Summary

  • Draft 3 did not rewrite the resource-recovery ladder. It amended the reported-complaint gate by requiring evidence, assigning AFRINIC staff to evaluate the facts and limiting repeat reported review of the same resource portfolio within 24 months.
  • Those changes were meaningful restraints, but evidence sufficient to open a review was not proof of non-compliance, staff evaluation was not independent adjudication, and the same private institution still sat across the downstream sequence of finding, correction, publication, appeal design and possible recovery.
  • A defensible private-registry protocol would tie each consequence to a particularised, verified defect; separate evaluation from adverse action; provide a proportionate cure and independent review with interim protection; preserve registry history; and permit no final action broader than technical accuracy, uniqueness, security metadata or recognition of a properly determined dispute requires.

L3 — The gate changed; the consequence ladder did not

The apparent paradox in AFPUB-2016-GEN-001-DRAFT03 is easy to miss if the proposal is read as a single block. Draft 3 made the entrance to one class of review more demanding. A community complaint could no longer stand merely because somebody had lodged it and it appeared to warrant investigation. It had to be backed by evidence, and AFRINIC staff were to evaluate the facts as appropriate to conducting the review. A further sentence withheld that reported-review route from a member whose same resource portfolio had already completed a full review during the preceding 24 months. These were concrete changes.

They addressed the costs of unsupported allegations and repeat scrutiny, and they deserve to be described as safeguards rather than dismissed as cosmetic drafting.

But they were changes to section 3.3.3, not a rewriting of the high-consequence recovery provisions. AFRINIC’s revision history records Draft 3 on 18 November 2016 and identifies the relevant update to section 3.3.3, alongside an acknowledgement update. A comparison with Draft 2 shows the material sequence in sections 3.4, 3.5 and 3.6 carried forward. The proposal still moved from a review and a finding of non-compliance to attempted contact and correction; if the situation could not be rectified, it contemplated publishing resources proposed for recovery for three months, during which the organisation could seek compliance.

It then contemplated possible recovery, removal of the former holder’s records from AFRINIC databases and possible reallocation under existing policy. It retained an appeal through an arbitration process defined by AFRINIC and annual reporting that would identify reviewed members and their level of compliance.

This distinction corrects the tempting but inaccurate shorthand that Draft 3 “rewrote the remedy ladder”. It did not. The important question is subtler: what happens when a more disciplined gate feeds into an unchanged chain capable of producing public exposure and irreversible registry consequences? The answer cannot be found by praising the gate or condemning the consequence in isolation. It requires tracing the vertical connection between them. A complaint may be adequately evidenced for the limited purpose of asking whether a review should begin, yet remain far short of proving a contractual or registry discrepancy.

If the opening evidence quietly becomes the basis for every later step, the new threshold merely gives concentrated downstream discretion a more respectable point of entry.

Draft 2 supplies the necessary baseline. Its reported lane permitted a member to request its own review, or allowed a community complaint that warranted investigation to prompt one. It did not say that the complaint itself had to carry evidence, identify staff fact evaluation or give a recently reviewed, unchanged portfolio a 24-month respite. Draft 3 therefore answered recognisable procedural objections. Evidence screening could discourage complaints made for tactical, competitive or personal reasons. Staff evaluation could stop an accusation from moving automatically into a burdensome exercise.

A cooling-off period could spare a stable portfolio from repeated demands for the same records after a full review had already been completed.

The restraint was nevertheless narrower than an institution-wide immunity. Draft 3 retained random, selected and reported review classes. The 24-month sentence appeared within the reported class. It did not expressly say that every random or selected review was barred, and it should not be read as though it did. Nor did “same resources portfolio” settle what would happen after any change to the portfolio. During the mailing-list discussion, an author explained that a new allocation altered the portfolio and that the new resources could be reviewed. A participant pressed for the chosen rule to be stated explicitly.

That exchange is useful evidence of the question the drafters were debating, but an author’s explanation cannot silently add precision that the published text lacked. The public words remained the public rule under discussion.

That ambiguity matters in practical rather than merely semantic terms. If an organisation completed a full review and then received a modest addition, would only the addition be in scope, or could the altered “portfolio” reopen everything? Would a transfer, return, renumbering or administrative correction be material? The text did not say. A 24-month protection reduces compliance cost only if the protected unit and the event that ends protection are knowable in advance. Otherwise the argument simply moves from whether a complaint warrants investigation to whether the portfolio is still the “same”.

A threshold rule intended to reduce friction can itself become a fresh site of discretion.

The gate also contained two decisions that must not be conflated. The complainant had to supply evidence. Then staff had to evaluate the facts as appropriate to conduct the review. Neither decision, by itself, amounted to proof that the member was non-compliant. Opening evidence answers a modest question: is there enough grounded information to justify asking for a closer look? An adverse finding answers a much more demanding one: after testing the evidence and the member’s response, has a specified obligation or registry fact actually been shown to be wrong?

