Summary

  • The surviving official record establishes a definite act on 21 May 2009: the IPv4 Soft Landing proposal did not reach consensus at AFRINIC-10 in Cairo and was sent back to the RPD mailing list. It does not disclose the wording or authors of objections, support or opposition counts, the population considered, the qualitative assessment applied, the reasoning of the assessors or which co-chairs expressly concurred.
  • Non-majoritarian consensus can sensibly protect substantial minority concerns and prevent premature closure. But an open meeting is an opportunity to participate, not proof of representation. Cairo’s result should be treated as a real, cautious and reversible status in AFRINIC’s private workflow, without enlarging it into rejection by “the community,” authorization by affected operators or a public mandate.

The crucial distinction in Cairo is between a result and the authority later attached to it. A result plainly exists. On 21 May 2009, the official minutes and meeting report say that Soft Landing did not reach consensus. They also state the immediate consequence: the proposal went back to the RPD mailing list for more discussion. Vincent Ngundi, writing two days later as Chair of the AFRINIC PDP-MG, repeated the date, place, finding and next step. It would be wrong to say that nothing was decided.

It would be equally wrong to treat the four-part notice—proposal, no consensus, return, further discussion—as a complete account of collective judgment. The surviving official material does not reproduce a Cairo objection. It does not say who raised one, whose interests a speaker was authorized to express, what response was offered, or whether an issue remained unresolved. It gives no support, opposition or abstention count. It supplies no Soft Landing room count and no account of remote or mailing-list participation incorporated into the assessment.

It does not state the test used in the room, explain how the test produced the finding, or name the co-chairs who separately concurred.

Those omissions do not prove that objections, explanations, deliberation or concurrence were absent in Cairo. They prove something narrower and important: those elements cannot be reconstructed from the sealed official record. The evidentiary gap blocks a later reader from auditing the path between discussion and label. It also blocks the stronger claim that the label represented the people and organisations affected by a rule for administering a finite pool of unallocated IPv4 addresses.

The stakes are institutional as well as archival. Returning a proposal to discussion is reversible. It keeps an unresolved question open and can protect against adopting a rule too quickly. A representative public decision carries a different burden: it requires an identifiable basis for connecting the people who participated to the population on whose behalf authority is claimed. Cairo’s record contains no such basis. A reversible pause inside a private procedure is therefore not the same thing as a mandate from operators, resource holders, customers, governments, a continent or the public.

What the disposition establishes

The official materials put a precise frame around the act. AFRINIC-10 took place in Cairo, with the controlled policy disposition on 21 May. The agenda placed Soft Landing in shared policy-discussion blocks with two other proposals. Vincent Ngundi was listed for those blocks, and a continuation was scheduled after the break. The agenda does not assign a duration to Soft Landing alone, so it cannot establish how much of either block the proposal occupied.

The meeting report and the current official minutes page record three discussed proposals, one of them described as global. They say that the global proposal was approved, while Soft Landing and the IPv6 proposal for non-profit networks did not reach consensus and were returned to the RPD list for further discussion. That comparison establishes that the record distinguished procedural outcomes among the three items. It does not supply the missing reasoning for Soft Landing, and the count of three proposals is not a count of positions, objections or possible outcomes within the Soft Landing discussion.

Official meeting materials identify Douglas Onyango as the author or proposer of Soft Landing. A policy-summary presentation gives a short account of the proposal’s stated incentive: extending the life of IPv4 during transition. That is enough to explain why the subject had operational and economic weight. It is not permission to reconstruct the proposal’s detailed text, to infer which clause was debated, or to judge its merits. The Cairo evidence before us concerns the disposition, not a complete history or technical exegesis of the proposal.

The most contemporaneous concise account came on 23 May. Ngundi signed a message to the RPD list as Chair of the PDP-MG. He wrote that Soft Landing had not reached consensus at the 21 May meeting in Cairo and would now be returned to the list for further discussion, in line with the policy development process. The notice corroborates the procedural result and consequence. Its signature does not disclose Ngundi’s private reasoning, demonstrate a particular objection, or establish a separately recorded concurrence by another co-chair.

