Summary

  • AFRINIC’s official archive records two distinct acts on 28 July 2022: consensus was announced on AFPUB-2021-GEN-002-DRAFT03, and the PDWG Chairs sent a ratification report to the Board.
  • That consensus was a real decision within a private policy-development procedure, but it did not itself ratify the proposal, instruct staff, alter the policy manual, change a registry record or direct a network operator.
  • The public record used for this analysis does not disclose the report, the exact version hash, the complete objection ledger, its attachments, proof of Board receipt or any Board disposition.
  • A separate Board gate has a defensible purpose: participants can test technical objections, while the corporation’s governing organ decides whether to accept implementation risk and commit staff, systems and budget within its private competence.
  • A compact consensus-to-ratification receipt could preserve both gates while making every transition reconstructible and expressly disclaiming sovereign authority or automatic routing effect.

Two verbs on one day

The official entry for 28 July 2022 uses a sequence that rewards close reading. Consensus was announced. A ratification report was sent. The first verb records the chairs’ conclusion about the state of discussion in AFRINIC’s Policy Development Working Group. The second records a transmission from that process to the corporation’s Board. Neither verb says that the Board ratified AFPUB-2021-GEN-002-DRAFT03. Neither says that management received an implementation instruction. Neither says that an amended policy appeared in a manual, that software changed, that a registry record moved or that an operator altered a route.

That is not an argument that nothing happened. Something consequential did happen. The proposal moved from deliberation to a chair-classified consensus state, and the chairs carried that result across an institutional boundary by sending a report to the Board. A proposal that remains in discussion is not in the same procedural position as one on which consensus has been announced. A conclusion that remains within the working group is not in the same corporate position as one formally handed toward a governing organ. The two recorded acts therefore deserve to be treated as acts, not as mere chatter.

But precision runs both ways. If it is wrong to dismiss consensus as inconsequential talk, it is equally wrong to stretch it into an executable rule. The archive identifies the proposal by its exact instrument number, AFPUB-2021-GEN-002-DRAFT03. It records the milestone and the handoff. It does not record the later corporate acts that would be needed to turn that procedural result into an organisational commitment. The gap is visible because AFRINIC’s own process description separated those stages.

This separation matters most when public accounts compress a policy’s life into a single label. “Approved,” “adopted” or “passed” can hide several different questions. Did participants and chairs close the discussion? Was the exact text transmitted? Did the Board receive it? Was an appeal pending? Did the Board ratify it or return it? Did management delegate implementation? Did staff deploy a change? Did operators observe that change and decide what to do locally? A truthful account answers each question with the evidence available for that stage. It does not allow proof of the first to stand in for proof of the last.

The narrow record here answers only the opening part of that chain. AFRINIC’s proposal archive says consensus was announced on 28 July and that the PDWG Chairs sent a ratification report to the Board the same day. It also records a same-day withdrawal and a 1 August archive. Those later entries bound the time in which the handoff might have advanced, but the procedural or legal meaning of the withdrawal is a separate question. It should not be recruited as an explanation of the consensus, a theory of motive or a substitute for evidence about what the Board did.

The better description is almost architectural. The working group had a door through which a proposal could leave community deliberation. The Board had another door through which the corporation could accept or return it. On 28 July, the record shows the proposal reaching the space between those doors. It does not show the second door opening.

The short path to the threshold

The essential chronology begins with Draft 3’s submission by Jordi Palet Martinez on 18 May 2022. The proposal was discussed at AFRINIC-35 on 1 June, and AFRINIC’s milestone summary labels that meeting stage “Rough Consensus.” That label did not end the process. The official account dates Last Call from 23 June through 14 July. A preserved RPD mailing-list message dated 12 July shows that substantive objections were still being expressed during that period.

The message is useful because it prevents the Last Call from becoming an empty box on a flow chart. It records an exchange involving concerns about temporary Board policy action, whether actions could survive after a later failure to reach consensus, and Board appointment of an Appeal Committee. Those are not trivial topics. They concern what happens when authority shifts between discussion, temporary action, appeal and institutional control. Their presence during Last Call shows that the process was still carrying contested governance questions close to its formal end.

The message must also be kept within its evidentiary limits. One preserved exchange is not the complete Last Call record. It does not establish how many objections existed, how many people held them, whether other objections were resolved, or what denominator should be used to measure participation. It does not by itself invalidate the chairs’ later classification. Nor does it reveal the reasoning by which each material objection was judged answered, non-sustained or otherwise treated. It proves something narrower and still important: specified objections were being voiced on 12 July, inside the recorded Last Call window.

