Summary

  • Draft 1 was posted to the RPD mailing list on 9 February 2016. Draft 2 followed on 16 February, seven calendar days later. Both preserved pages carry the same 4 February submission date, name Omo Oaiya, Joe Kimaili and Alain P. AINA as authors, and show the proposal as Under Discussion.
  • The decisive documentary change was architectural. Draft 1 said it amended selected clauses of AFPUB-2010-v4-005 and obsoleted nothing. Draft 2 supplied a complete Section 3 and said it obsoleted the existing IPv4 Soft Landing Policy. Its revision history gives two reasons: Section 3 was made completely new, changing the proposal from amendment to replacement, and the title was changed for clarity.
  • The visible scarcity parameters at issue remained substantially intact: a /15 Phase 1 maximum, the /11 trigger for Phase 2, a /22 Phase 2 maximum, 90-percent prior-resource utilisation, an IPv6-resource precondition, the regional-connectivity condition for out-of-region use, and the /16, /14 and /13 reserves.
  • Publication demonstrates drafting initiative and records what was proposed. It does not demonstrate community consensus, AFRINIC adoption, implementation, actual allocation effects or public-law authority. AFRINIC's role here was that of a private technical bookkeeper, service operator and coordinator, not a sovereign legislature or regulator.

A seven-day change in the shape of the proposal

The first public interval is short and exact. Draft 1 was posted to the RPD mailing list on 9 February 2016. Draft 2 was posted on 16 February. The elapsed time was seven calendar days. The two preserved draft-detail pages retain a submission date of 4 February 2016, list the same three authors—Omo Oaiya, Joe Kimaili and Alain P. AINA—and mark the proposal Under Discussion. Those facts establish continuity of authorship and a rapid version sequence. They do not establish why the authors returned with a differently structured text after a week.

What changed was the proposed instrument's relationship to the policy already on the books. Draft 1 described itself as an amendment to AFPUB-2010-v4-005, the existing IPv4 soft-landing policy, and said it obsoleted none. It offered proposed replacement wording for selected clauses: 3.5.1, 3.6, 3.8 and 3.9. In practical reading terms, Draft 1 was a patch. A reviewer had to place those proposed clauses beside the existing policy to understand the complete exhaustion mechanism, but the limits of the authors' stated intervention were comparatively visible.

Draft 2 took another form. It said it would obsolete the existing IPv4 Soft Landing Policy and supplied a complete Section 3. Instead of asking the reader to splice several replacement clauses into the earlier instrument, it presented the relevant policy architecture together. That was a larger documentary move even where the numerical settings remained the same. The proposal no longer announced, “change these selected pieces while leaving the base instrument in place.” It announced, in effect, “read this complete section as the proposed successor instrument.”

That distinction matters because form determines the work a reviewer must do. A patch foregrounds differences but requires cross-reading. A replacement foregrounds coherence but may conceal provenance. Once old, revised and newly supplied sentences appear in one uninterrupted text, a reader needs an additional comparison record to know which is which. The central event on 16 February was therefore not legislation, nor an exercise of sovereign allocation power. It was the publication of a revised private proposal whose architecture had expanded from selective amendment to full replacement.

The official pages establish the publication dates, authors, wording, status and reasons recorded in the revision history. They do not show a settled wider process. There is no basis in this record for turning Under Discussion into consensus, ratification or commencement. There is equally no basis for treating the authors' ability to revise their own text as a public mandate. Their initiative was real and documentable: they chose the wording and form of a proposal and issued a second version. The authority that can reasonably be inferred stops there.

What Draft 1 actually put forward

Draft 1's structure makes its claimed scope readily legible. It identified an existing instrument, said that instrument would be amended, specified that none would be obsoleted, and then presented replacement language for four selected clauses. That format matters independently of the policy substance. It tells a reviewer where to look for the proposed differences and implicitly leaves the surrounding architecture of the base policy in place.

The two early drafts shared materially the same problem statement and the same four proposed responses. They sought to reduce the Phase 1 maximum, require IPv6 resources, reserve space for critical infrastructure and new entrants, and remove an explicit minimum. This shared agenda is important because it prevents the second version from being misdescribed as a wholly new response to scarcity. The record instead shows a continuing proposal expressed through a different documentary form.