The archived process text explains why return to the list followed the finding. It says proposals could be submitted by anyone, discussed on an open mailing list and, after at least 30 days, brought to an open face-to-face meeting. It describes consensus as general agreement rather than a majority measurement, assigns the determination to MG co-chairs, and provides that mailing-list discussion repeats when consensus is absent. These are AFRINIC’s stated procedural terms. They connect the no-consensus label to the next step without converting either term into proof about what people in Cairo actually said.

That distinction between procedural rule and event evidence matters. A rule may authorize co-chairs within an institution to assess consensus. It does not by itself show how they assessed a particular room. A presentation may describe a forum as open or bottom-up. It does not show who attended a particular discussion, who spoke, whether views were independent or coordinated, or how closely participants resembled all people affected by the prospective rule. The process document explains what was supposed to happen; the outcome records say what status was announced. Neither fills the missing middle.

The five records and their limits

The agenda establishes scheduling. It shows that Soft Landing shared policy-discussion time with two other proposals and that Ngundi was named for the block. It cannot identify the proposal’s speakers or objections, measure the time actually used, or show the assessment method. A scheduled continuation after the break is still not a transcript.

The PDP overview presentation establishes institutional self-description. It portrays the process as coordinated and moderated by the PDP-MG, open to anyone and free of a formal participation prerequisite. It identifies the 2009 management group as Vincent Ngundi, Hytham El Nakhal and Paulos Nyirenda, with AFRINIC staff member Alain Aina providing support. That roster proves named roles in the described structure. It does not identify which person spoke as assessor, which co-chair concurred, or whether the roster members shared a single reasoning path for Soft Landing.

The policy-summary presentation establishes proposal identity, authorship, the stated transition incentive and a generic sequence in which the author introduces a proposal, the public discusses it and a way forward is determined. It does not preserve Onyango’s exact Cairo presentation, his answer to any objection, the words discussed, or the merits on which the outcome rested. A generic slide about deciding a way forward cannot substitute for the actual decision record.

The meeting report and the current minutes excerpt establish the recorded outcome. They place the meeting in Cairo, identify Soft Landing among the proposals, state that it did not reach consensus and say it went back to the list. The report also says Brian Longwe chaired the morning session on 21 May. Chairing a session is not the same fact as making the consensus assessment. The record does not assign that assessment to Longwe, and his chairing must not be used to bridge the gap.

Finally, the chair’s list notice establishes a timely restatement. It is strong evidence that the institution recorded no consensus and intended further list discussion. But it gives no objection, count, denominator, qualitative test, explanation or separate concurrence. The notice’s brevity makes the next procedural step easy to understand while leaving the route to the finding opaque.

Read together, these records support confidence about the outer edges of the event: what proposal was at issue, who authored it, when and where the disposition occurred, what status AFRINIC announced, and where the proposal went next. Their convergence does not increase the detail of the missing centre. Repetition of the same conclusion in several official documents is not independent evidence of the unrecorded reasons that produced it.

Three claims that must remain separate

The Cairo material invites three propositions that sound similar but carry different evidentiary burdens. The first is historical: AFRINIC recorded no consensus. This is directly supported. The second is procedural: the recorded status was the kind of finding that, under the described PDP, returned a proposal to mailing-list discussion. This too is supported by the process text and the reported consequence. The third is representative: affected people collectively withheld agreement. That claim is not supported because the affected group, its relationship to participants and the assessment inputs are not reconstructed.

Keeping those propositions separate prevents two opposite errors. One error erases the institutional act because its reasons are poorly documented. That would understate the record: the proposal’s path changed, and the official notice announced how. The other error treats evidence of the act as evidence of a mandate. That would overstate the record: an administrator can conclusively record what its own procedure did without conclusively recording what a population wanted.

The difference is especially important when an official source speaks in the institution’s own vocabulary. “Consensus” may have a defined procedural function inside the PDP, while “community” or “bottom-up” may express the institution’s conception of participation. An analyst may report those descriptions as attributed claims. The terms cannot do additional evidentiary work merely because they appear in an official document. Institutional authorship authenticates the statement as AFRINIC’s statement; it does not authenticate every implication later readers may attach to it.

This disciplined separation also protects reasonable defenders of the process. One does not need to prove continental representation to defend a cautious internal pause. Nor must one deny that an internal decision occurred in order to challenge claims of public authorization. The smaller procedural account is both more secure and more useful: it tells authors and participants what happened next while leaving broader legitimacy questions where the evidence places them—unresolved.

