Summary

  • Ratification begins with an instrument that identifies the drafting authority, the functions in scope, the institutions that can bind themselves, the rights that cannot be suspended through ordinary policy, and the questions that remain outside the charter. A general call for reform is not a drafting mandate.
  • The drafting commission must publish membership, appointment route, funding, clients, provider affiliations, litigation interests, resource holdings and recusal decisions. Expertise is necessary, but undisclosed dependence makes even technically sound language difficult to trust.
  • Consultation needs a versioned public docket. Every submission receives an identifier, attribution and disposition: accepted, partly accepted, rejected, referred, withdrawn or unresolved. Attendance, silence, mailing-list volume and repeated organizational submissions do not count as consent.
  • Voting should occur article by article before a final package vote. The approval record must distinguish operators and resource holders that supplied explicit mandates from institutional delegates, advisers and observers. A high aggregate majority cannot cure failure of a protected clause or an affected constituency threshold.
  • Ratification should require several independent keys: a qualified operator or holder threshold, approval by the institutions that will assume duties, confirmation from the IANA-facing coordination layer, independent rights and security review, and the legal instruments needed in the relevant jurisdictions. No single conference, RIR board, government or advocacy group can supply every key.
  • Recognition is not implementation. Before entry into force, authorised registries and technical providers must prove data export, state reconciliation, provider succession, RDAP continuity, RPKI recovery, court-order preservation, rollback and communications under failure conditions. A failed rehearsal blocks the affected phase.
  • Entry into force should be staged by function and cohort. Documentary duties can begin first, followed by shadow operation, limited voluntary service and only then broader activation. Existing holder state remains valid throughout; migration requires positive verification rather than an assumption that old records copied correctly.
  • Emergency clauses must name triggers, permitted acts, decision-makers, evidence thresholds, notice, review, maximum duration and restoration duties. Every emergency power expires automatically unless renewed through the ordinary ratification or amendment route.
  • Amendment, repeal and migration are part of ratification, not later housekeeping. The charter needs protected clauses, ordinary and heightened amendment tracks, a termination vote, a successor-selection procedure, portable records and a final reconciliation rule that prevents two authoritative settlements from surviving at once.
  • Number Resource Society may research the proposal, campaign for it, convene debate and represent members that give it specific power of attorney. NRS cannot ratify the charter for the Internet, approve operators, maintain authoritative state, execute migrations, recognise legal effect or decide appeals. Those acts remain with RIRs, IANA-facing institutions, authorised operators, courts, public authorities and independent reviewers acting under their own mandates.

Ratification begins where a declaration ends

Internet governance is good at producing declarations. A working group publishes principles, a meeting records broad support, a board welcomes further work, and advocates describe the result as a new consensus. Each step can be valuable. None proves that an operational institution accepted a duty, that a resource holder authorised a change, that a court will recognise a new custody arrangement, or that a successor can restore registration and security services during failure.

A continuity charter is especially vulnerable to this gap because its subject sounds constitutional. Words such as uniqueness, portability, due process and institutional recovery invite agreement at a high level. Disagreement returns when the text specifies who pays for standby capacity, which evidence a gaining provider receives, when a losing provider may entity, who can freeze a disputed instruction, which court order travels with a record, or how a compromised RPKI service is replaced. Ratification must force those decisions into the open.

The procedure should therefore treat a charter as a package of linked but separately attributable instruments. The public charter states the durable settlement. Participating RIRs and authorised operators adopt corporate and contractual duties. The IANA-facing layer accepts only defined state transitions. Courts and public authorities recognise custody, continuity and review arrangements through applicable law. Technical operators implement and test interfaces. Holders opt into service changes through authenticated instructions.

Independent reviewers receive jurisdiction from an actual instrument rather than from the moral importance of the text.

That arrangement is less dramatic than a founding conference. It is also more honest. Ratification is complete only when the public can identify which clause binds which actor, through which instrument, from which date, after which evidence, and with which route to challenge or exit.

Step one: publish a bounded drafting mandate

The first document should not be the charter. It should be the mandate to draft one. The mandate identifies its issuer, the authority the issuer actually possesses, the problem to be addressed, the functions in scope, the intended entities, the expected outputs, the timetable, the budget and the decision that will follow.

