Summary

  • AFRINIC Resolution 201608.297, associated with the Board meeting of 10 August 2016, said the Governance Committee needed to be elected at the next AFRINIC meeting in November and called a Special General Members’ Meeting as part of AFRINIC-25.
  • The later notice turned that sparse call into a meeting scheduled for 09:00 on 30 November in Flic-en-Flac, Mauritius. It placed twenty proposed bylaw special resolutions before a 95-minute Governance Committee election block and offered participation in person or by arranged proxy.
  • The notice made action possible, but the published record reviewed here does not establish delivery to every eligible member, the frozen electorate, candidate materials, proxy instruments, ballot and quorum rules, custody, reconciliation, objections or remedies.
  • The strongest fair reading is that the Board used a corporate power, paired the delayed election with an existing gathering and reduced participation friction. That benign explanation still does not supply a reproducible chain from call to result.
  • AFRINIC remained a private technical bookkeeper and coordinator. The committee was designed to give non-binding advice and take no direct action; neither a meeting nor a ballot created sovereign, regulatory, police, punitive, confiscatory or adjudicative authority.
  • The right remedy is fuller publication: authority, reasons, notice timing and delivery, exact voting objects, voter and proxy controls, ballot rules, aggregate participation, reconciliation, minutes, results and reasons, with ballot secrecy and personal data protected.

A sentence in August, a crowded morning in November

One sentence can open a door without showing what lies on the other side. That is the useful way to read Resolution 201608.297. AFRINIC’s 2016 Board register associates it with the Board meeting held on 10 August. The published record states the need to elect the Governance Committee at the next AFRINIC meeting in November 2016 and calls a Special General Members’ Meeting as part of AFRINIC-25. At that level, the act is plain: delay would no longer be the last public position. Members were to be convened, and the still-unformed advisory committee was to move toward an election.

The later notice revealed a much larger decision surface than that sentence disclosed. The SGMM was set for Wednesday 30 November 2016 at 09:00 in the Plenary Room of the Sofitel Imperial Resort & Spa in Flic-en-Flac, Mauritius. The first substantive block, from 09:05 until 10:30, contained twenty proposed special resolutions concerning AFRINIC’s bylaws, numbered SR1 through SR20. A tea break occupied the next half-hour. Only then, at 11:00, did the Governance Committee election begin.

The notice allowed ten minutes for a Nomination Committee report, thirty-five minutes for candidate presentations at five minutes per candidate, twenty minutes for voting, twenty minutes for counting and five minutes for announcement of results. Closing remarks were scheduled from 12:35 until 13:00.

That is a remarkable compression. The August call named a month, a host meeting and an election need. The November notice described a room, a starting hour, two distinct families of member action and a sequence measured almost to the minute. The first voting object was a block of twenty bylaw propositions. The second was an election to fill the membership-elected part of a Governance Committee whose advice would not bind the Board. They shared a gathering but were not the same kind of choice.

A member preparing to judge constitutional propositions faced a different informational task from a member preparing to compare candidates for an advisory committee.

The contrast is not proof that the meeting was unlawful, invalid or manipulated. It is proof of something more modest and more useful: the resolution alone was radically incomplete as an account of what members would actually be asked to do. A call can be valid as an initiating corporate act and still be inadequate as the full public record of member control. The analytical question is therefore not whether a short resolution should have carried every later detail.

It is whether the chain of documents, delivery, authorisations and records ultimately let an affected reader reproduce the route from the Board’s decision to convene members to the choices placed before them and the result announced in their name.

That distinction matters because the meeting belonged to a private company, not a polity. AFRINIC’s task was narrow and useful: bookkeeping and coordination for Internet number records. A special meeting could organise the company’s affairs. It could not legislate for Africa, govern networks, turn number records into a public franchise or create authority over states, non-members, operators and end users. The committee election concerned an internal accountability interface. It did not create a continental mandate.