A denominator that 135 cannot supply

The meeting report gives a figure of 135 participants for the entire AFRINIC-10 event. The event ran from 10 to 21 May and included training, presentations, policy forums and elections. That number is therefore an event-wide attendance figure. It is not the number of people in the room for Soft Landing, the number who followed the relevant mailing-list exchange, the number who commented, or the number whose views assessors considered.

Nor is 135 an electorate. The process was described as consensus-based rather than a majority vote, and the record provides no basis for treating every event attendee as eligible for, aware of or engaged in the Soft Landing disposition. The report’s election figures concern a different institutional activity and cannot be imported into policy consensus. No percentage can responsibly be computed because neither a Soft Landing numerator nor a relevant affected-population denominator is available.

The available official outcome narratives disclose zero numerical counts for Soft Landing supporters, objectors, abstentions, room participants, remote participants, mailing-list participants incorporated into the assessment or eligible affected operators. “Zero” here describes disclosed numerical fields, not human conduct. It does not mean that nobody supported the proposal, nobody objected, or nobody abstained. Converting missing data into a social fact would commit the same error as converting 135 into a decision body.

There are at least three different denominators that later claims can blur. One is attendance: the people physically or remotely present during a specific discussion. Another is process participation: the people whose list messages, meeting interventions or other input the assessors took into account. A third is the affected population: existing and prospective resource users, network operators, their customers and others exposed to the consequences of registry allocation rules. Cairo’s official record supplies none of these for Soft Landing.

Even if a future record established the first two, they would not automatically equal the third. An open door removes a formal barrier; it does not compel attendance, fund participation, cure information gaps, select delegates or confer authority to speak for absentees. A participant may bring expertise and a useful operational warning without being a representative. Conversely, a person authorised by one organisation does not thereby represent other operators, an entire national market or the African public.

This is why “open” and “representative” cannot be treated as synonyms. Openness is a property of access to a forum. Representation is a relationship between a speaker or decision body and a defined constituency, ordinarily supported by selection, delegation, accountability or another authorising link. The official presentation’s claim of openness may be accepted as AFRINIC’s description of its procedure. The record does not identify a rule that turned self-selected participation into representation of affected interests.

The missing assessment chain

A reconstructible consensus finding needs more than the final label. It needs a chain connecting inputs to assessment and assessment to consequence. In Cairo, the first and last links are partially visible: a discussion was scheduled and the proposal was returned to the list. The middle links are not disclosed.

The record does not give the wording or even the subject of a Soft Landing objection made in the room. It does not identify an objector or supporter, record an affiliation, or state whether a contribution was personal or authorised by an organisation. It does not preserve the author’s exact response. It does not say whether an objection was regarded as major or minor, resolved or sustained, answered or unanswered.

It also does not state whether assessors used a show of hands, a hum, a poll, an informal reading of discussion, private consultation or some other qualitative method. That absence must not be used to assert that any one of these methods was used or that none was. The process text says consensus meant general agreement rather than majority measurement. It does not define the event-specific observations that led the co-chairs to conclude that general agreement was absent.

The identity of the assessment is similarly bounded. The process assigned consensus determination to MG co-chairs. The available roster names Ngundi, El Nakhal and Nyirenda as elected group members and Aina as staff support. Ngundi was listed for the policy block and signed the later notice. Yet the surviving event record does not attribute an assessment statement or a separate concurrence to El Nakhal or Nyirenda, and it does not identify precisely which named individuals acted as co-chair assessors for this finding. Ngundi’s signature should not be stretched into evidence of unrecorded individual acts.

This gap prevents reproduction, not recognition. A later analyst can recognize that AFRINIC recorded a status under its process. The analyst cannot reproduce the reasoning by asking: What was the substantial objection? What answer was offered? Why did the answer not resolve it? Which assessors agreed? What set of contributions did they consider? The answer to each question is absent from the sealed record.

That is an accountability problem even without evidence of misconduct. An opaque record may result from ordinary minute-taking limits, a preference for concise summaries, missing archives or a process culture that values conversation over documentation. The seal does not choose among those explanations. It supplies no basis for allegations of manipulation, capture, bad faith or hidden vetoes. Documentary insufficiency is itself the finding; intent need not be invented to make it consequential.