That distinction protects both scepticism and fairness. The existence of an objection is evidence that a question was raised; it is not automatic proof that consensus was impossible. A consensus declaration is evidence that the designated chairs classified the process as having crossed its threshold; it is not automatic proof that every objection vanished or that every affected party authorised the result. An adequate record would connect the two by showing the material objection ledger and the chairs’ reasons. The official milestone summary used here does not reproduce that bridge.

On 28 July, after the Last Call window had closed, the archive records the final consensus announcement and the sending of the ratification report. This was not simply a repeat of the 1 June rough-consensus label. AFRINIC-35’s own published explanation laid out a staged procedure: meeting rough consensus, Last Call, final consensus, a report to the Board, Board ratification and Secretariat implementation. That sequence is valuable because it prevents a later narrator from retroactively turning the earliest encouraging signal into the final act.

Each date therefore does a particular job. The 1 June entry shows the meeting-stage assessment. The 23 June to 14 July interval shows the additional review period. The 12 July message preserves evidence of a live objection exchange. The 28 July entry shows the chairs’ final consensus classification and the report handoff. None of those entries supplies a Board resolution. None supplies an implementation record.

This sparse timeline is enough to make the central institutional point without pretending to reconstruct every discussion. The commission is not a general history of AFRINIC’s policy machinery, nor a clause-by-clause treatment of Draft 3. The instrument itself proposed a broader design, but the event under examination is narrower: the moment a chair-classified community decision became a documented corporate handoff, without a documented corporate disposition in the official milestone record used here.

A decision with real but limited force

Calling consensus “real” and “bounded” is not a rhetorical compromise. It is a description of what kind of decision it was. Within the policy-development procedure, chairs had a designated role: moderate discussion, evaluate the treatment of objections, declare whether consensus had been reached and transmit the result. Participants supplied technical experience, arguments, warnings, support and objections. On 28 July, the chairs’ announcement closed one procedural question in the affirmative and enabled the report to move forward.

That procedural force matters. Organisations depend on defined moments when deliberation changes state. If every disagreement kept a proposal perpetually open, no voluntary coordination process could finish its own work. A consensus procedure does not need to be a legislature to have internal consequences. It can set a status, authorize a handoff and establish what the next institutional actor is expected to consider.

Its limit follows from the same institutional design. The working group was not the corporation’s Board. Discussion participants did not, by speaking or remaining silent, become directors with fiduciary responsibility for AFRINIC. Chairs did not acquire authority to allocate staff time, change production systems, assume contractual exposure or accept corporate liability merely by classifying consensus. A report could present the result to the organ capable of making a corporate choice; it could not make that choice on the Board’s behalf.

Draft 3’s own proposed approval clause reflected this distinction. In broad terms, it contemplated that after consensus the chairs would submit the proposal and a short discussion report, and the Board would then either ratify after analysing implementation details and timing with staff or return the proposal with reasons and possible alternatives. Its proposed conflict-resolution provision also contemplated that ratification would be held while an appeal was pending. These provisions identify the kind of machinery being handed over, but proposal text is not proof that any step was performed.

A design for ratification does not establish ratification.

The bounded character of consensus also guards against a larger conceptual error. AFRINIC is a private, member-based technical registry and service provider. Its legitimate functions are consequential: maintaining accurate, unique number-resource records, supporting contactability and security services, documenting changes and helping interoperable networks continue to coordinate. Those functions require discipline and can materially affect the costs faced by resource holders and operators. But importance does not convert a private coordinator into a sovereign.

The 28 July decision was not a statute, a court judgment, a regulatory order, a confiscation, a police action or a routing command. It did not prove authorisation from all AFRINIC members, all African operators, African states, Internet users or a regional electorate. The record supplies no audited denominator against which such a claim could even be measured. “Community” can describe a venue and a method of technical participation. It cannot be allowed to imply a constitutional people whose attendance or silence transfers public power.

The exact claim that survives scrutiny is stronger because it is smaller. A private policy community, operating through an identified process, reached a chair-classified decision about a proposal. That decision triggered the recorded transmission of a report to the private corporation’s governing organ. The handoff was institutionally meaningful. It was not yet an executable corporate instruction, and it created no sovereign legitimacy.