Summary

  • The Number Resource Organization's own published description presents it as the coordinating body for AFRINIC, APNIC, ARIN, LACNIC and RIPE NCC, and places an "ICP-2 Documents" category in its policy section.
  • No criterion, assessing body, amendment clause, version marker or recognition record was obtained as published text in this review, so the operative content of the criteria document stays outside what this article can support.

What the record establishes

Four statements can be carried from the material this review actually obtained. The first is identity: the organisation's own site describes it as a coordinating body for five Regional Internet Registries and names them — AFRINIC, APNIC, ARIN, LACNIC and RIPE NCC (nro.net). The second is purpose as the organisation itself frames it: providing and promoting a coordinated Internet number registry system, and coordinating the registries' joint activities. The third is placement: the site lists an "ICP-2 Documents" category within its policy section.

All three come from the same page, and the evidence quality matters to how they are worded here. What this review holds of that page is a brief summary rather than the page's own text, so the statements are attributed to the organisation's published description as obtained here, and they are not presented as quotations from the site.

The fourth statement is narrower and differently sourced. ICP-2 is designated Coordination Policy 2, and its title is rendered as "Criteria for Establishment of New Regional Internet Registries" (encyclopaedia fragment). That source is a third-party reference work, and the passage retained is reported second-hand rather than read from the document itself. It is used here for one purpose only: naming the instrument.

Note what those four statements do not include. Nothing in the obtained material states a numbered or lettered criterion, says which conditions are thresholds rather than matters of judgment, names the body that assesses an applicant region, or records which body takes a recognition decision. Nothing states who may amend the text or by what procedure. Nothing carries a version, a date, or a status marker separating a text in force from a draft under consultation. Nothing reproduces the founding agreement among the registries, or any recognition decision.

What this review did not obtain

Six searches were made for this briefing. One returned usable material and yielded the two sources above; five returned nothing usable. Four document classes were attempted: the ICP-2 criteria text in force; the registries' founding instrument among themselves; the organisation's own ICP-2 index or policy page carrying a version, date or stated status; and a Regional Internet Registry's own institutional statement about recognition. None of the four was obtained as published text.

The pattern of failure is worth stating plainly because it explains why this article has a boundary where its argument should sit. Retrievals aimed at the criteria text and at a registry's own material returned provider capacity errors with unknown outcomes, and the single attempt reserved for the organisation's index page failed the same way. The gaps below are therefore retrieval facts dated to this review, not claims about what the bodies involved have chosen to publish.

Why an index entry is not a rule

The analytical point is not that the criteria text is absent from the world. On the material obtained here that cannot be said, and this article does not say it. The point is narrower and it is institutional: a document listed under a policy category on a coordination body's site is legible to a reader as a rule, while the authority that wrote and can revise the rule sits in an agreement among the registries themselves.

Consider who reads the same entry. An applicant region, a government weighing whether to support a regional registry proposal, and a network operator planning against an expected allocation all see the same listing. Each is being asked, in effect, to plan against conditions that look settled because they are catalogued. What this review could not establish is whether those conditions are thresholds with a named assessor, or expectations that five members may revise — a difference that decides whether a reader is looking at an obligation or at a description.

The cost of that confusion arrives late. If a criteria document is treated as operative before its text is read, the divergence between announcement and obligation surfaces at the moment a revision takes effect or an assessment is refused, which is precisely when the relying party has already committed. An index entry cannot be cited in a dispute or checked against a decision; an instrument can.

What this review did not obtain, and what would settle it

Three document classes would move the question from open to answerable.

First, the ICP-2 criteria text in force, carrying a version or date and an amendment clause. That would establish what an applicant region must satisfy, who assesses it, and whether the criteria bind or merely advise.

Second, the registries' founding instrument among themselves, with its purpose clauses, decision rules and member obligations. That would establish what authority the coordination body actually holds, and whether it can act on its own account or only by member agreement.

Third, a recognition decision record naming an applicant and an outcome. That would establish whether the criteria have been applied, and by whom the decision was taken.

Each is a public-instrument question rather than an investigative one. None of the three requires access to anything the bodies involved do not ordinarily publish.

Retrieval boundary

This briefing is scoped to material read on 18 October 2026. The criteria text, the founding instrument and a recognition record were not obtained as published text in this review. Any statement about what ICP-2 requires, who may amend it, or whether it has been applied therefore falls outside what this article can support, and the claims above are bounded to that extent. What remains is a dated observation, a named test, and a directory record whose own description of the organisation is treated here as a record description rather than as evidence about the organisation (directory record).