What the August act established—and what it did not

The resolution did three important things. First, it acknowledged that election of the Governance Committee remained necessary. Second, it located that election at the next AFRINIC meeting in November 2016. Third, it selected an SGMM held as part of AFRINIC-25 as the corporate vehicle. Those points converted an earlier postponement into a renewed commitment to convene members. They also gave the Secretariat and Board an identifiable basis on which to prepare later notice materials.

The surrounding chronology helps explain why even that sparse commitment mattered. AFRINIC’s 2015 register records Resolution 201511.254 as creating the Governance Committee and aiming to complete the first election and appointment at the May or June 2016 Annual General Members’ Meeting. The 13 April 2016 minutes record a resolution previously passed by email, Resolution 201602.270, postponing the initial election and appointments until a November or December AFRINIC meeting at the request of the Nomination Committee. That postponement is not the subject here.

Its relevance is simply that, by August, the institution needed a concrete act that shifted the public record away from waiting and toward a members’ forum.

Resolution 201608.297 supplied that act, but little more. Its published text did not state the exact day, hour or venue. It did not set out an agenda. It did not name candidates or define the electorate. It did not mention proxy voting, ballot form, quorum, counting, custody, challenge or remedy. It did not identify a proposer or seconder, record a vote count, describe debate or state an exact time of adoption. Silence on those points is not evidence that nobody addressed them elsewhere. It means only that the published resolution does not answer them.

This is the first discipline of institutional reading: never make one instrument carry facts that appear only in another, and never turn absence from one page into an allegation about the entire organisation. The August act should be credited for what it did. It ended ambiguity about whether the Board intended to call a special meeting for the November election. It should not be credited with details that emerged later, and it should not be condemned for misconduct that the published materials do not establish.

The 2012 bylaws supply an important part of the authority chain. They describe AFRINIC as a private company limited by guarantee, and Article 11.6 states that an SGMM may be called at any time by the Board. It must also be called on a qualifying written request by members, but the record here concerns a Board call. The August decision therefore sat within a stated corporate power to convene such a meeting. That answers the basic “who could call?” question at a general level.

It does not resolve every notice question. Article 12.2 says that notice pursuant to Article 11 should describe the business in enough detail for a member to form a reasoned judgment and should reproduce the text of special resolutions. Yet that clause and nearby provisions refer in part to an AGMM called under Article 11.1. The materials reviewed here do not settle how every part of that drafting applied to this SGMM under Article 11.6. It would be careless either to ignore the notice principle or to pronounce a definitive legal result from language whose exact application is not established here.

The defensible conclusion is narrower: sufficient detail for reasoned member judgment is a strong governance measure, and the interaction of the relevant clauses deserves explicit explanation in a complete meeting record.

By 9 November, the documentary chain had taken another step. Board minutes for that date say the Board reviewed the SGMM notice as presented by the Secretariat. Through Resolution 201611.308, the Board resolved to circulate it to members. This establishes review and authorisation to circulate. It does not establish the dispatch timestamp, the channel used, the completeness of the recipient list, delivery success or receipt by each eligible voter. A public notice page is evidence that notice content existed in public. It is not by itself a delivery ledger.

Those distinctions may sound formal, but they decide what a reader can actually know. “The Board called a meeting,” “the Board approved a notice,” “a notice appeared publicly,” “the notice was sent,” and “each eligible member received usable notice in time” are separate claims. A strong record links them. A weak record invites readers to slide from the first true statement to the last unproven one. Resolution 201608.297 began the chain. It did not complete it.

The notice changed the object of preparation

The November notice did far more than insert a date into the August sentence. It changed what an attentive member had to prepare for. The short resolution pointed to a Governance Committee election. The notice presented a two-part decision morning: first twenty proposed changes to the bylaws, then the committee election. Even without examining the substance of SR1 through SR20, which lies outside this inquiry, their placement matters because agenda structure shapes the practical conditions of participation.

Bylaw propositions demand close reading. A special resolution is not merely a preference among personalities; it asks members to judge proposed changes to the company’s governing text. The notice scheduled all twenty between 09:05 and 10:30, an 85-minute block. After the break, the advisory-committee election followed. A distant member deciding whether to travel, designate a representative or arrange a proxy needed to know both that the agenda had expanded and how the two forms of choice would be handled.