A sound mandate might authorize preparation of rules for continuity of number-resource registration, provider succession, evidence portability and dependent-service recovery. It should not quietly authorize new allocation policy, ownership adjudication, route control, sanctions policy or general Internet regulation. Those subjects may intersect with continuity, but intersection does not create jurisdiction.

The mandate should divide questions into three columns. The first contains matters the charter can settle directly because the participating institutions can bind themselves. The second contains matters requiring separate legal, contractual or technical recognition. The third contains matters expressly reserved to operators, courts, governments, standards bodies or existing global policy processes. This scope map prevents later drafters from treating every discovered dependency as permission to govern it.

The issuer must also state what ratification can mean. A board can approve duties for its corporation. A service provider can accept contractual conditions. A legislature or court can create legal effect within jurisdiction. A resource holder can appoint a provider for its own records. None can ratify on behalf of every network merely by using global language.

If several institutions issue the mandate jointly, each should sign a schedule identifying its contribution and limit. Joint sponsorship is not pooled sovereignty. It is a documented agreement to run one procedure while preserving the source of each entity's authority.

Step two: freeze an evidence baseline before drafting preferences harden

The commission should begin with a public baseline rather than with favored constitutional language. The baseline records the existing allocation and registration architecture, service dependencies, continuity arrangements, known failure modes, legal entities, contracts, data flows, key custody, funding, audit rights, review routes and unresolved disputes. It separates verified facts from claims and proposals.

Historical sources matter because they show why the present arrangement exists. RFC 790 records early central assignment practice. RFC 1174, RFC 1366 and RFC 1466 document scaling pressure and the movement toward distributed and regional management. ICP-2 records the criteria used to recognise Regional Internet Registries. RFC 7020 describes the hierarchical Internet Numbers Registry System and the boundary between registration and routing. These documents establish context, not a permanent answer to the ratification question.

The operational baseline should include current IANA service commitments, RIR coordination arrangements, mutual assistance, data custody, RDAP discovery, RPKI dependencies, reverse DNS, transfer processes, litigation and insolvency exposures. It should identify which claims have been tested. A statement that backups exist is weaker than evidence that a separately controlled operator restored from them. A stability fund is not the same as a lawful power to transfer functions. A public API is not a complete continuity export.

Every baseline item should carry a source, date, owner, confidentiality status and confidence level. Disputed claims remain visible with competing evidence. The commission should publish a cutoff date and a change log for later material events. Otherwise the factual foundation will shift silently whenever new evidence inconveniences a preferred clause.

The baseline is not a verdict against incumbents. It is the common record against which all proposals, including claims made by entrants and advocates, must be tested.

Step three: appoint a commission whose conflicts are legible

Drafting requires people who understand registry operations, number policy, organizational law, cross-border litigation, insolvency, cybersecurity, RPKI, RDAP, public administration, human rights and the daily constraints of networks. Expertise inevitably creates relationships. The response should be disclosure and balanced appointment, not a fiction that qualified drafters have no interests.

Each commissioner should disclose current and recent employment, clients, board roles, resource holdings, registry membership, provider investments, litigation, grants, political appointments, advocacy positions and close institutional relationships. The record should identify who nominated and appointed the commissioner, who pays their costs and which matters require recusal. Updates remain public throughout the process.

Composition should prevent any one bloc from controlling the text. Incumbent RIRs have essential operating knowledge but should not possess a veto over portability or succession. Prospective providers can identify entry barriers but should not lower the security floor for their own benefit. Large holders understand scale but cannot speak automatically for small networks. Governments bring public-law authority but not global operational command. Civil-society and academic entities add scrutiny but do not inherit operator mandates through independence claims.

The commission needs an independent chair, a technical editor and a public-docket secretariat. The editor records wording and provenance but does not decide contested policy privately. The secretariat preserves submissions, vote credentials, recusals, meeting records and versions. Removal procedures should cover misconduct, concealed conflicts and persistent non-participation without allowing sponsors to dismiss commissioners for an unwelcome view.

Observers may attend and advise, but the record must distinguish observers from decision-makers. A crowded room cannot be converted into a claim that every attendee authored the charter.