What “no consensus” did—and did not do

Within AFRINIC’s stated process, no consensus functioned as a safety valve. The proposal did not advance on 21 May. Instead, mailing-list discussion was to repeat. The status therefore preserved the possibility of further work and avoided treating an unresolved proposal as settled. Relative to adoption, that consequence was reversible.

Calling the result reversible does not make it trivial. A pause affects timing, expectations and the operating environment around a finite unallocated resource pool. Existing and prospective users may care about when administrative rules change; operators may plan around anticipated scarcity conditions; authors and participants invest attention in the next round of discussion. The sealed evidence does not quantify any cost and does not permit a judgment about whether delay was beneficial. It does establish that the consequence was continued deliberation rather than final closure.

Scarcity makes precision more valuable because procedural shorthand can influence conduct even when it does not have public authority. A network operator hearing that a proposal “failed” may form different expectations from one hearing that unresolved questions returned to a list. A prospective resource user may read finality into a label that the stated procedure treated as an invitation to continue. The archive cannot tell us whether anyone actually made those inferences in 2009, but it shows why the distinction between status and substantive rejection matters operationally.

The interests potentially exposed to allocation rules were not uniform. Existing resource users, prospective applicants, network operators and customers could face different incentives. Yet the record contains no census of these interests and no allocation of voices among them. It would therefore be speculative to claim either that the meeting protected a vulnerable minority or that a minority frustrated a broader need. Both narratives require information about the objection and affected population that the record does not provide.

What can be said is structural. A pause preserves option value: discussion can continue, a concern can be clarified, and a proposal can be reconsidered. At the same time, delay can carry opportunity costs, especially around a finite administrative pool. Because the record does not quantify those costs or identify the blocking concern, it cannot resolve the balance. The fact that both possibilities exist is a reason to document the assessment, not a licence to select the possibility that suits a later policy preference.

The no-consensus notice therefore performed two jobs of unequal clarity. It communicated the workflow change very clearly. It communicated the substantive obstacle hardly at all. The first job allowed the process to move back to the list; the weakness of the second could make that next discussion less efficient because participants lacked a canonical account of the issue to address. This is a consequence of the documentary design, not proof that the underlying discussion was defective.

The result cannot prove that the proposal was technically defective, economically unfair or badly drafted, because no Cairo objection is preserved. It cannot prove that the proposal was sound and defeated by a minority, because support and assessment evidence are equally absent. It cannot prove that “Africa rejected” Soft Landing or that affected operators authorised the finding. It cannot even establish that all 135 event participants knew of the disposition.

The careful vocabulary is therefore procedural. AFRINIC recorded that Soft Landing did not reach consensus under its private policy development process, and the proposal returned to the RPD list. “Rejected by the community” would be stronger in two ways the record does not support: “rejected” would imply a more final substantive judgment, while “the community” would imply a defined and represented constituency. Neither addition is necessary to describe what occurred.

Likewise, it would be inaccurate to say that Cairo produced no decision. A decision about procedural status and next step did occur. The limitation concerns the kind of decision and the authority it can bear. It is reconstructible as an institutional disposition but not as a representative entity decision by all affected operators, registrants, users or the public.

The strongest case for the process

The best defence of Cairo’s result begins by rejecting the assumption that every legitimate technical decision must be a counted majority vote. Internet coordination questions can be complex, and a bare majority may overlook a technically serious objection. Participants may hold different forms of expertise, and the people most exposed to a failure may be fewer than those who find a proposal acceptable. Requiring general agreement can slow a proposal until a material concern is answered rather than allowing numerical weight to bury it.

Open list discussion can also widen the time available for analysis. People unable to attend a meeting may contribute asynchronously; written exchange can expose assumptions and improve language. When agreement is genuinely uncertain, returning a proposal to the list is a cautious act. It prevents premature closure without declaring the proposal permanently defeated. Nothing in the sealed record establishes that adoption on 21 May would have produced a better outcome.

This defence has real force. The absence of majority measurement is not itself a flaw, and demanding a ballot could create false precision. A 51-to-49 split would not show whether the minority had identified a fatal operational issue. A room count would still not prove that the room represented everyone affected. Consensus can therefore be a useful internal coordination technique, particularly when paired with reasoned attention to substantial objections.

