Summary

  • Resolution 201808.446 ratified AFPUB-2018-V6-001-DRAFT01 on 8 August 2018; it did not ratify the separate V6-002 proposal, and it did not itself put revised clauses into operation.
  • The Board minutes made the remaining work explicit. Action Item 201808.22 assigned notification to the chair, while Action Item 201808.23 assigned implementation to staff. AFRINIC’s records date incorporation into CPM 1.3 to 23 November, 107 calendar days after ratification, and the public implementation notice to 29 November, six days later.
  • The sequence is a strong example of competent private coordination: recommendation, numbered resolution, recorded vote, assigned execution, versioned manual and public notice created a reconstructable service-change record.
  • That procedural competence has a strict boundary. AFRINIC is a private technical bookkeeper and uniqueness coordinator. Its internal acts can govern how it performs its registry service, but they create no sovereign, legislative, regulatory, police, prosecutorial, punitive, confiscatory or public-law adjudicative authority.

L3 — The vote that left two jobs unfinished

Inside the AFRINIC Board meeting of 7–8 August 2018, agenda item 9.1 presented what could easily look, at first glance, like a completed institutional act. The Policy Development Working Group co-chair recommendation had arrived. The proposal had passed through the policy development process and, according to the resolution’s own recital, reached consensus. A director identified in the minutes as SO proposed Resolution 201808.446, SI seconded it, and the Board passed it unanimously on 8 August. Yet the minutes did not close the file with that vote.

They immediately recorded two pieces of work still to be done: Action Item 201808.22 for the chair to inform the PDWG co-chairs, and Action Item 201808.23 for staff to implement the ratified policies.

That juxtaposition is the key to the entire record. The resolution was decisive, but it was not self-executing. It authorised the next institutional stage; it did not erase the need for that stage. AFRINIC’s own later records date implementation of the relevant proposal in Consolidated Policy Manual version 1.3 to 23 November 2018. Counting from 8 August to 23 November gives 107 calendar days. AFRINIC then issued its public implementation notice on 29 November, six days after the recorded implementation date. None of those intervals, by itself, shows delay, concealment or misconduct.

What the dates prove is narrower and more useful: recommendation, ratification, execution and publication were distinct acts, performed by different roles and evidenced by different records.

Getting the policy’s identity right matters before interpreting any of those acts. Resolution 201808.446 concerned the IPv6 Policy and References Update, AFPUB-2018-V6-001-DRAFT01. It did not ratify V6-002, which was the separate IPv6 sub-assignments proposal. The distinction is not clerical trivia. A numbered resolution is valuable only if a reader can connect it to the correct proposal, then follow that proposal into the correct operating text. Confusing V6-001 with V6-002 would attach the Board’s vote to a different policy question and corrupt every later conclusion about what was approved and when it became active.

The pre-ratification record reinforces that distinction without requiring a retelling of the technical redline. At AFRINIC-28, V6-001 was recorded as moving to Last Call on 9 May 2018. V6-002, by contrast, required more discussion. AFRINIC staff stated that V6-001 could be implemented as written without an impact on AFRINIC operations. That was a feasibility assessment, not the ratification itself, and it was not a universal finding that no operator could ever encounter a cost or question. It established that staff saw no obstacle to carrying out the proposal as written.

The later co-chair referral and Board vote then occupied their own places in the chain.

The Board materials describe the referral with unusual traceability. The minutes connect the co-chair recommendation to an archived item and place it under a numbered agenda entry. Resolution 201808.446 recites the procedural predicates on which the Board relied: completion of the policy development process, consensus and referral by the PDWG chairs. Those recitals tell a later reader what the Board believed had arrived at its desk. They are evidence of AFRINIC’s internal basis for action.

They are not evidence that an internal consensus amounted to political consent by a continent, nor are they a substitute for the later implementation receipt.

The vote record is similarly specific. The named initials for proposer and seconder, the unanimous result and the resolution number make the corporate adoption act reconstructable. Reconstruction is important because institutional memory otherwise tends to compress a series of events into a single phrase: “the policy was adopted”. That phrase can conceal several questions. Did a working group merely recommend the proposal? Did the Board approve it? Did staff change the operating manual? Which version contained the change? When were service users told? Resolution 201808.446 answers only the approval question.

The action items and later implementation records answer the remaining questions.