Step four: release a zero draft with an authority map

The zero draft should be deliberately incomplete. Its purpose is to expose architecture, definitions and unresolved choices before institutional positions become attached to polished prose. Each article should carry an annotation with five fields: the problem addressed, the actor expected to perform, the proposed source of authority, the evidence required for implementation, and the consequence of refusal or failure.

The draft needs a definitions schedule. Holder, operator, registrar, RIR, common coordinator, authorised provider, reviewer, custodian, emergency operator, current state, transfer, sponsorship change, legal hold and dependent service should not be used interchangeably. Ambiguous institutional nouns are a common route to accidental power.

An authority matrix should list every consequential verb. Who registers, authenticates, commits, publishes, certifies, preserves, stays, reviews, corrects, funds, activates, restores and terminates? A clause that says “the community shall ensure continuity” has no executable subject. The matrix should name the institution and the instrument that could lawfully give it the verb.

The zero draft should also contain brackets around unresolved alternatives. For example, one option may use an existing RIR-controlled common layer with functional separation; another may use a separately authorised neutral coordinator. Publishing alternatives prevents editors from presenting an early preference as settled consensus.

Every article needs a dependency note. A portability clause may depend on authenticated holder instructions, a common provider pointer, complete export, recognition by dependent services and a review route. Ratifying the headline without the dependencies would produce a right that fails at first use.

Step five: run consultation as a docket, not a sentiment exercise

Consultation should accept public submissions, protected evidence, oral testimony, technical demonstrations and statements from affected cohorts. Every contribution receives a stable identifier, date, attribution, represented constituency, declared interests and the draft provisions addressed. Anonymous public comments may inform risk discovery but should not be counted as mandates.

The commission should ask precise questions. Can the proposed export reconstruct a disputed holder record? What objection may a losing provider raise? Which law permits protected evidence to cross borders? Can a court restrain transfer without disabling ordinary service? What happens to hosted RPKI when registration sponsorship moves? How is a substitute paid after insolvency? General invitations to “share views on continuity” produce broad approval and little implementable evidence.

Consultation must reach actors who cannot attend global meetings. Small operators, public networks, universities, recent market entrants, organizations in sanctioned jurisdictions, holders with active disputes and networks using hosted security services face different risks. Translation, asynchronous submissions, regional hearings and limited travel support can improve access. Funding and selection should be disclosed.

Silence is not consent. Failure to comment may reflect lack of awareness, language, cost, litigation sensitivity or a belief that the proposal cannot affect the organization. Repeated submissions by related companies should be linked. A mailing-list majority is evidence of participation, not an electorate.

The docket should remain open through a stated period, followed by a short correction window for attribution and factual errors. Late material events can trigger a supplementary round, but sponsors should not prolong consultation indefinitely to avoid a difficult decision.

Power of attorney is the bridge between advocacy and member choice

An association can explain the charter and organize common positions. It cannot infer authority over a member's resources from membership alone. Where an operator or holder wants representation, the record should include a specific power of attorney or equivalent mandate identifying the principal, representative, subject, duration, permitted acts, confidentiality terms and revocation route.

This is the legitimate role available to Number Resource Society. NRS may research the draft, publish comparisons, convene members, campaign for safeguards and submit evidence. It may represent named members that give it power of attorney for defined consultation or review acts. The mandate does not let NRS vote unmandated resources, bind non-members, approve a provider, operate a registry, hold continuity data or convert a policy submission into an authoritative instruction.

The same rule applies to trade groups, government delegations, technical associations and coalitions. A representative may aggregate common reasoning while preserving the list of principals and differences among them. Confidential principal lists may be verified by an independent credential officer where public attribution would create legal or security risk, but the aggregate claim must state how many valid mandates were checked and what they covered.

Revocation matters. A principal should be able to withdraw future representation without erasing an earlier valid submission. The docket records the effective time and which acts remain attributed. This prevents a representative from turning a temporary consultation mandate into permanent governance authority.

NRS earns a positive place in ratification by making operator concerns legible and accurately attributed. It loses that place if it presents advocacy reach as a substitute for assent.