In the visible text, Draft 1 set a /15 maximum in Phase 1. It provided for movement to Phase 2 when no more than a /11 of non-reserved space from the Final /8 remained. It used a /22 Phase 2 maximum, required 90-percent utilisation of prior resources, made IPv6 resources a precondition, and limited out-of-region use to circumstances supporting connectivity back to the AFRINIC region. Its reserve design included a /16 for critical infrastructure, a /14 for new local Internet registries or End Users, and a /13 for unforeseen uses.

Those settings describe a proposal about the administration of a scarce pool. If the text had eventually been adopted and applied, criteria of that kind would have shaped the size, timing and eligibility of requests. But the conditional is essential. The Draft 1 page is evidence of proposed wording, not evidence that an allocation occurred, that a request was refused, that a network changed its plan, or that the address pool moved in a measurable way because of the proposal.

The patch format also gives Draft 1 a specific audit advantage. A reader can begin by asking why clauses 3.5.1, 3.6, 3.8 and 3.9 were selected. The stated amendment boundaries narrow the comparison. They do not make interpretation effortless—a reviewer still needs the existing policy to reconstruct the whole mechanism—but they make the proposed touchpoints conspicuous. The document's own posture says that the rest of the base instrument is not being displaced.

That posture is the proper starting point for comparison with Draft 2. It would be misleading to compare the total length of Draft 2's Section 3 with only the selected replacement passages in Draft 1 and then label every additional line a new policy choice. Draft 1 depended on an existing instrument for context. Draft 2 reproduced a fuller architecture because it proposed to stand in place of that instrument. Difference in displayed volume is therefore not identical to difference in substantive invention.

What Draft 2 changed—and what it did not

Draft 2 changed the proposal's declared effect. Where Draft 1 said “amends” and “obsoletes none,” Draft 2 said it obsoleted the existing IPv4 Soft Landing Policy. It then provided a complete Section 3 covering affected documents, definitions, the current phase, exhaustion phases, recurrence, a planning period, allocation criteria and reserves. That complete restatement is the clearest evidence that the authors had moved from a patch model to a replacement model.

The revision history confirms this interpretation in unusually direct terms. It says Section 3 was made completely new and that, as a result, the proposal now obsoleted the existing policy rather than amending it. It separately records that the title changed for clarity. Those two entries are the stated reasons available for the second version: structural reconstruction of Section 3, with a consequent change in the proposal's formal relationship to the prior policy, and clarification of the title.

The phrase “completely new” must be read in its documentary setting. It describes the construction of a complete new Section 3 in the proposal. It does not, by itself, establish that every sentence in that section was invented between 9 and 16 February. A replacement needs to carry forward definitions, transition mechanics, criteria and reserves that a patch could leave in the underlying instrument. The archived material does not provide a clause-origin table capable of identifying every sentence as inherited verbatim, edited from earlier language or newly drafted.

The responsible conclusion is about presentation and architecture, not universal novelty.

The visible scarcity settings reinforce that conclusion. The Phase 1 maximum is /15 in both drafts. Both place the transition to Phase 2 at the point when no more than a /11 of non-reserved Final /8 space remains. Both retain a /22 Phase 2 maximum. Both use the 90-percent prior-resource utilisation test, the IPv6-resource precondition and the condition that out-of-region use must support connectivity back to the AFRINIC region. Both retain the /16 critical-infrastructure reserve, the /14 reserve for new LIRs or End Users, and the /13 reserve for unforeseen uses.

This is not to say that form is cosmetic. Turning selected amendments into a complete successor text can change how a proposal is interpreted, maintained and later cited. It can decide whether readers regard a sentence as part of the live base policy or part of the proposed replacement. It can alter the amount of material that must be reviewed before anyone can sensibly assess the draft. It can also reduce the chance that a future reader accidentally combines amended and unamended provisions in the wrong way. Structural change is real change; it is simply not the numerical tightening story that the visible comparison does not support.

Draft 2's fuller Section 3 includes definitions and current-phase mechanics, an express Phase 2 clause, language addressing repeat requests, an eight-month allocation and assignment period, criteria and reserve administration. The eight-month period is described as promoting short-to-medium-term requests and equitable distribution. Its presence in Draft 2 shows part of what the complete section contained. It does not prove that the authors first conceived that period during the seven-day interval. Without an origin map against the base policy, the sentence cannot safely be assigned a 16 February birth date.