The Board chair’s announcement on 21 August makes the unfinished status particularly clear. It reported that the Board had ratified V6-001 at the 8 August meeting and said staff would proceed with implementation, after which the community would be informed. The grammar of that sequence matters. “Would proceed” points forward. It does not describe a manual already changed on the day of the vote. The announcement therefore supports the minutes’ division of labour: the Board had completed ratification, staff had implementation ahead of it, and a completion notice would follow.

There is no need to manufacture drama out of the 107 days. The sealed record does not identify the engineering, documentation or release tasks carried out during that period. It does not show whether some internal deployment occurred before 23 November. It therefore cannot support a claim about why the interval lasted as long as it did. Ordinary implementation and release work is a sufficient benign possibility. More importantly, AFRINIC eventually recorded a determinate implementation date and a determinate CPM version. Visibility turns an interval into a manageable part of change control.

The institutional problem would be an unmarked effective date, a retroactive demand or an inability to identify the governing edition, none of which is established here.

The distinction also prevents an inverse error: treating an announcement as though it were the source of operational change. The public notice appeared on 29 November, but the implementation report and AFRINIC-29 record date the policy’s activation to 23 November. Publication supplied public notice and an accessible account of what had been incorporated; it did not move the recorded implementation date forward by six days. A precise lifecycle can contain an approval date, an implementation date and an announcement date without forcing them into one artificial “effective” moment.

Each record is strongest when used for the proposition it can actually prove. The Board minutes and resolution register establish the identity of the proposal, the procedural recital, the vote and the follow-up assignments. The 21 August announcement establishes that implementation was still prospective at that point. The staff implementation report and AFRINIC-29 record identify 23 November and CPM 1.3. The 29 November notice establishes public communication and describes the incorporated changes. None of those documents alone carries the whole lifecycle, but together they create a linked evidential chain.

That chain is also the answer to a superficially awkward archival detail. The proposal page retained an “Under Discussion” status label even though the Board, implementation and CPM records show a later lifecycle. A single field on a preserved web page cannot reasonably outweigh the subsequent documentary sequence. The right response is not to edit history in one’s head or to privilege the latest-looking interface element. It is to read the proposal page as one record created at one stage, then test it against the resolution, action items, implementation report and versioned manual.

Lifecycle status belongs to the chain, not to an isolated label.

The public implementation notice presents a second documentary imperfection that must remain visible. The proposal and other records use section 6.5.4.1, while the implementation notice prints 6.6.4.1. The materials establish an inconsistency; they do not establish its cause. It might be tempting to declare one number an obvious typographical mistake and silently replace it, but that would convert inference into fact. The sound approach is to preserve both readings, flag the mismatch and refrain from deciding what an unexamined source was intended to say. Documentary discipline is most valuable when it resists easy tidying.

The same discipline applies to the September approval of the August minutes. Resolution 201809.450 later approved those minutes with amendments. The available record does not identify the amendment text and does not show that an amendment changed Resolution 201808.446. It is therefore proper to disclose the qualification and improper to invent its significance. The numbered resolution, vote and action items remain evidenced; the unproduced amendment wording remains unknown.

Seen at this close range, the 8 August meeting was neither empty ceremony nor instant implementation. It was a bounded corporate handoff. The PDWG chairs’ recommendation allowed the proposal to reach the Board in the form AFRINIC’s process expected. The Board’s resolution recorded corporate approval. The chair’s task carried notification back towards the policy participants. The staff task carried the approved change towards the operating manual. CPM 1.3 and the implementation report created the operational receipt. The later notice made that completed stage public.

That division is not bureaucratic fussiness. It distributes responsibility and makes error easier to locate. If the wrong proposal identifier appears, the referral-to-resolution link can be checked. If approved language does not appear in the manual, the execution stage can be checked. If users are uncertain which version is active, the implementation receipt and notice can be checked. An institution that collapses all of these acts into “consensus decided” deprives users of those diagnostic handles. Here, despite the surviving status label and section-number mismatch, AFRINIC left enough handles to reconstruct the essential path.

The close factual conclusion is therefore modest but firm. On 8 August 2018, Resolution 201808.446 ratified V6-001 through a unanimous Board vote. That vote did not itself revise the working manual. Action Item 201808.23 demonstrates that implementation still belonged to staff, just as Action Item 201808.22 shows that notification remained assigned to the chair. On 23 November, AFRINIC’s implementation records place the update in CPM 1.3. On 29 November, AFRINIC publicly announced it. The 107-day and six-day intervals are properties of that documented sequence, not evidence of fault.