Step six: answer every material comment with a disposition

After consultation, the commission should publish a resolution table rather than a narrative claiming broad support. Each material submission maps to the affected article and receives a disposition: accepted, accepted in part, rejected, referred to another authority, withdrawn by the submitter or unresolved. The table identifies the reason, evidence used, commissioner recusals and resulting text change.

Grouping duplicative comments is efficient, but the grouping method must remain visible. Ten identical form submissions can demonstrate organized concern without becoming ten independent technical findings. One submission from a small operator can expose a decisive failure even if nobody repeats it. Ratification is not comment counting.

Unresolved objections deserve special treatment. The commission should identify whether the objection concerns fact, law, technical feasibility, cost, distributional burden or institutional authority. A factual dispute may require new evidence. A legal dispute may require an opinion from relevant jurisdictions. A technical dispute may require a prototype. A mandate dispute may require removal or narrowing of the clause.

Minority reports should be published with the final draft. A commissioner should not be forced to endorse text merely to preserve the appearance of unanimity. The report should state whether dissent affects a protected clause or implementation precondition.

The revised draft should be redlined against the zero draft and linked to the resolution table. Readers should be able to travel from final language back to the evidence and objection that shaped it. Text without provenance makes later amendment and judicial interpretation unnecessarily speculative.

Step seven: vote article by article before voting on the package

A single up-or-down vote invites logrolling. Entities may accept a defective emergency clause to secure portability, or reject useful evidence rules because they oppose the proposed coordinator. Article-by-article voting makes the source of support and dissent visible.

The voting instrument should list eligible decision-makers, credential requirements, represented principal, conflict status, voting period, quorum and challenge procedure. Ballots should be independently counted and preserved. Secret ballots may protect individuals in some corporate settings, but institutional votes that claim public authority should normally identify the institution and authorized officer.

Protected articles require separate approval. These include the singular authoritative state, holder continuity, provider exit, independent review, emergency limits, evidence custody, amendment, repeal and migration. Failure of one protected article cannot be cured by a high average across the document. The package returns for revision or proceeds without the failed function.

The final package vote answers a narrower question: are the approved articles coherent enough to operate together? It should not reopen rejected language through an annex or implementation note. Any material post-vote change triggers a defined reconsideration rule.

Votes should record abstentions and reasons where supplied. An abstention caused by a legal conflict differs from indifference. Institutions that lack authority to vote should be listed as observers rather than added to the denominator to enlarge the ceremony.

Approval needs several independent keys

No universal voting formula exists for number-resource governance, but a legitimate threshold should prevent one institutional class from ratifying duties for another. A multi-key design is stronger than one undifferentiated supermajority.

The first key is operator and holder support verified through eligible organizations or specific mandates. The record should be disaggregated by region, scale and resource type so a few large portfolios do not silently substitute asset volume for principal count. The second key is approval by the RIRs and authorised providers that accept operational duties. A charter cannot compel an institution merely because users favor it.

The third key is confirmation from the IANA-facing coordination layer that the required global state transitions and authoritative discovery can be supported under applicable policy and contracts. The fourth is independent review of rights, security, competition and continuity. Reviewers do not ratify policy; they state whether protected requirements and evidence thresholds were met.

The fifth key is legal recognition where implementation depends on corporate authority, insolvency treatment, evidence custody, data transfer, court orders, public procurement or statutory power. Recognition may come through several jurisdiction-specific instruments rather than one global law.

A proposed threshold might require two-thirds approval within each decision class plus a minimum number of regions and no failure of a protected-article review. The exact numbers are contestable. The essential design rule is that one government, registry board, provider bloc, resource-rich coalition or advocacy organization cannot supply every key.

The published certificate of ratification should list which keys were satisfied, by whom, on which evidence and with what reservations. “Adopted by consensus” is inadequate if the underlying procedure did not define consensus or record objections.

Legal recognition must attach duties to real entities

A charter is not a legal person. Its duties have to enter contracts, bylaws, corporate resolutions, service agreements, trust or escrow arrangements, legislation, regulatory orders and court-recognized procedures. The legal implementation schedule should map each article to the required instruments and jurisdictions.

Participating RIRs may need bylaw and contract changes to support portability, independent review or continuity custody. Authorised providers need enforceable service and exit duties. A data custodian needs lawful authority to hold and release protected evidence. A continuity fund needs ownership, triggers and controls outside the failing provider's ordinary reach. An independent reviewer needs jurisdiction, appointment protection and enforceable remedies.

