Summary

  • Resolution 201404.196, listed by AFRINIC under April 2014, records that the Board approved Board-revised Election Guidelines submitted by staff for use in the 2014 Board elections. The register does not establish an exact April adoption day, the meeting at which the decision was taken, the vote, or the identity of the final approved file.
  • The act mattered because it moved choices about nomination, eligibility, voter authority, proxies, ballots, counting and announcement upstream of the next ballot. Rules fixed before candidates become winners or losers can constrain improvisation and give members a common standard against which administration can be examined.
  • The official preserved election-process page describes a detailed architecture, but it cannot be treated as a verified byte-for-byte copy of the instrument approved in April. AFRINIC’s own 2018 retrospective says the previous process was adopted in 2014 and then received minor changes.
  • Better evidence would combine a dated final file and cryptographic hash with a redline, a decision receipt, a role-and-recusal matrix, authority checks, ballot configuration and custody records, aggregate reconciliation, exception records and bounded review. These measures can legitimate a private corporate election; they cannot give AFRINIC sovereignty, regulatory jurisdiction, police, punishment, confiscation or adjudication power.

A consequential act in a very short entry

Resolution 201404.196 appears in AFRINIC’s Board register under April 2014. Its recorded act is concise: the Board approved Election Guidelines that staff had submitted and the Board had revised, for use in the 2014 Board elections. The next identified election setting was the annual general members’ meeting in Djibouti on 6 June 2014. That sequence places the approval before the next ballot and gives it practical significance. The rules were not merely an after-the-event description. They were intended to govern a corporate choice that was approaching.

That is the value of the resolution. It located consequential choices in a pre-ballot instrument, when the identities of winners and losers had not yet hardened around the application of each rule. Eligibility, nomination support, voter standing, proxy authority, ballot format, secrecy, custody, counting and announcement all create moments at which discretion can alter participation or confidence. Approving a rulebook in advance can turn those moments from improvisations into applications of a known standard.

The limit is equally important. AFRINIC Ltd was identified in the then-effective bylaws as a Mauritius-incorporated private company limited by guarantee. Its election could authorize officeholders within that private corporate arrangement. It could not manufacture governmental jurisdiction. AFRINIC remained a technical bookkeeper and coordinator: consequential because operators rely on a registry ledger, but not sovereign, legislative, regulatory, judicial or coercive. No internal ballot, however well run, could give it police power, a right to punish operators, a power to confiscate property, or authority to adjudicate title.

The public record therefore supports a disciplined inquiry. What was fixed before the ballot? Which private actor was assigned each function? Which version can be proved? What evidence would show that the approved rule was the rule actually configured and applied? Those questions take the resolution seriously without converting corporate procedure into public law.

What the register proves—and what it does not

The Board register is the primary proof of the institutional act. It supports three propositions. First, the resolution is grouped under April 2014. Second, it says staff submitted Election Guidelines that the Board revised. Third, it records approval for use in the 2014 Board elections. These propositions are modest, but they are firm enough to establish a pre-ballot rulemaking decision.

The resolution identifier is not a calendar date. The register separately mentions approval of minutes for a meeting held on 16 April, but it does not expressly attach Resolution 201404.196 to that meeting. It would therefore be inaccurate to turn 16 April into the resolution’s adoption day. “April 2014” is the defensible description. The exact day remains unproved.

The register also does not supply the underlying staff submission, a comparison showing the Board’s revisions, a final file, or a cryptographic fingerprint. It does not identify the meeting in which the resolution was passed, attendance, quorum, individual votes, abstentions or recusals. Nor does it record a legal review or explain why particular language changed. None of those absences proves that AFRINIC lacked such records internally. None proves invalidity, concealment, unfairness or misconduct. They are limits on what an outside reader can authenticate from the public record.

That distinction between absence of public proof and proof of absence is essential. A sparse register may truthfully summarize a sound process. A full internal archive may exist even when the public cannot inspect it. The responsible conclusion is not accusation. It is that the public entry performs only one evidentiary job: it attests that the Board recorded an approval for intended use. It does not by itself authenticate the bytes of the rulebook, reveal the chain of revision, or demonstrate execution.

The 6 June meeting record performs another limited job. It establishes the proximity of a Board-election setting after the April approval. It does not establish which precise provisions were invoked, how any nomination was treated, who cast a ballot, how voting systems were configured, how paper and electronic totals were reconciled, or whether a review was requested. The meeting page is temporal context, not an implementation certificate.

