.1 supplies an eligibility filter. It does not, on the public record, supply a complete selection method among those who remained eligible.

The chain therefore has two selection stages whose visibility differs. First, the Board selected three community members and one of its own members for NomCom. Second, NomCom selected candidates from nominations. At both stages, a public rule is visible. At both stages, the eventual names are visible. Between rule and result, however, the public material is aggregate and incomplete. The Board vote is recorded, but its evaluation of volunteers is not. NomCom describes its methods, but individualized screening reasons and the full nomination flow are not disclosed.

Partial visibility should not be mistaken for proof that hidden steps were defective. Institutions reasonably protect personal data, references and candid candidate-merit discussion. People may be less willing to volunteer or stand if every exploratory exchange is broadcast. Confidentiality can support a fair process. Yet “confidential” is not a complete architecture. It must be possible to disclose enough about stages, criteria, conflict handling and correction routes to distinguish protected deliberation from an untraceable decision.

Where the votes and reporting sat

The internal structure of the committee makes those missing receipts important. The three community representatives chose the NomCom chair. The committee decided by simple majority. Dr Abdalla Omari, the Board representative, was a full voting member rather than a silent liaison. The published process also gave that representative a reporting function: general activity and progress, including whether qualified people had been found, could be reported to the Board, while confidential discussion of candidate merit remained outside the reporting line.

This arrangement tried to maintain two boundaries at once. NomCom needed a connection to the Board that appointed it and approved the election guidelines. It also needed a confidential space in which candidate assessments were not relayed as running commentary. Allowing general progress reporting while excluding confidential merits discussion is a recognizable attempt to separate institutional oversight from candidate-level influence.

But the representative’s vote means the Board connection cannot be understood as observation alone. One quarter of the panel was a Board member with the same formal voting status as the other members. Every committee appointment, including the three labelled as community representatives, came through the Board’s act. The community members chose the chair, which supplied an internal counterweight, but the public record does not show a broader appointing constituency or an independent confirmation step.

None of that demonstrates domination by the Board. A four-person committee can contain disagreement, and labels do not tell us how anyone voted. The minutes and report do not expose NomCom’s candidate-level deliberations. It would be an error to manufacture a voting alignment from institutional roles. The defensible claim is structural: the design concentrated appointment authority in the Board, gave its representative a committee vote, and asked the remaining three appointees to select a chair and reach majority decisions.

That design heightens the value of clear criteria and conflict records because personal independence cannot simply be presumed or denied.

The geographic rule complicates the structure further. It screened out residents of the two regions most directly connected to the regional seats. This lowered one obvious category of conflict. It did not test every other possible connection: professional alliances, organizational dependencies, past disputes, shared networks or interests in the region-independent seat. Geography can be a practical proxy, but a proxy is not a complete conflicts system. An appointee from outside a region can still have strong connections to a candidate or issue; a resident can sometimes contribute relevant knowledge while recusing from a home-region decision.

The public evidence does not show that any such conflict existed in 2022. It shows why domicile alone cannot bear the entire burden of proving independence.

The reporting boundary likewise needs process support. General progress reports can reassure the Board that the committee is functioning without exposing protected discussion. Yet a public accountability record should say what categories were reportable, who received them, whether conflicts triggered recusals and whether the Board could request action beyond approving guidelines. A formal statement that merits discussion stayed confidential is useful. A stage map would make it more useful by showing where oversight ended and finalization authority began.