Summary

  • AFRINIC’s publication of Draft 1 on 13 August 2020 made five Board interventions imaginable: initiate, vary, suspend, reject and require policy action. Publication did not establish consensus, adoption, contractual incorporation, implementation or use.
  • The strongest case for Board involvement is a demonstrated conflict with company law, a proven contractual commitment, technical feasibility, uniqueness, fraud control or running-network continuity. Even then, fiduciary duty identifies a responsibility to respond carefully; it does not supply an otherwise missing power.
  • Each verb needs a separate authority and consequence test. The minimum protections are a named source of authority, a narrow trigger, evidence, written reasons, notice, preserved community versions and objections, a time or return limit, continuity protection and independent review.
  • AFRINIC remains a private member-based technical registry and coordinator. Neither its Board, an RSA nor a PDP consensus result can become a source of sovereignty, legislation, regulation, policing, prosecution, punishment, confiscation or public-law adjudication.
  • The most useful practical control is an intervention receipt that records exactly what was done, by whom, under which instrument, for how long, with what evidence, safeguards, review and result—while leaving every unverified field visibly unknown.

L3 — Five verbs, five different powers

A risk signal is not an authority source

Corporate directors are expected to notice serious risks. That proposition is easy to accept and, in a technical registry, difficult to avoid. A proposed rule might be impossible to implement safely. It might conflict with a lawful obligation that the company can identify. It might put the uniqueness of number-resource records at risk, weaken fraud controls or expose running networks to avoidable disruption. A responsible Board cannot turn away merely because the risk appeared during a community process.

But noticing such a risk and possessing a power to intervene are different institutional events. The first is a judgement about facts and consequences. The second is a claim about who may do what, under which governing instrument, at which stage, against which object and subject to which remedy. A Board paper declaring the first cannot silently establish the second. Nor can a general appeal to stewardship fill in the missing steps.

That distinction is the proper starting point for AFPUB-2020-GEN-004-DRAFT01. The established record fixes 13 August 2020 as the publication date of Draft 1 and identifies its object as Board power to initiate, vary, suspend, reject or require policy action. It does not provide the complete clause language in the material available here. It does not establish definitions, thresholds, burdens of proof, deadlines, exceptions or an order of precedence among instruments.

The analysis therefore has to resist two temptations at once: treating the proposal as if it had no conceivable legitimate purpose, and treating the five verbs as if the mere act of proposing them settled their legitimacy.

The five verbs occupy different positions in a decision process. They alter different things and expose different people to different risks. “Initiate” can add a question to the community’s agenda. “Vary” can alter the text on which that community is deliberating. “Suspend” can stop the clock and interrupt participation. “Reject” can displace an outcome through a veto. “Require” can direct another actor and may attach consequences to non-compliance. Calling all five “oversight” would conceal precisely what needs to be examined.

Initiate: the least intrusive verb, if the ordinary process still governs

Initiation has the strongest claim to fit a private technical coordinator’s legitimate role. If the Board identifies a conflict between a proposal and a specific corporate obligation, it may be reasonable for it to put the issue into the ordinary policy process. Initiation can mean asking the community to consider a documented problem while leaving authorship, deliberation and the answer with that process. It can expose a risk without dictating the solution.

That benign reading depends on structure. The Board would need to identify the procedure that permits it to initiate. Its submission should disclose the acting organ, the proposal and version affected, the concern, the evidence, the superior instrument said to be in conflict and the alternatives it considered. A Board-originated proposal should be labelled as such, not presented as though it emerged from community consensus. The community should remain free to examine the evidence, dispute the premise, improve the text or choose a different lawful answer.

Initiation becomes more consequential if it secures procedural priority, bypasses normal requirements or obliges others to process a preferred solution. It changes again if “initiate” means not merely placing a question before participants but putting operative text into effect. Those distinctions cannot be resolved from the record available here. They are questions for the full wording and governing hierarchy, neither of which is established in this account.

The suitable check is therefore simple but firm: initiation should open deliberation, not predetermine it. Written reasons and a preserved version history would let later readers distinguish the Board’s risk notice from the community’s response. Where immediate continuity is genuinely at stake, a separately proven temporary corporate measure may be possible within contract and law, but it should not be smuggled into permanent policy by calling it initiation.