The two stages need different records, different levels of confidence and different protections. Treating them as interchangeable would allow suspicion to mature into consequence without the intervening work that makes a finding reliable.

Draft 3’s downstream wording made that separation especially important. Its non-exhaustive indicators included an unjustified absence from the global routing table, breach of policy or agreement, a non-operating organisation that retained untransferred blocks and unauthorised transfers. These situations are not equivalent. They can raise different factual questions, carry different degrees of urgency and call for very different responses. Routing visibility may have technical explanations and, as AFRINIC’s own embedded staff assessment noted, its status as a policy requirement was disputed.

A stale contact record is not the same kind of defect as a proven duplicate claim. An organisation that has ceased operating presents questions unlike those raised by an active member contesting a contractual interpretation. A non-exhaustive list increases the need for a particularised finding because it denies both the member and the decision-maker the discipline of a closed catalogue.

The institutional status of the text also sets a firm limit on interpretation. Draft 3 was a proposal under discussion. The documentary record used here does not establish adoption, implementation, the opening of a review under it, recovery of resources or application to any named member. It would be wrong to narrate the proposal as an episode of completed enforcement or to use subsequent disputes as retroactive evidence of its operation or intent. November 2016 must be examined on its own terms: a published design for private registry action, carrying AFRINIC’s own staff and legal assessments, not a demonstrated history of use.

The date itself requires similar care. AFRINIC’s revision history and policy index record Draft 3 on 18 November 2016. The generic “Submitted” field on the Draft 3 page displays 9 August 2016, but that field is not the date of Version 3. The distinction is small but instructive. When a private registry proposes serious consequences, version identity is part of due process. The applicable text, its date and its precise delta should not depend on a reader resolving inconsistent page furniture. A particularised process begins with knowing which words were operative at which stage, even where the instrument never proceeded beyond discussion.

The official assessments add context without settling the normative question. AFRINIC’s staff assessment warned that investigations would substantially affect staff and reviewed organisations, identified the risk of unjustified complaints, questioned routing-table visibility as a requirement, said deletion of prior-holder records conflicted with practice, raised privacy concerns and estimated that creating an appeal process would take three to six months, followed by two months to appoint arbitrators. These observations reinforce the practical significance of the proposal’s design.

They also show that the institution itself recognised that review, disclosure, record handling and appeal were not costless administrative details.

AFRINIC’s embedded legal assessment characterised the registration service agreement as binding under Mauritian contract law and regarded reclamation following failed review as lawful. That is an institutional legal position, not an independent court judgment. It may illuminate how AFRINIC understood its private contractual footing; it cannot manufacture public-law power.

A later official agreement redline, which says it reflects changes from a January 2016 agreement, preserves baseline language about reasonably required information, investigation or review, contractual consequences, written termination notice, a 30-day response opportunity and remedial action. It is relevant evidence of contractual architecture, but it is not the signed agreement of every member and does not prove Draft 3 lawful, adopted or used.

The proper constitutional description is therefore deliberately modest. AFRINIC is a private, membership-based technical bookkeeper and coordinator for Internet number-resource records. Accurate records, reliable contact information, fraud controls and orderly coordination are valuable. None confers sovereignty. A registry does not become a government because its database entries matter to network operators, or a court because its staff can assess documents.

Policy discussion does not create legislative power; a membership contract does not create public jurisdiction; technical influence does not turn recovery language into confiscation authority. The institution’s legitimate task is to maintain a dependable record and coordinate unique identifiers within defined private arrangements.

That boundary does not forbid review. A bookkeeper may check whether its ledger is accurate, ask a member to explain an apparent discrepancy, verify a claim and correct a record. It can preserve evidence of a dispute and, within an exact contractual framework, take narrowly necessary steps to protect uniqueness or prevent incompatible entries. What it cannot properly do is let the practical importance of its ledger inflate those technical steps into open-ended investigation, punishment or adjudication.

Draft 3 should be judged by whether its entire sequence kept every burdensome act objective, evidenced, bounded, contestable and proportionate to a registry need.

On that test, the stronger gate was a real improvement but an incomplete one. It could reduce the number of weak complaints entering the machinery. It did not itself specify how a preliminary evaluation became a reasoned finding; what exact notice the member received; how a cure would be matched to the defect; whether appeal paused publication, deletion or reallocation; how conflicts would be managed in AFRINIC-defined arbitration; or why removal of historical records would ever be the narrow response to an accuracy problem. The gate changed. The consequence ladder did not. The unresolved issue lay in the joints between them.