The same caution applies across the restatement. A reviewer may identify language present on the Draft 2 page but absent from Draft 1's selected clauses. That observation alone cannot answer whether the language was inherited from AFPUB-2010-v4-005, adjusted during the rewrite or genuinely introduced in Draft 2. Draft 1's patch did not attempt to reproduce every provision it expected to remain operative. Silence in a patch is not reliable evidence of nonexistence in the base instrument.

The two stated reasons, and the space beyond them

The version history gives a small but valuable explanation. First, Section 3 was made completely new, so the proposal shifted from amending the existing soft-landing policy to obsoleting it. Second, the title was changed for clarity. These statements should be taken seriously because they explain how the authors represented their own revision. They are not merely labels attached by an outside commentator.

They are also the limit of the available causal account. The preserved pages contain no author memorandum describing the drafting meeting, no sequence of RPD messages tied to the rewrite, no named requestor, no recorded objection that the new structure answered, no vote, no consensus record, no staff assessment and no clause-by-clause rationale. Nothing in the available material identifies a particular comment, constituency, staff request, Board instruction or private pressure as the cause.

That absence should produce restraint, not insinuation. It is possible to explain the mechanics of the change without inventing a hidden motive. The authors had a proposal in amendment form on 9 February and a proposal in replacement form on 16 February. Draft 2's own history says the newly complete Section 3 produced the switch in formal effect. A reasonable benign interpretation is that the authors concluded a patch would force readers to combine too many pieces and that a single coherent section would be clearer.

This interpretation fits the stated structural reason, but it remains an interpretation rather than a preserved account of a specific conversation.

The title entry requires even tighter handling. The revision history states that the title changed “for clarity.” The preserved material does not supply a reliable earlier title string from which the before-and-after wording can be reconstructed. The sound finding is therefore limited: a title change was recorded, and clarity was the stated reason. Supplying a guessed earlier title would create evidence rather than report it.

The distinction between a stated reason and a reconstructed motive is central to institutional reporting. “Section 3 was made completely new” describes what the authors say they did. “Therefore a particular actor demanded the rewrite” would require evidence of an interaction that is not present. “The title changed for clarity” records a stated purpose. “Therefore the old title caused confusion among a named group” would require messages or testimony that the pages do not contain. Good analysis can draw consequences from the form of the revision; it cannot backfill missing participants and conversations.

A parameter-by-parameter reading

The surest way to avoid a false thresholds story is to separate the proposal's declared instrument effect from each visible scarcity control. On instrument effect, the delta is sharp: amendment with no obsolescence became replacement with obsolescence. On the central controls visible in both drafts, the comparison is continuity.

The Phase 1 maximum remained /15. The move into Phase 2 remained tied to no more than a /11 of non-reserved Final /8 space being available. The Phase 2 maximum remained /22. These three prefixes are easy to seize on because they provide an apparent numerical summary of the proposal. Yet their evidentiary value in this comparison lies precisely in their stability. They show that the seven-day version event should not be narrated as a new adjustment to those thresholds.

The eligibility and use conditions also persisted. Both drafts retained a 90-percent utilisation test for prior resources. Both retained the IPv6-resource precondition. Both retained the condition that address space used outside the region serve connectivity back to the AFRINIC region. Again, this does not settle whether those proposed requirements were wise, burdensome or effective. It establishes the narrower fact needed here: they were visible in Draft 1 and remained visible in Draft 2.

The reserve structure supplies the same lesson. A /16 was reserved for critical infrastructure, a /14 for new LIRs or End Users, and a /13 for unforeseen uses in both versions. No claim that Draft 2 newly created or resized those reserves can be supported by the comparison. Their repeated presence instead helps distinguish substantive continuity from the wider scope of the replacement document.

The four proposed responses stated by both drafts likewise remain aligned: reduce the Phase 1 maximum, require IPv6 resources, reserve address space for critical infrastructure and new entrants, and remove an explicit minimum. The consistency of the problem statement and response list shows that the proposal's policy programme carried across the week. Draft 2 repackaged that programme within a complete section.