Vary: when advice becomes authorship

To vary is not merely to object. It changes the thing under consideration. If the community has produced text and the Board can substitute, delete or add words, the institutional authorship of the result becomes mixed. That may occasionally be defensible—for example, if a specific provision would require the company to do something demonstrably unlawful or technically impossible—but the justification must meet the intervention actually made.

The first control is a visible redline. Readers should be able to compare the community text, the Board’s proposed change and any later revision without losing the original record. The second is a written explanation tied to a precise authority source. “Fiduciary duty” is not precise enough. Directors’ duties may explain why the Board could not ignore the problem; they do not establish why this Board could select these words instead of returning the issue, seeking advice, disclosing the constraint or declining an unlawful implementation.

The third control is a route back to deliberation. If Board variation becomes the final text simply because the Board made it, community drafting has been displaced. A more proportionate design would treat the variation as a reasoned proposed amendment, preserve objections and ask the ordinary process to consider it. If an exceptional mechanism makes a variation temporarily controlling, that mechanism would need a clear trigger, duration, scope, return condition and independent review.

There is also a consequence question. A change to wording can alter technical operations, contractual expectations or the treatment of resource holders. The more directly the variation changes real-world service, the more important it becomes to separate proposal, policy, contract and applicable law. A private corporate organ may manage the company within lawful instruments. It does not acquire legislative power because its edit has broad technical consequences.

The source record examined here does not reveal whether Draft 1 required a redline, reasons, renewed deliberation, a time limit or review. It would be wrong to say those protections did not exist. It would be equally wrong to assume them. Their status is unverified here, and that uncertainty matters because “vary” is the point at which oversight can become authorship.

Suspend: stopping time can change the result

Suspension sounds temporary, and temporary action may sometimes be the least harmful response. If implementation would create an immediate technical conflict or if reliable evidence cannot yet be tested, a short pause can preserve options. A pause may be preferable to allowing an unsafe change or to inventing a permanent solution under pressure.

Yet suspension is not neutral. It determines which version remains operative, who must wait, whose costs continue and whether participation loses momentum. A process halted without an expiry can become a process defeated without a formal rejection. A suspension ordered after consensus but before implementation differs from one imposed during initial discussion. A pause applied to the whole process differs from one confined to a disputed provision. Each needs its own explanation.

A credible suspension mechanism would state the emergency or conflict, the evidence standard, the decision-maker, the affected stage, the start time and the automatic return or expiry condition. It would explain what can happen during the pause: whether evidence may be challenged, whether drafting continues, whether the community can respond and who decides that the condition has cleared. It would protect the current community record from being rewritten while attention is elsewhere.

Continuity protection belongs at the centre of this verb. LARUS’s operator perspective is useful because it asks what discretion means outside the meeting room. Registry decisions can affect planning, customer commitments and the dependable operation of networks even when they are described as internal governance. That observation does not prove Draft 1 caused an outage, a loss or any particular disruption. It explains why a suspension should identify the running-network baseline it protects and the operational change it forbids while the question is reviewed.

The sealed material does not establish Draft 1’s temporal limits, return conditions, hearing arrangements or continuity safeguards. Nor does it establish that the suspension power was ever used. The right conclusion is not that the proposal necessarily authorised an indefinite freeze. It is that duration and return are indispensable questions, and the answers remain unknown here.

Reject: a corporate veto must show the instrument above it

Rejection is the clearest veto among the five verbs. It extinguishes or refuses an outcome rather than asking for more work. That makes the authority chain especially important. A private Board may sometimes have to decline to implement a decision that would require unlawful conduct, breach a proven superior commitment or exceed the company’s technical capacity. It cannot be compelled by an internal procedure to do what law forbids.

Even in that strongest case, rejection should be framed as a bounded corporate response, not as a superior legislative judgement. The Board should identify the precise obligation or limitation that prevents implementation, provide the evidence, explain why narrower measures are insufficient and publish reasons that affected participants can examine. The original consensus record, if one exists, should remain intact. The Board’s refusal and the community’s result are two records, not one rewritten history.