Three kinds of evidence must not be collapsed

The 2014 resolution, the then-effective bylaws and the preserved election-process page answer different questions. Treating them as interchangeable would create a confidence that the surviving materials do not support.

The bylaws supplied the corporate framework. Article 9 provided for a Nomination Committee chaired by a person selected by the Board, with three other people appointed by the Board. It said the committee would operate under guidelines prescribed by the Board. It also assigned functions involving the solicitation of candidates, qualification, possible interviews, the final candidate list and supervision of the polls. Article 10 placed pre-election administration with an Election Committee made up of staff designated by the chief executive, working in liaison with the Nomination Committee chair for election-day polls.

Those allocations help explain why Board approval of guidelines mattered. The Board did not need to conduct every check, carry every ballot or announce every result itself. Its role at the rulemaking layer was to prescribe the instrument under which other private bodies would act. The bylaws describe the institutional channels, but they do not reproduce the final text approved in April 2014. They cannot answer a version question that belongs to the guideline itself.

The official preserved process page supplies far more operational detail. It presents an election architecture extending from seat rotation and candidate eligibility through nominations, public comment, voting, counting and announcement. Yet its evidentiary role must be stated with care. AFRINIC later described the previous process as having been adopted in 2014 and subjected to minor changes thereafter. The preserved page also contains timeline material that reflects later movement. It is therefore evidence of the instrument family and its operational design, not a verified frozen image of the exact April-approved file.

The later retrospective is useful only for that lineage point. It confirms a 2014 baseline and warns that the process did not remain textually motionless. References there to changed features demonstrate why version control matters; they do not invite analysis of later elections. The subject remains the 2014 approval and the auditability of the instrument it put in place.

This three-part separation prevents overclaiming. The register proves approval and intended use. The bylaws explain the private corporate allocation under which guidelines were prescribed and administered. The preserved page illuminates the kind of rule architecture associated with the 2014 process, subject to version uncertainty. The later retrospective confirms adoption lineage and intervening movement. No single item proves the entire chain from rulemaking to final file to configuration to implementation.

The private allocation of electoral power

An election rulebook is most useful when it does more than state aspirations. It should divide authority so that no one title silently absorbs every function. The materials surrounding Resolution 201404.196 reveal several distinct roles that should be understood separately.

The Board occupied the prescribing role. It approved a version that it had revised after staff submitted it. That placed the Board above the operational workflow in one precise respect: the Board fixed the governing guidance. It does not follow that the Board performed eligibility checks, maintained the voter register, validated every proxy, held electronic credentials, counted paper ballots or reviewed its own decisions. Those tasks require their own allocations and evidence.

The Nomination Committee occupied a candidate-facing and poll-supervision role under the bylaws. Its functions included soliciting candidates, examining qualifications, interviewing where appropriate, producing a final list and supervising polls. The preserved process presentation adds architectural detail about review against process and criteria, possible interviews or background checks, and notices where nominations are found ineligible. But these descriptions do not establish how any named person was evaluated, and this analysis makes no such claim.

The Election Committee was a staff-based administrative body. The bylaws describe it as staff designated by the chief executive and associate it with pre-election administration, in liaison with the Nomination Committee chair for election-day polling. Administration is not a minor residue. It can include calendars, forms, completeness checks, voter records, proxy workflows, ballot preparation, credential arrangements and coordination of physical proceedings. A good rulebook should make clear which of those actions is mechanical, which permits judgment, and which must be escalated.

Members, registered contacts and authorized voters formed another layer. Their standing and authority determined who could nominate, support a nomination or vote. A proxy holder did not acquire a free-floating right to vote; the authority derived from a member and had to pass the process’s validation requirements. Candidates had a different status again. The preserved architecture barred candidates from carrying proxies, a separation that appears designed to prevent one form of role concentration.

Electronic-ballot trustees, paper counters and the person announcing results performed still other functions. Custody of voting credentials is not the same as counting paper. Counting is not the same as reconciling channels. Reconciliation is not the same as declaring a result. Announcement is not review. When these functions are bundled under the vague label of “the election,” the public cannot see where an error could be detected or who had power to correct it. A role map should instead make every handoff explicit.

These are private roles organized for a private corporate election. The word “authority” in this setting means authority derived from the company’s constitutional and membership arrangements. It does not mean territorial jurisdiction over Africa, operators or the Internet. A member’s vote can help select a director. A committee can apply a corporate eligibility rule. A trustee can hold election credentials. None of those functions is a public licence to regulate, police, punish, confiscate or adjudicate.