None of these comparisons proves what happened beyond the page. A proposed maximum is not an allocation. A proposed utilisation threshold is not a refusal. A stated regional condition is not enforcement. A reserve described in a draft is not proof that space was set aside or issued. The available pages are strongest when used to compare texts and weakest when asked to demonstrate real-world consequences they do not record.

This separation protects both accuracy and significance. It prevents exaggeration of a parameter change that did not occur, but it does not trivialise the rewrite. Instrument topology can matter to anyone who must decide what language is controlling, what older text survives and how each provision entered the proposal. The week changed the map by which those questions would be answered, even though the principal numerical landmarks stayed put.

Reading a complete Section 3 without inventing a complete rewrite

A full replacement creates a visual temptation. Because Draft 2 displays a complete Section 3 while Draft 1 displays selected clauses, the later page looks more extensive. Readers may assume the visible increase represents a week of wholesale substantive creation. That inference is not warranted. The documents use different presentation logics.

Under the Draft 1 logic, a clause outside the four selected targets could remain operative through the existing policy without being repeated in the proposal. Under the Draft 2 logic, a clause intended to survive needed to appear within the proposed replacement. Definitions, statements about the current phase, the Phase 2 mechanism, repeat-request language, the planning horizon, eligibility criteria and reserve administration all contribute to a self-contained section. Their inclusion is consistent with replacement architecture even when individual language may be inherited.

This is why an origin table would be so useful. For each Draft 2 clause, it could identify one of three basic statuses: carried forward from the existing policy, modified from an earlier provision, or newly supplied in the second draft. A line-level comparison could then show the exact textual delta. A short reason entry could explain why each substantive modification was made. With those materials, readers could appreciate the coherence of the replacement without mistaking completeness for novelty.

The available version history offers only a high-level structural receipt. It tells readers that Section 3 was made completely new and that this converted the proposal into an instrument that would obsolete rather than amend the existing policy. That is enough to identify the broad change. It is not enough to allocate authorship dates to every sentence inside the completed section.

The eight-month allocation and assignment period illustrates the point. Draft 2 contains it and associates it with short-to-medium-term requests and equitable distribution. But because Draft 1 was selective, its failure to reproduce that part of the architecture cannot prove the period originated in Draft 2. The proper report describes the period as part of the complete restatement and leaves its precise textual origin unresolved.

The same method applies to definitions and current-phase mechanics. Their appearance in Draft 2 helps make the proposed instrument readable on its own. It may represent continuity, revision or a mixture. Without the base-policy comparison laid out sentence by sentence, one cannot confidently sort every provision. The larger lesson is simple: completeness answers “what text would the replacement contain?” It does not automatically answer “what text changed during this week?”

What the official record can prove

The archived AFRINIC pages have a specific evidentiary strength. They preserve how the proposal was presented: its identifier, dates, authors, Under Discussion status, selected wording, complete Section 3 and revision-history claims. The proposal-history page records the postings of Draft 1 and Draft 2. Together these materials make the seven-day sequence and the amendment-to-replacement delta verifiable.

Their limits are equally specific. A publication page cannot prove community consensus merely by hosting a proposal. A status label of Under Discussion cannot be read as adoption. A revision history cannot prove implementation unless it records implementation, which this one does not. Proposed criteria cannot prove actual allocation outcomes. The fact that AFRINIC published the pages does not reveal the authors' unrecorded motives or the views of mailing-list participants.

This is not a reason to disregard official material. It is a reason to use it for the propositions it actually establishes. Official records are authoritative evidence of the acts and claims they record. They are not a substitute for missing records of assent, institutional approval, operational action or measured effect.

The contextual analyses used here illuminate why those limits matter. They describe AFRINIC as a private number-resource service and coordination institution, distinguish registry recognition from sovereign title, and explain the operational exposure created when recordkeeping processes become uncertain. They also articulate a bounded model of uniqueness coordination: accurate records, continuity and orderly administration are valuable, but those functions do not make the bookkeeper a state.

Those propositions stand as substantive institutional analysis, not decorative “balance” against an official account. At the same time, they cannot be used as evidence that a particular threshold changed, that a hidden instruction drove Draft 2 or that the proposal produced a concrete loss. Keeping documentary proof and institutional analysis in their proper lanes produces a stronger account than either official-page literalism or unsupported suspicion.