Insolvency deserves explicit opinions. Can critical records, credentials and contracts transfer without becoming trapped in an estate? Can continuity funds be used for restoration rather than general creditors? Which court can authorize temporary operation? The Financial Stability Board's resolution framework offers only a comparison, but its emphasis on preserving critical functions, advance planning and legal powers shows why good intentions are limited public evidence at failure.

Cross-border data and sanctions questions require jurisdiction-specific schedules. A charter cannot promise portable evidence if applicable law prohibits the proposed transfer. The solution may involve local custodians, controlled access, verified summaries or court-supervised release. Concealing the conflict until crisis is not a solution.

Legal opinions should be public where possible, with privileged or security-sensitive material summarized. Reservations become entry conditions, not footnotes. A phase does not begin in a jurisdiction until the identified legal gap is closed or the scope is narrowed.

Technical acceptance proves conformance, not political legitimacy

Technical testing determines whether implementations behave as specified. It does not decide who should govern. The ratification record should preserve this distinction while making conformance a hard precondition for operational effect.

RFC 7020 supplies the core expectation of unique and accurate registration within the Internet Numbers Registry System. RFC 9224 demonstrates authoritative discovery for RDAP. RFC 6480 and RFC 8181 describe distinct certificate and publication functions in RPKI. These standards help define interfaces and observations. They do not appoint NRS, a new provider or a drafting commission to perform any function.

The conformance suite should test ordered state changes, stale-version rejection, holder authentication, provider-pointer updates, complete export, protected evidence references, dispute holds, rollback and public discovery. RPKI tests need separate key, certificate, entity, publication and relying-party observations. Court restraints and sanctions controls should be represented as state that survives an otherwise valid provider move.

Independent implementations should run the suite. A provider's private demonstration cannot prove interoperability. Failures and waived cases remain in the report. Waivers need expiry and cannot cover protected invariants such as duplicate current authority or loss of the only recoverable record.

Technical acceptance is therefore one ratification key with a bounded claim: these implementations can execute these approved articles under these conditions. It is not a credential to make law or represent operators.

Rehearsal is the final vote cast by the system itself

Before entry into force, the participating institutions should run adversarial continuity exercises. A tabletop is useful for discovering unclear roles, but at least one rehearsal must move verified test or bounded live state through the actual interfaces and independently controlled systems.

The first scenario removes the ordinary provider's management, primary systems and discretionary funds. The substitute receives only the material available under the charter. It must establish current state, preserve pending disputes, restore RDAP discovery, communicate with holders and return to ordinary service. The exercise records every manual dependency.

The second scenario assumes recent data corruption. The newest backup cannot be trusted. Independent witnesses, holder receipts, historical versions and protected evidence must reconcile a justified state. Blind restoration is failure.

The third scenario combines a court restraint with a holder request to change providers. Service should move only if law permits, while the restrained transfer or disputed ownership remains preserved. The charter must demonstrate that portability is not an eraser.

The fourth scenario tests hosted RPKI or publication failure. Registration state may be correct while certificates, manifests or entities create operational consequences. The authorised certification and publication operators—not NRS or the commission—must restore the relevant function under observed controls.

The fifth scenario activates an emergency clause and lets its clock expire. The exercise tests notice, independent review, restoration and automatic loss of temporary authority. A system that can activate but cannot deactivate emergency power has failed ratification.

Each failure maps to an article and owner. Cosmetic defects can receive bounded remediation. Contradictory authority, unrecoverable evidence, unauthorized credential action, inability to preserve a court order or dependence on the failed operator blocks the affected phase. Sponsors should not relabel a failed rehearsal as a useful workshop and proceed unchanged.

A readiness board can say no, but it cannot rewrite the charter

The final go-or-no-go decision should belong to a readiness board independent of the principal implementing providers. Its mandate is evidentiary: confirm that ratification keys, legal instruments, conformance results, rehearsals, funding and communications meet the approved conditions. It cannot waive protected articles or invent new policy.