But the case for non-majoritarian judgment increases, rather than eliminates, the need for a reasoned record. If a number does not determine the result, the substantive objection and the assessors’ treatment of it become the main evidence. Readers need to know what remained unresolved, how it mattered and who accepted responsibility for the assessment. Otherwise a process designed to protect a technically important minority can become indistinguishable in the archive from an unexamined veto.

The proper answer is not to retroactively demand a crude vote in Cairo. It is to treat the recorded finding as valid for the narrow purpose the surviving evidence supports: a status in AFRINIC’s own workflow that triggered more discussion. The evidence does not justify promoting that status into a representative verdict. Openness supports an opportunity to participate; it does not prove participation, delegation, authorization or public consent.

AFRINIC’s institutional boundary

AFRINIC’s role here was that of a private technical coordinator and bookkeeper for Internet number-resource administration. Its records can support coordination by describing registrations and administrative decisions. That useful common-layer function does not make the registry the creator of the Internet’s operation, the owner of networks, or the owner of number resources already issued for use.

More specifically, AFRINIC is not a sovereign, legislature, regulator, police force, prosecutor, court, punisher or confiscator. It cannot acquire any of those powers by labelling a meeting outcome consensus or no consensus. A mailing list cannot create public authorization merely because anyone could in principle join it. A co-chair assessment cannot turn self-selected participation into a mandate for a continent. A private process may govern what the private coordinator does next without governing society at large.

This boundary does not strip the Cairo disposition of all meaning. Within the process, the finding controlled whether the proposal advanced or returned to discussion. Participants, authors and registry staff had practical reasons to respond to it. The boundary instead prevents institutional category error: administrative coordination is not sovereign lawmaking, and procedural standing is not public representation.

Nor does the boundary settle the merits of Soft Landing. One can believe a private registry needs an orderly method for considering allocation rules while insisting that the method’s outcomes be described accurately. One can value open technical participation while acknowledging who is missing. One can defend the caution of a no-consensus pause without pretending the pause expresses a collective will that the record cannot define.

The available sources from NRS, Heng Lu, LARUS and BTW supplied no direct evidence of the 21 May Cairo event in this research. The event account therefore rests on official AFRINIC records within their limits. The private-bookkeeper doctrine controls how authority is interpreted; it is not a substitute source for who spoke, what they said or how the co-chairs assessed the room.

What the archive permits us to conclude

The official record allows a firm statement with a deliberate stop. Soft Landing, authored by Douglas Onyango, was among the proposals discussed at AFRINIC-10. On 21 May 2009 in Cairo, AFRINIC recorded that it did not reach consensus. Under the described process, lack of consensus led back to mailing-list discussion, and that is the consequence the meeting record and Ngundi’s 23 May notice report.

The archive does not allow a reconstruction of the objections, speakers, responses, counts, assessment method, reasoned basis, co-chair concurrence or relevant denominator. The 135-person total belongs to the full event and cannot repair those absences. The three-member elected PDP-MG roster, with one staff support member, identifies a procedural structure but not a recorded vote or individual concurrence on Soft Landing. Brian Longwe’s chairing of the morning session does not identify him as consensus assessor.

These limits leave several possibilities open. There may have been one substantial objection or several. The discussion may have been careful or cursory. Assessors may have agreed explicitly or through an undocumented practice. People in the room may have spoken for themselves or with organisational authority. None of those possibilities can be selected from the sealed material. A missing record is not proof of a missing event.

The conclusion is therefore neither that consensus existed nor that the proposal deserved to advance. It is that a procedural no-consensus finding occurred and caused a return to discussion, while the chain needed to treat it as a reconstructible representative decision does not survive in the official record. The distinction protects both sides of the evidence: it respects the documented act and refuses to invent the authority behind it.

That distinction is the durable lesson of Cairo. Consensus can be a prudent way for a technical forum to avoid pushing through an unresolved rule. But a consensus label is not self-authenticating. Its public meaning depends on what was assessed, who assessed it, whose participation informed it and how far the institution’s authority actually extends. When those foundations are absent from the record, the honest description is smaller than the rhetoric often built around such moments: a private coordinator paused a proposal and reopened discussion. Nothing more representative has been proved.