The published record here does not reproduce the Board authority chain for adding the twenty special resolutions to the SGMM whose August recital named the need for the committee election. That absence does not show that authority was lacking. It identifies a link that should be visible. A complete record would explain who placed the propositions on the agenda, under which corporate provision, when the texts were finalised, and how the expansion related to the original call. Such an explanation would let members distinguish a properly developed agenda from a meeting whose scope simply grew between the initiating resolution and the notice.

The timed sequence offered real information. It told a prospective participant that the Governance Committee portion would begin at 11:00 and that voting was scheduled for 11:50. It allocated a NomCom report before candidate presentations. It allowed five minutes per candidate and set aside distinct periods for voting, counting and results. That is materially better than a bare promise of an election sometime during a conference. Time, place and order reduce uncertainty and help members decide whether and how to attend.

Yet a timetable is a plan, not a transcript. It cannot prove that the meeting began at 09:00, that each segment ran for its allotted period, that every presentation took place, that voting closed at the stated time or that results appeared at 12:30. Nor can the thirty-five-minute presentation block be reverse-engineered into a candidate count. The instruction “five minutes per candidate” and the block length do not establish how many candidates actually stood, whether unused time existed or whether the schedule changed. Institutional precision requires resisting arithmetic that the record does not support.

The notice also widened the stated participation channel by saying voters could participate and vote in person or arrange for a proxy vote. That sentence matters because the meeting took place alongside AFRINIC-25 at a physical venue in Mauritius. For a membership dispersed across many jurisdictions, travel cost, scheduling and distance could otherwise turn formal eligibility into practical exclusion. Proxy access can reduce those frictions. It can allow an authorised person to carry a member’s vote when the member cannot be in the room.

But the availability of a proxy is not the same as evidence about proxy use. The notice does not identify proxy holders, reveal how many appointments were made, state whether any one holder carried several proxies, reproduce the instruments, show when they were lodged, explain validation decisions or state whether any appointment was rejected. Nothing in the reviewed record supports a claim that proxies were concentrated, abused or decisive. The proper question is what record would allow authorised participation to be confirmed without exposing private personal data or compromising the ballot.

The difference between the call and the notice can therefore be expressed simply. The call created an institutional destination. The notice described a route. Neither, standing alone, provides the complete travel log. To understand member control, a reader needs to see not only that a meeting was summoned and an agenda published, but also who was entitled to act, which materials they received, how representatives were authorised, what ballots were issued, how totals were reconciled and how objections could be raised.

Two kinds of member choice shared one room

The presence of two voting objects is central to the August-to-November transformation. It is tempting to treat the SGMM as a single event and stop there. That loses the fact that members were asked to exercise judgment in two different modes.

The special resolutions concerned the company’s bylaws. The Governance Committee election concerned the selection of three membership-elected voting members for an advisory body. The committee’s September 2016 Terms of Reference described five voting members in total: two appointed by the Board and three elected by AFRINIC Membership. Additional roles included a non-voting Board liaison, a non-voting legal adviser and Secretariat support. The first election was meant to produce staggered expiries across three consecutive years.

Those structural facts define what the election could accomplish without importing candidates, winners, totals or term allocations into this account.

Each voting object called for different evidence. For the bylaw propositions, a reasoned voter would need the exact texts, an account of what each change would do and time to compare the proposals with the existing bylaws. For the committee election, the voter would need candidate names, eligibility, statements, conflict information and a clear voting method. A single notice can serve both, but only if its attachments and delivery make the distinct materials intelligible.

Combining the objects might have been efficient. Members already attending AFRINIC-25 could address significant corporate business in one morning. A shared venue could reduce travel expense, meeting cost and participation fatigue. One properly notified SGMM could be more practical than separate gatherings. That is part of the strongest benign explanation and deserves serious weight.