Members should disclose conflicts and technical roles. Findings identify the evidence inspected, excluded material, residual risk and dissent. The board can approve a phase, approve with conditions already authorized by the charter, require remediation, narrow a cohort or refuse activation.

The implementing institution retains responsibility for its own lawful decision. A readiness certificate is not borrowed authority. If an RIR board, court or IANA-facing body must approve a step, it should issue its own instrument and cite the readiness evidence. This prevents the board from becoming an unelected supreme coordinator.

No institution should be penalized for a reasoned refusal outside a committed obligation. Refusal becomes part of the public record and may require redesign of the phase. Ratification is stronger when it can survive an honest no than when every doubt is pressed into ceremonial unanimity.

Entry into force should move by function, cohort and evidence

The first effective provisions should be documentary: conflict disclosures, service maps, incident reporting, standardized export requirements, review appointments and funding segregation. These duties improve visibility without moving authoritative state.

The second phase is shadow operation. Qualified substitutes ingest consented copies, calculate proposed state transitions, serve non-authoritative responses and rehearse recovery while existing RIR and IANA-facing state remains current. Differences are investigated rather than exposed to relying systems.

The third phase permits bounded voluntary service for uncomplicated records and excludes disputed holder changes, new allocations and high-risk cryptographic migration. Every entity gives an authenticated instruction. Cohort, resource types, functions and duration are capped. Automatic stops apply to conflicting state, evidence loss and failed restoration.

Later phases can add more providers, complex portfolios or distinct dependent services after evidence supports them. RPKI administration should not move merely because registration sponsorship did. Each added function receives its own authorization, rehearsal and rollback.

Grandfathering must preserve existing holders. Ratification should not force organizations to reconstruct identity or migrate on an arbitrary date. Existing state remains recognized until a verified transition, correction or lawful decision changes it. At the same time, indefinite grandfathering should not let incumbents avoid new transparency, export or review duties.

The entry notice lists effective articles, actors, cohort, interfaces, review route, stop conditions and next decision date. A global press release does not substitute for service-specific notice to affected holders and counterparties.

Emergency clauses need an expiry that does not depend on the emergency actor

Continuity systems need temporary power for compromised credentials, conflicting current claims, destructive unauthorized changes, cyber incidents and sudden provider failure. The charter should state the trigger, evidence threshold, permitted acts, decision-maker, notice, review, maximum initial duration and restoration duty for each emergency power.

Temporary action should preserve the last verified state where possible. It should not decide ownership, allocate new resources, erase history, punish criticism or turn a continuity operator into a permanent provider. Separate functions should be isolated separately: a compromised RPKI service does not automatically justify freezing ordinary registration support.

Every emergency grant expires by time, not by the actor's declaration that conditions are safe. Renewal requires fresh evidence and approval from an independent authority named in the charter. Repeated renewal reaches an absolute long-stop and must enter the ordinary amendment or legal process.

The public record should state activation time, legal or contractual basis, affected functions, aggregate cohort, reviewer and outcome. Sensitive attack detail can remain protected. After-action review determines whether the trigger was met, the action stayed within scope and the restoration duty was completed.

An emergency clause itself should sunset after a defined period unless re-ratified using evidence from exercises and real activations. Threats change, but temporary fear should not become the permanent source of institutional power.

Amendment should have ordinary, heightened and emergency tracks

Implementation details need to evolve. Data formats, cryptographic profiles, service measures and communication methods cannot wait for a constitutional convention every time. The charter should permit ordinary amendments to these details through public notice, compatibility testing, reasoned disposition and a defined approval threshold.

Protected subjects need the heightened track: current-state uniqueness, holder continuity, provider exit, independent review, evidence access, emergency limits, amendment thresholds, repeal and successor selection. A heightened amendment repeats the relevant consultation keys, legal review, article vote and rehearsal. The institution benefiting from the change cannot approve it alone.

Emergency amendments are narrower than emergency operations. They may temporarily adjust a technical parameter necessary to prevent immediate harm, but they cannot create a new permanent function or remove a protected right. They expire automatically and must be replaced, rejected or ratified through the ordinary or heightened track.

Every proposal should identify the problem, evidence, alternatives, affected cohorts, cost, security effect, backward compatibility, migration, dissent and review date. The docket links the changed text to its provenance. Silent changes through provider terms, implementation notes or API behavior are invalid where they alter charter duties.