The candidate path: eligibility, nomination and publication

The preserved process presentation shows why a pre-ballot instrument can constrain decisions long before a vote is cast. It describes a recurring seat architecture in which three Board seats open each year within a rotation covering seats one through eight. Because that presentation is version-caveated, these figures should be understood as features of the preserved architecture rather than certified wording of the exact April file.

At entry, the page presents legal-capacity exclusions. It refers to minority, undischarged bankruptcy, director disqualification, non-natural-person status and adjudged unsoundness of mind. Such filters are consequential because they can prevent a name from reaching the ballot at all. Their legitimacy depends on more than their existence. A workable process needs a defined source for each criterion, a consistent method of checking it, a responsible decision-maker, notice to the affected person and a bounded route for correcting factual error.

Nothing in the public materials establishes that a particular person was subject to one of these exclusions. The point is architectural: eligibility sits upstream of voter choice, so its administration must be traceable. If criteria are vague, if evidence requirements shift, or if the decision-maker is unclear, a later ballot cannot repair the earlier exclusion. Conversely, when criteria and proof requirements are fixed in advance, candidates can understand the threshold and committees can show that the same standard was used.

The preserved page also allows self-nomination or nomination by another person. It links nomination authority to registered contacts of members in good standing and describes support from at least two authorized registered contacts. Again, the surviving presentation does not prove how any actual nomination was handled. It reveals the intended control points: who may submit, who may support, whether the member relationship is current, and whether the proposed candidate consents.

Staff completeness checks and candidate consent belong to administration. Assessment against eligibility and process criteria belongs to the committee role assigned that judgment. Interviews or background inquiries, where permitted, require an especially clear boundary because they can introduce information not visible on the face of a form. A rejection notice should identify the applicable rule and enough of the reason to permit correction or review, while withholding unnecessary personal information.

Publication of the candidate list, an online comment period and an opportunity for candidates to speak extend the process beyond private committee assessment. Notice allows members to know whom they may choose. Comment can supply relevant information, but it also needs moderation and evidentiary rules so that unsupported assertions do not become unofficial disqualifiers. Candidate speaking opportunities should be arranged on comparable terms. These are not decorative steps: they translate an approved rulebook into an informed membership decision.

The public materials do not preserve a complete 2014 record for publication, comments, speaking opportunities or individual nomination decisions. It would be wrong to fill that silence with assumptions. The defensible finding is that the preserved architecture contemplated these stages, while the available evidence does not prove each stage’s implementation for the 2014 ballot.

Voter standing and proxy authority are separate proofs

The preserved process page ties voting eligibility to resource-member good standing, registered membership or an approved proxy. Those concepts should not be compressed into a single list of names. Each asks a different question: does the member have the required status, is the person registered to act for that member, and, where the person acts by proxy, was that delegation validly made and approved?

A reliable voter register should carry a timestamp. Status may change, contact records may be updated, and proxy instructions may arrive near a deadline. The rulebook should identify the relevant cutoff and the system whose record controls. A receipt need not disclose a secret choice. It can confirm that a particular authority check was completed under a specified version of the rule, by a specified role, at a specified time, with any exception separately recorded.

The preserved presentation requires proxy approval before election day, caps a voter at five proxies and bars candidates from carrying proxies. It displays a 48-hour advance interval, but that timing is version-caveated and cannot safely be attributed to the exact April-approved file. The number five likewise describes the preserved architecture, not an actual proxy count or a certified copy of the April rule.

Proxy controls show how procedural design can separate participation from concentration. Advance approval gives administrators time to verify the delegating member and the appointed person. A cap limits accumulation. A candidate bar separates the pursuit of office from the carriage of other members’ votes. Yet written controls are only one half of auditability. The other half is an aggregate record showing how many proxies were submitted, accepted, rejected, duplicated, superseded or withdrawn, and how many each voter carried, without publishing secret selections or unnecessary personal data.

The public materials do not supply those 2014 authority receipts. That absence does not establish that unqualified people voted or that proxy rules were ignored. It means that the public cannot move from a description of the rule to independent confirmation of its application. The gap should be named at exactly that scale.

Ballot design, secrecy, custody and count

The preserved page allows both electronic and paper voting and describes an electronic period before the annual meeting. It displays an interval of at least seven days, but later timeline material prevents treating that interval as verified April wording. The distinction matters because opening and closing times affect who can participate and whether duplicate voting across channels can be prevented.

Electronic voting requires a configuration record. The candidate names and ordering, positions open, voter roll, credential issuance, opening time, closing time, trustee structure and tabulation settings should all refer to the same rulebook version. A hash of the configuration, a record of authorized changes and an exception list can show that the system voters used corresponded to the approved design. A screenshot or general description is not equivalent to that evidence.