Efficiency, however, is not self-proving legitimacy. A crowded agenda can favour those who already follow institutional affairs closely. Regular participants may know which proposals are coming, which candidates are likely to stand and which forms must be lodged. Distant or occasional members rely more heavily on the formal notice and linked materials. When twenty constitutional propositions precede an election, the quality and timing of information become especially important because attention is finite. The issue is not that combined business is inherently improper.

It is that the record should show how each choice remained intelligible and genuinely available to the eligible membership.

This is where notice must be understood as a control, not ceremonial stationery. A notice does more than announce a room. It allocates preparation time, tells members which powers they are being asked to exercise, identifies the documents they need and signals the channels through which they may participate. If notice is precise and delivered reliably, it distributes information away from the institutional centre. If the public record does not show delivery and supporting materials, knowledgeable insiders retain an advantage that cannot later be measured.

The answer is not to infer capture. There is no evidence here of fraud, manipulation or an invalid agenda. The answer is to make unequal access to information less consequential by publishing the chain. An organisation confident in its procedures should be able to show when the agenda became final, when exact texts and candidate materials were distributed, which voter classes were eligible, how proxy instruments were validated and how the count reconciled. Transparency substitutes evidence for suspicion.

An advisory committee, not a new authority

The Governance Committee’s Terms of Reference impose the most important boundary on the meaning of the election. Version 3.1, dated 5 September 2016, described a body intended to advise the Board, membership and community on governance matters. Its advice was non-binding. It was not to take direct action. When formal advice went to the Board, the Board was to consider it in a timely way; if the Board declined to follow it, the Terms of Reference required publication of reasons.

That design can matter. A body that cannot command may still improve accountability by framing questions, examining institutional conduct and making it harder for the Board to depart silently from a reasoned recommendation. The reason-for-departure requirement creates an interface between advice and Board responsibility. It gives members and affected operators something concrete to inspect: the committee’s stated position and the Board’s explanation for choosing another course.

But the committee was not a regulator, tribunal, enforcement agency, police force or public legislature. It did not own or steward number resources. Its election did not make its members representatives of Africa as a political community. It could not punish an operator, confiscate an allocation, adjudicate rights or bind governments and end users. Describing the body accurately is not a semantic nicety; it determines the scale of legitimate scrutiny. Members were choosing part of a private advisory mechanism, not transferring sovereignty.

The same boundary applies to AFRINIC itself. NRS’s description of AFRINIC, Heng Lu’s governance doctrine, LARUS’s account of infrastructure exposure and BTW’s reporting on direct member participation together support a coherent institutional view. AFRINIC performs consequential technical coordination, and defects in its governance can impose real operational and economic costs. Yet consequence does not manufacture public authority. A bookkeeper can affect a business profoundly while remaining a bookkeeper. A coordinator can disrupt infrastructure confidence while remaining a private coordinator.

Heng Lu’s argument holds two propositions together. Informed, direct member participation can check opaque insider concentration and force corporate officers to give reasons. At the same time, “community” rhetoric cannot turn an internal ballot into ownership of the Internet or a mandate over a continent. A meeting room is not a democratic state. The membership is not identical to every operator, network or end user that may bear the consequences of registry conduct.

That separation is vital here. If the Governance Committee is inflated into a public watchdog with coercive authority, the election looks like a continental constitutional moment and invites claims the evidence cannot bear. If the committee is dismissed as irrelevant because its advice was non-binding, the real accountability mechanism disappears. The accurate middle is stronger: this was a limited advisory body whose ability to demand and publish reasons could improve the conduct of a private registry company. Because the power was limited, the record should be exact about the limited power members were exercising.

Member votes can authorise internal corporate choices within the governing instruments. They cannot authorise AFRINIC over people outside that arrangement. They cannot make number resources the collective property of meeting participants, confer police powers, or bind states and independent networks. The legitimacy question is therefore internal and evidentiary: did the company provide a sufficiently traceable path for authorised members to form part of an advisory committee under its own rules?