Versioning must preserve one applicable settlement for each state transition. The record identifies which charter version governed an act. Migration between versions cannot leave providers applying incompatible authority rules to the same current record.

Repeal requires a destination, not merely a majority

A failed charter may need repeal. It may concentrate power, prove technically unworkable, lose legal recognition or become inferior to another settlement. Repeal without migration would destroy the continuity it was designed to protect.

The repeal article should identify who can initiate termination, the evidence required, the voting threshold, protected notice period, review and the conditions for accelerated action after catastrophic failure. It should distinguish repeal of an optional function from termination of the entire settlement.

A successor plan must identify the receiving institutions, legal authority, verified state, protected evidence, pending disputes, credentials, funds, service contracts, staff dependencies and public discovery changes. A complete dress rehearsal precedes cutover. Holders receive individual status and a route to challenge mismatch.

At the final commit, one settlement loses authority as the successor gains it. Parallel comparison may continue, but two systems cannot retain equal current authority over the same resources. A reconciliation report records exceptions and the institution responsible for resolving them.

If no successor is ready, repeal may remove disputed optional powers while preserving the minimum existing coordination under temporary bounded authority. Dissatisfaction does not justify rival allocation truth. The termination mechanism should make exit possible without making fragmentation easy.

NRS can campaign for repeal, publish evidence and represent specifically authorising members in the process. It cannot become the successor through advocacy, receive custody by default or execute the final state transition.

The public ratification ledger is the durable evidence of legitimacy

The charter needs a public ledger of procedure, not a blockchain slogan. The ledger contains the drafting mandate, baseline, commission appointments, conflicts, funding, drafts, submissions, disposition table, minority reports, credentials, article votes, package vote, legal instruments, conformance reports, rehearsal results, readiness findings, effective notices, amendments, emergencies and repeal decisions.

Protected material should be referenced through custody and integrity records rather than indiscriminately exposed. The public must be able to see that decisive evidence existed, who could inspect it, what claim it supported and which reviewer tested it. Privacy and security are compatible with procedural visibility when evidence classes are designed in advance.

Corrections remain visible. If a vote was misattributed, a conflict omitted or a test result revised, the ledger appends the correction and explains its effect. Legitimacy does not require a flawless historical presentation. It requires an honest record of how errors changed the decision.

The ledger should be mirrored and exportable under independent custody. No provider or advocacy organization should be able to erase the ratification history when leadership changes or litigation begins. Custody, however, does not confer the power to interpret or alter authoritative resource state.

A reader should be able to answer: who proposed this clause, who objected, what changed, who approved, what instrument made it effective, what test passed, which reservation remains, and how the clause can end? If those answers require personal access to insiders, the charter has not been publicly ratified in a meaningful sense.

Three decisions show why procedure matters

Consider a portability article supported by most consultation entities. One incumbent RIR identifies a court-order problem: a gaining provider in another jurisdiction may not preserve a confidential restraint. The commission should not call the objection resistance to competition. It records the issue, obtains legal analysis, drafts a controlled hold-transfer mechanism, tests it and seeks article approval again. The resulting delay is evidence of ratification working.

Consider an RPKI continuity article. A technical demonstration shows that registry data can move, but relying-party observations reveal a period of invalid route state during key transition. Registration portability may still enter force while the RPKI article remains in shadow operation. Package ratification does not require pretending every function passed at once.

Consider provider insolvency. The charter names a substitute and a reserve, but a rehearsal reveals that a cloud account and one vendor licence cannot transfer. The readiness board refuses activation. The authorised providers renegotiate the contracts, repeat the exercise and publish the result. A declaration of continuity would not have discovered the dependency.

These cases illustrate the central benefit of the procedure. Objection, partial approval and failed testing do not weaken the charter. They separate real authority from rhetoric before operators depend on it.

Ratification theater has recognizable failure modes

The first failure is premature consensus. Sponsors announce agreement after a general principles session, then negotiate the operative clauses privately. The cure is a versioned docket and article votes.

The second is constituency laundering. An organization cites membership, event attendance or mailing-list support as permission to bind operators. The cure is explicit credentials, power-of-attorney records and separate observer status.

