Summary

  • Resolution 202107.624 accepted Mark Elkins' resignation as a Board member; Resolution 202107.625 then instructed AFRINIC to notify the Registrar of Companies and update its Registers of Members and Directors.
  • The paired resolutions make the Board's intended administrative response visible, but the public record does not establish when the resignation letter arrived, when the vacancy became effective, when either internal register changed, or when Registrar notice was delivered or accepted.
  • This distinction matters because later quorum, committee and signing-authority calculations depend on a roster tied to a particular time. A stale record creates uncertainty and cost without automatically invalidating a later act.
  • The strongest benign reading deserves weight: the Board dealt with the resignation and named both internal and external record actions in the same dated instrument. The remaining weakness is evidentiary, because an instruction is not a completion receipt.
  • A privacy-safe vacancy receipt could preserve the document identity, effective time, decision record, old and new roster versions, internal update times, Registrar delivery and receipt, downstream authority changes, corrections and final reconciliation.

The moment a seat becomes an empty socket

The two resolutions passed on 26 July 2021 are easy to read as administrative housekeeping. Resolution 202107.624 says that AFRINIC's Board took cognisance of Mark Elkins' resignation letter, accepted his resignation as a Board member and thanked him for his contribution. Resolution 202107.625 begins from that acceptance and directs two kinds of maintenance: notification of the Registrar of Companies, and updates to the Registers of Members and Directors. One decision recognised a departure. The next translated it into work for the records.

That translation is the real institutional act worth examining. A resignation is not only a biographical event concerning one person. For a private company, it changes the set of people who may count toward a meeting, occupy a committee role, sign under a delegation or appear to an outsider as a current office holder. Once a seat becomes vacant, every record used to answer those questions needs to refer to the same state, or at least explain why one record is legitimately waiting on another.

The published material shows that AFRINIC's Board understood this basic requirement. It did not merely note the letter and move on. It paired acceptance with an instruction aimed at both the external corporate record and two internal registers. This is the strongest benign account of the episode, and it is a substantial one. The acts are timely in their published arrangement, specific in their destinations and restrained in their verbs. They look like the ordinary work of a private company trying to make a change in its governing roster legible.

But a direction to update a register is not evidence that the register was updated. A direction to notify a public office is not a delivery timestamp or a receipt from that office. The difference is not semantic fussiness. It is the difference between knowing what an institution intended to do and knowing which roster governed a later decision at a particular moment.

The public resolutions do not establish when Elkins wrote the resignation letter or when AFRINIC received it. They do not disclose whether the letter named an effective date distinct from the Board's acceptance. They do not establish the exact instant at which each relevant corporate status ended. Nor do they publish the attendance, quorum, proposer, seconder, vote, objections, abstentions, recusals or conflicts associated with the two resolutions.

On the execution side, there is no published extract showing the revised internal registers, no update timestamp, no approved Registrar form, no delivery time, no receipt and no acceptance or reconciliation record.

Those absences must be described accurately. They do not prove that AFRINIC delayed the work, that no private record exists or that the notice was mishandled. It is entirely possible that employees and advisers completed the required steps quickly and preserved adequate internal evidence. What the public material allows us to say is narrower: the completion time cannot be reconstructed from the resolutions themselves.

That limit also corrects a misleading paraphrase elsewhere in the selection record. The official acts do not direct a change involving a registered agent. Resolution 202107.625 names the Registrar of Companies and the Registers of Members and Directors. Accuracy matters here because each destination represents a distinct state, custodian and evidentiary function. Adding an actor that the instrument did not name would make the chain seem clearer only by changing it.

Mauritius company law reinforces the need to keep the states separate. Section 142 of the Companies Act distinguishes a change in directors from notice to the Registrar and ordinarily allows 28 days for the required notice with specified particulars. That outer legal period does not tell us when AFRINIC delivered anything. More importantly, a statutory maximum is not necessarily a sound operating target for internal governance.