Membership, representation and the proxy channel

The notice’s phrase “in person or arrange for a proxy vote” addressed a practical obstacle, but only at a high level. Representation at a company meeting is a chain of authority. An eligible member must identify an authorised natural person or proxy holder; an instrument must express the appointment; the company must receive and validate it under applicable rules; the holder must cast only the vote authorised; and the ballot record must allow reconciliation without revealing a secret choice.

The 2012 bylaws contain proxy provisions, including written appointment and production before a meeting. They also include a proxy form with notice. Nearby language contains a ceiling of five proxies per member during an election and prevents a candidate for director from acting as proxy in such an election. Yet the Governance Committee Terms of Reference say elected committee members should be chosen by a procedure similar to Board-director elections. Similarity is not identity. It would be wrong to import every director-election detail automatically into the committee election without the specific instrument that did so.

This leaves a precise evidentiary question. Which proxy and representative rules actually governed the 30 November Governance Committee vote? The public notice tells members that proxy participation was available. The materials reviewed here do not establish the applicable form, filing deadline, validation standard, concentration limit, treatment of corporate representatives, revocation rule or response to a defective appointment. Those details may have appeared in separate materials. Their possible existence is part of the fair contrary case. Their absence from this record means the article cannot claim either compliance or breach.

The electorate presents the same issue. The Terms of Reference assigned election of three voting members to AFRINIC Membership. That establishes the institutional selector. It does not identify the frozen voter list, eligible member classes, good-standing requirements, authorised representatives, suspended rights or number of eligible votes. Nor does the notice establish turnout or quorum. A later statement that the election occurred cannot fill these fields retrospectively.

Why insist on a frozen electorate record? Because eligibility is the denominator against which participation becomes meaningful. An attendance number without the eligible population says little. A ballot total without the number of authorised voters cannot be reconciled. A proxy aggregate without the rules governing appointment cannot show whether representation operated as intended. None of this requires disclosure of how named individuals voted. It requires institutional totals and preserved authorisations.

The candidate side also remains bounded. The notice allowed thirty-five minutes for presentations at five minutes per candidate. It included a NomCom report immediately before those presentations. It does not provide the candidate list, nomination record, eligibility determinations, conflicts or withdrawals. The timetable cannot establish the number of candidates. Nor does the existence of a NomCom report disclose what the committee evaluated or recommended.

A sound election record would connect the candidate list to the ballot that was actually issued. It would state when nominations opened and closed, publish appropriate candidate and conflict information, record any eligibility decision with a reason, identify the final slate and preserve the ballot configuration. If a candidate withdrew or a ruling changed the slate, the revision should be timestamped. This is not an invitation to examine individual biographies or the 2016 winners here. It is a description of the evidence needed to make the call-to-election chain reproducible.

Ballot control is the final link. The notice gave twenty minutes to voting, twenty minutes to counting and five minutes to results. Those allocations say nothing about paper or electronic format, issuance, custody, invalid ballots, recount thresholds, observers, reconciliation or objections. They do not reveal whether vote weights existed or how quorum was determined. The public record should therefore be read with restraint: a count was scheduled, but the mechanics are not established by the schedule.

The annual report and Governance Committee page state that the first election took place on 30 November and that the committee was constituted. That bounded confirmation is enough to close the call-to-event chronology. It is not a licence to reconstruct the result. Winners, tallies, term allocation and later performance lie outside this account. The distinction preserves both facts: the August commitment reached an actual election, and the public materials reviewed here do not reproduce the controls by which an outsider could verify every step.

The strongest benign account deserves full consideration

A fair judgment must present the most plausible explanation for the Board’s choices before identifying what remains unproven. That explanation is substantial.

The Board had a stated bylaw power to call an SGMM at any time. The initial Governance Committee election had already moved from the May or June meeting to November or December at the Nomination Committee’s request. AFRINIC-25 provided an existing event at which members, candidates and institutional staff could gather. Calling the SGMM as part of that meeting could reduce incremental venue expense and travel friction. A public notice later supplied the exact date, hour, room and agenda. It expressly allowed a proxy route for voters unable to attend.