The third is conflict concealment. Providers, brokers, litigants or governments shape text that affects their interests without disclosure. The cure is a live conflict register, recusals and minority reports.

The fourth is implementation by annex. A narrow approved article acquires broad powers through technical documentation, provider terms or emergency manuals. The cure is an authority map and a rule that material changes return to the appropriate amendment track.

The fifth is legal magical thinking. Drafters assume that a global label overrides corporate, insolvency, privacy, sanctions or evidence law. The cure is jurisdiction schedules and entry conditions.

The sixth is rehearsal capture. The implementing provider chooses clean records, normal management and friendly observers, then calls the demonstration proof of recovery. The cure is adversarial scenarios, independent control and publication of exclusions and failures.

The seventh is emergency permanence. Temporary authority renews itself until it becomes the settlement. The cure is external review, absolute long-stops and automatic sunset.

The eighth is repeal without migration. Opponents defeat the charter without identifying how current state and dependent services survive. The cure is to make verified succession a condition of termination.

A realistic ratification calendar is measured in decisions, not meetings

The process can be planned over roughly two years without pretending every jurisdiction or function will move on the same date. The first quarter establishes the mandate, commission, conflict register and evidence baseline. The second releases the zero draft and authority map. The third and fourth conduct regional, technical and protected-evidence consultation.

The next quarter publishes dispositions, redlines and minority reports. Article votes and the package vote follow only after unresolved mandate questions are narrowed. Legal implementation and technical conformance can begin earlier on stable provisions, but neither should assume final approval.

The second year is dominated by instruments and rehearsal: corporate approvals, contracts, custody, review appointments, funding, interfaces, shadow systems and adversarial exercises. Readiness findings determine which functions enter the first effective phase. Failed functions remain pending without holding documentary improvements hostage.

Deadlines should be public, but evidence gates matter more than calendar prestige. A delayed phase is preferable to false activation. Sponsors should record why time changed: missing legal authority, technical failure, incomplete cohort representation, unresolved funding or an external event. This creates knowledge for the next amendment or charter process.

Ratification ends with several dated acts, not one ceremony. The public certificate summarizes them; it does not replace them.

Conclusion: the procedure is the charter's first continuity test

A number-resource continuity charter should be judged first by whether its own adoption survives concentrated power. If an incumbent can veto the record, an entrant can lower the safety floor, a government can convert presence into jurisdiction, or an advocate can convert membership into operator consent, the drafting process has already reproduced the problem the charter claims to solve.

The alternative is demanding but legible. Issue a bounded mandate. Freeze an evidence baseline. Appoint a balanced commission and publish conflicts. Release an annotated zero draft. Consult through a versioned docket. Verify representation. Resolve every material comment. Vote article by article. Require independent approval keys. Attach duties to real legal entities. Test technical conformance. Rehearse institutional failure. Activate by phase. Sunset emergency power. Preserve amendment, repeal and migration from the beginning.

RIRs and IANA-facing institutions remain responsible for the authoritative coordination they are authorised to perform. Qualified providers execute only the services granted to them. Courts and public authorities supply legal force within jurisdiction. Independent reviewers decide only matters assigned by enforceable instruments. Network operators retain routing decisions and authenticated choices over their own service relationships.

NRS can make a valuable contribution by researching the record, campaigning for a fair procedure, convening participation and representing members under specific power of attorney. Its influence should be measured by accuracy and faithful representation, not by borrowed execution. NRS does not ratify for the Internet and does not become the registry, custodian, continuity operator, accreditor or tribunal after ratification.

The final proof is not the signatures on a page. It is a public chain from clause to authority, from authority to tested action, and from tested action to a safe route out. A charter that contains its own lawful correction and replacement is more than a declaration. It is a continuity institution capable of surviving the institutions that first adopted it.

Sources and analytical limits

The ratification procedure in this analysis is a governance recommendation derived from the documented history, technical standards and continuity comparisons. It is not a claim that the procedure has already been adopted, that one global vote can displace applicable law, or that any organization acquires authority merely by supporting the charter. Each operational duty requires its own lawful instrument, observed capability, independent review and safe route to amendment or termination.