A board could need to know its current denominator at its next meeting, a bank could need a current mandate before processing an instruction, and a committee could need to know whether its composition remains valid well before a filing deadline expires.

Section 190, meanwhile, requires company records that include minutes and resolutions, certificates and current directors' full names and addresses. That recordkeeping frame supports disciplined maintenance, but it should not be used to collapse distinct kinds of register into one another. AFRINIC's rules connect a director's office with Registered Member status, and Resolution 202107.625 separately names the Registers of Members and Directors. A sound account therefore follows both internal destinations without mechanically treating the member register of a guarantee company as though it were a share register.

The practical risk begins when those records tell different stories. Imagine a later minute that gives only the names present and says quorum existed, without identifying the roster version used as the denominator. If one register has removed the departing director, another still contains the name, and an external record has not yet been reconciled, a reviewer must guess which state controlled. The same problem can touch committee eligibility, delegated authority, contract signatures and bank mandates. Nothing in the available evidence shows that such a mismatch affected an actual decision in 2021.

The point is that an unversioned gap makes the answer expensive to reconstruct when confidence is later needed.

This is where the useful function of AFRINIC must be stated precisely. NRS describes AFRINIC as a private, member-based body incorporated in Mauritius. Heng Lu's analysis supplies the sharper institutional boundary: the keeper of a ledger should be judged by the accuracy and continuity of its records, not mistaken for the source of public authority. AFRINIC coordinates technical and corporate records within a private framework. It is not a state, sovereign, legislature, regulator, police force, prosecutor, punisher, confiscator or public-law court. Its Board and membership do not acquire those powers through dependence on the registry.

That limitation does not make record quality less important. It makes record quality the centre of the job. When an organisation's legitimate authority is private and bounded, it must be especially clear about who held which office, under what internal rule, at what time and for which decision. The ledger does not create reality, but it helps participants, counterparties and reviewers identify the reality on which private action relied.

LARUS's operator-facing analysis explains why apparently small governance uncertainties travel. Network operators do not need every corporate detail to run a router, but they do need the institutions around critical registries to behave predictably. If a governance record becomes hard to interpret, operators, counsel, banks and counterparties spend more time monitoring, verifying and protecting against uncertainty. This episode does not prove an outage, routing loss or customer harm. It demonstrates the mechanism by which unclear institutional state can impose cost even when the underlying technical ledger continues to function.

BTW's earlier analysis of an election-to-register handoff provides a useful, brief contrast. An election moves in the forward direction, from a result toward corporate implementation and external notice. A resignation moves in reverse, from an occupied seat toward a vacancy that must be reflected across the same kinds of records. The comparison should stop there. The July 2021 episode is not a retelling of that election, a survey of Mauritius company law or an account of later institutional crises. Its distinctive question is how a departing name leaves every authoritative private record without leaving uncertainty behind.

A simple completion test follows. For each state transition, can a later reviewer identify the input, the accountable action, the time, the resulting roster version and the evidence of handoff? The resignation letter needs an immutable identity and receipt time. Any stated effective time needs to be preserved. Board cognisance and acceptance need their own decision record. The prior and successor rosters need versions rather than a silent overwrite. The two internal registers need separate update timestamps. The Registrar notice needs a delivery record and, when available, receipt or acceptance.

Committee roles, delegations, bank mandates and signing authorities need reconciliation. Later decisions need to cite the roster version used for quorum and authority.

Most of this can be disclosed without exposing a private letter, address, signature or confidential discussion. A public receipt can report that a verified document was received at a stated time, that a vacancy became effective under an identified corporate basis, that specified registers moved from one version to another, and that an external notice was delivered and acknowledged. Sensitive fields can remain restricted while the state transition remains visible.

The lesson of the paired resolutions is therefore neither accusation nor celebration. They show the correct instinct: turn a resignation into explicit record actions. They also expose the limit of a resolution-only public trail: intention is visible, completion is not. For quorum and authority to remain trustworthy, a vacancy must travel quickly through every relevant register, and the journey must leave timestamps, versions and receipts behind.