Paper voting requires a different but parallel chain. The preserved presentation calls for secrecy, treats personal identifying marks as invalidating and awards a position to the highest vote total. Secret voting protects the member’s choice, but secrecy cannot mean recordlessness. Blank ballot quantities, issued ballots, spoiled ballots, unused ballots, sealed containers, custody transfers and count sheets can be recorded without connecting a voter to a choice.

The architecture also describes a paper re-vote for a tie and a None of the Above option even for a single candidate. The later retrospective identifies None of the Above among features that subsequently changed, which is precisely why the preserved page cannot authenticate the exact April text on its own. The important 2014 question is not whether a later design was better. It is whether the applicable rule for ties and a non-affirmative choice was fixed, identifiable and consistently applied before the ballot.

Custody should be separated from decision-making. The preserved page distinguishes electronic trustees from the paper-counting and announcement roles. It describes a counting group involving the Nomination Committee, Election Committee and up to four volunteers from other regional Internet registries, followed by an announcement from the chair and later publication. Those numbers and role descriptions remain version-caveated, and the page does not establish who actually held credentials or counted in 2014.

For an auditable election, custody evidence answers simple but decisive questions. Who could access electronic material, and under what threshold? When were credentials created and destroyed? Who received sealed paper materials? Were electronic and paper channels checked for duplicate participation? Did totals reconcile with accepted voting authorities and issued ballots? Were discrepancies or invalid papers logged? Who approved the final aggregation before announcement?

These questions do not demand exposure of secret choices. Aggregate numbers and signed handoffs can demonstrate control without deanonymizing voters. Nor do they presume a defect. Their function is prospective: a rulebook should specify the records that ordinary administration will generate so that confidence does not depend on memory after the result.

The approved instrument and the implemented election are different propositions

Resolution 201404.196 demonstrates a rulemaking act. It does not demonstrate that each rule was executed. The difference is familiar in any controlled process. A policy may be properly approved while a configuration departs from it; a configuration may be accurate while an exceptional event is handled incorrectly; an action may be correct while its evidence is incomplete. Each proposition requires its own receipt.

For the 2014 election, the public materials do not establish rule-by-rule nomination administration, voter or proxy validation, ballot configuration, electronic custody, paper custody, reconciliation, exception handling, announcement timing or review. That conclusion should not be inflated. It does not say the rules were not followed. It says the public record presently described cannot prove that they were.

The same discipline applies to fairness and validity. An official rulebook is relevant evidence, but it does not certify fairness merely by existing. An event page proves that an event was presented, not that every underlying action met the applicable standard. Conversely, the absence of a public implementation log does not invalidate a ballot. The careful position lies between institutional self-certification and speculative accusation.

This is why the exact final instrument matters so much. Without it, even a complete implementation archive would have an unstable reference point. With it but without receipts, the public could know the rule but not the performance. Auditability requires both: a fixed standard and evidence of action tied to that standard.

Internal legitimacy stops at the corporate boundary

There is a temptation to treat operational centrality as public authority. AFRINIC maintains and coordinates records upon which network operators rely. Decisions by its corporate organs can therefore have serious practical effects. That consequence makes good internal governance more important; it does not alter the legal nature of the institution.

An election can answer who is entitled to occupy a Board seat under AFRINIC’s corporate arrangements. It can authorize that person to exercise the powers lawfully available to the company and its Board. It cannot confer sovereignty over Internet number resources, operators, national jurisdictions, Africa or the Internet. Participation does not amount to a transfer of governmental power. Technical expertise does not create a legislature. Consensus language does not create police power. Operational dependence does not create a court.

The registry ledger should likewise not be confused with title itself. Confidence in the election affects confidence in the people overseeing a high-consequence coordination system. It does not make ledger entries a sovereign grant of property, and it does not give the bookkeeper a freestanding power to confiscate or adjudicate ownership. The election’s proper stakes are substantial but bounded: competent, transparent and accountable administration of a private coordinating institution.

That boundary protects both governance and analysis. If AFRINIC is described as a sovereign, every internal procedure is burdened with claims it cannot legitimately carry. If it is dismissed as inconsequential because it is private, the dependence of operators on accurate coordination is ignored. The correct description is more exacting: AFRINIC is a private technical bookkeeper and coordinator whose decisions matter operationally, whose internal elections therefore deserve strong evidence, and whose power remains non-sovereign and non-coercive.