The schedule separated a NomCom report, presentations, voting, counting and results rather than collapsing the election into an undefined agenda item.

Adding twenty bylaw propositions might also have been rational. A special members’ meeting already being convened could address other business requiring member action, avoiding another meeting and the fatigue of repeated mobilisation. Consolidation can be a sign of administrative economy, not concealment. It is also entirely possible that candidate packs, voting instructions, proxy forms, membership lists, ballot protocols and other materials existed outside the published items examined here.

Every part of that account should temper accusation. There is no proper basis here to say the call, notice, meeting or election was fraudulent, captured, unlawful or invalid. There is no basis to say proxies were abused, that the agenda expansion was unauthorised or that the count was defective. The evidence does not support those allegations.

But the benign account remains an explanation rather than a reproducible record. Authority to call a meeting does not identify who authorised each added voting object. A detailed public agenda does not prove timely delivery. The availability of proxies does not show the instruments and validation rules. Scheduled counting does not show reconciliation. Confirmation that an election occurred does not disclose how authorised voters, ballots and totals matched.

This is why evidentiary restraint and accountability are allies. Restraint refuses sensational claims. Accountability refuses to fill institutional gaps with trust. The answer to an incomplete record is not a verdict of invalidity; it is a request for the missing links. If the efficient, member-friendly explanation is correct, a complete call-to-result record should reinforce it.

What the published materials cannot establish

The first gap concerns adoption detail. The Board register associates the August resolutions with the 10 August meeting, but the published material does not identify the proposer, seconder, vote count, debate or exact adoption time for Resolution 201608.297. Those omissions do not negate the resolution. They limit what can be said about how the Board reached it.

The second concerns the expansion from an election recital to a two-part agenda. The August resolution refers to the need to elect the Governance Committee. The later notice begins substantive business with twenty bylaw special resolutions. The record reviewed here does not reproduce the authority chain, reasons or intermediate decision that placed SR1 through SR20 on the SGMM agenda. It would be wrong to assume the expansion lacked authority; it would be equally wrong to pretend that the initiating resolution explained it.

The third concerns notice delivery. On 9 November, the Board resolved to circulate the notice. The public page shows notice content. Neither establishes when dispatch occurred, by which channels, to which recipient list, with what attachments or with what delivery success. A timestamped publication and dispatch ledger would answer these questions without revealing sensitive member information.

The fourth concerns the exact legal and corporate notice standard. The bylaws express a principle that business should be described in enough detail for reasoned judgment and that special-resolution texts should be reproduced. Some surrounding language is tied specifically to the AGMM provision. The materials do not settle the application of every clause to this SGMM. A contemporaneous explanation of the relied-upon provisions would have reduced later ambiguity.

The fifth concerns candidates and voters. There is no candidate slate in the notice, and the presentation timetable cannot establish one. There is no frozen electorate, eligible count, turnout or documented quorum in the materials examined. Membership as the selecting body is clear; the individuals and authorisations comprising the electorate at the relevant time are not.

The sixth concerns proxy and representation controls. The notice offers the channel but does not reproduce appointments, validation rules, totals or concentration data. The general bylaw language cannot be applied selectively to the committee election without confirming the actual election instrument. Proxy access should therefore be credited as an inclusion mechanism, while its operation remains unproven here.

The seventh concerns the ballot and count. The timetable distinguishes voting, counting and announcement. It does not establish the ballot medium, issuance control, custody, invalid-vote rule, observers, reconciliation, objection route or remedy. A scheduled twenty-minute count is not a counting protocol.

The eighth concerns execution. A notice describes planned order and duration. It does not prove that every block occurred exactly as written. The annual report and Governance Committee page confirm the election and constitution of the committee, but they do not convert the agenda into minutes or disclose the details excluded from this subject.

These gaps form a single pattern. Each arises at a handoff: resolution to notice, notice approval to delivery, membership to voter list, voter to proxy, candidate list to ballot, ballots issued to ballots counted, count to result, advice to Board reasons. Institutional accountability is often lost at handoffs because each document proves only one side. A complete record joins them.