Meaningful review is essential because the Board is both identifying the conflict and benefiting from the power to decide it. Review need not convert the policy process into a court. It should, however, provide an independent route to test whether the cited authority exists, whether the trigger is met, whether the evidence supports the finding and whether rejection was necessary rather than merely convenient. A remedy might include reconsideration, correction of reasons, a return to the process or another response authorised by the governing instruments.

The exact remedy would have to come from a proven source, not from institutional preference.

“The Board has fiduciary duties” is insufficient as a complete explanation. Duty can require care, inquiry and lawful action. It cannot generate a veto with no identified textual or legal basis. If the rejection affects service under an RSA, the company must also distinguish the policy judgement from whatever private contractual right it says governs implementation. A contract may carry obligations and remedies between proven parties under applicable law. It is not a statute and cannot turn the rejecting Board into a public regulator or court.

The available record does not show the hierarchy clause, reasons requirement, objection process, appeal, court interface or remedy associated with Draft 1. It also does not show that rejection occurred. Those absences impose discipline on the analysis: the veto risk is real as an institutional design question, while any claim about an actual veto or its result would go beyond the evidence.

Require: a direction needs a recipient, an act and a lawful consequence

“Require” is the most open-ended verb because it raises an immediate grammatical question: require whom to do what? A Board might require staff to provide technical analysis, require a proposer to clarify an ambiguity or require a stage of the internal process to be repeated. Those possibilities are not institutionally equivalent. A direction to company employees under ordinary management differs from a demand imposed on members, resource holders, participants or outsiders.

Every exercise of “require” would therefore need a named recipient, an exact act, a deadline, the source of the duty and the consequence of refusal. Without those elements, a request can be presented as a command and a command can acquire unstated penalties. The authority to ask for information does not automatically include the authority to alter rights, interrupt service or impose a punitive consequence. Nor does an ability to administer records become police, prosecutorial, punitive, confiscatory or adjudicative power.

The safest form of requirement would remain close to the company’s proven organisational and contractual relationships. It would use the least intrusive measure, give notice, allow a response, protect live service, record conflicts of interest and provide review proportionate to the consequence. If the consequence reaches resource administration, the Board would need to show the applicable lawful and contractual boundary rather than rely on the rhetorical importance of the registry’s role.

Heng Lu’s analysis supplies the controlling institutional warning here. A registry’s useful function is that of a technical bookkeeper and coordinator. Private process and regional language cannot launder that function into a public mandate, and registry administration must not drift into enforcement reserved to public institutions. This does not deny ordinary company management or bounded contractual remedies. It insists that each consequence remain in its correct legal category.

The full Draft 1 definitions, recipients, deadlines and consequences are not established in the available material. It is therefore not possible to say which version of “require” the proposal contained in every circumstance. The question to carry forward is exact: for every proposed direction, who was bound, by what instrument, to perform which act, and what would lawfully follow if they did not?

Six separate thresholds, not one completed event

The record also requires chronological restraint. Publication of Draft 1 established that a proposal existed. It did not establish PDP consensus. Even consensus would not itself prove adoption by the necessary corporate mechanism. Adoption would not prove that the result entered an RSA or otherwise became part of a private contractual relationship. Contractual incorporation would not prove technical implementation. Implementation in the abstract would not prove invocation against any particular party.

These are six separate thresholds: publication, consensus, adoption, incorporation, implementation and invocation. Evidence for an earlier step cannot be carried forward as if it proved the next. The policy-proposal index and AFRINIC’s annual reporting are primary institutional sources for the proposal’s identified existence and date in this account; they do not, by that fact alone, establish the remainder of the chain.

The sealed record does not establish whether Draft 1 reached consensus, whether a later draft changed the five powers, whether the Board adopted anything, whether an RSA incorporated a rule, whether the company implemented it or whether any intervention was invoked. It identifies no affected party and no operational outcome. That means neither triumph nor abuse can responsibly be narrated. What can be assessed is the design problem made visible by the five verbs: one institutional actor was contemplated as initiating, editing, pausing, vetoing or directing policy action, and each move would require controls fitted to its distinct effect.