Mailing-List Volume Is Not Support

Summary

• Mailing lists widen access and create an authoritative public record. RIPE describes working-group lists as open and archived; APNIC requires pre-meeting list discussion to inform chairs; IETF procedure treats list review as essential when meeting conclusions are new or materially different.
• Raw message totals are structurally misleading. A few people can dominate a thread, replies can duplicate the full history, staff posts can inflate activity, and one unresolved objection can produce dozens of messages without adding breadth.
• Low volume is equally ambiguous. It may reflect agreement, rational delegation, low notice, language cost, fatigue, fear of conflict, uncertainty or a belief that the outcome is predetermined.
• Chairs need an argument map rather than a popularity count: unique contributors, relevant affiliations, independent operational contexts, propositions supported or challenged, evidence, unresolved objections, text versions and channel reach.
• Quantitative indicators can diagnose concentration and access, but no metric should become a vote. One well-supported technical or fairness objection can outweigh many simple endorsements; repeated posts do not gain weight through persistence.
• Consensus announcements should state what the archive proves and what it cannot: which reasons were considered, why objections were resolved or accepted, how missing perspectives were addressed and why progression is justified without claiming that list traffic represents the whole region.

The archive invites an arithmetic illusion

A mailing-list page looks measurable. It displays dates, senders, subjects and thread depth. Search can return every message. Compared with the ambiguity of a meeting room, the archive appears to offer a complete dataset from which support might be counted.

The appearance is deceptive. A post is not a person, an organisation, a constituency, an argument or a unit of preference. It is a transmission event. One person can send twenty messages; twenty people can sign one joint statement; a staff member can answer ten procedural questions without expressing a policy view. Quoted text can make a short reply look long. A subject split can make one debate look like several.

Even a carefully cleaned count answers only limited questions. It can show activity and concentration among visible contributors. It cannot show how many subscribers read silently, delegated attention, discussed internally or never saw the thread. It cannot tell whether a brief endorsement rests on deep examination or courtesy. It cannot convert an employer affiliation into a formal mandate.

The arithmetic illusion becomes politically attractive when a decision is contested. Supporters cite “hundreds of messages” to show momentum. Opponents cite the same volume as evidence of unresolved controversy. Both claims can be true at the level of activity and wrong at the level of consensus.

The archive's real strength is qualitative. It preserves claims, evidence, revisions and responses so chairs can judge whether objections were adequately considered. Counting can help navigate that record. It should not replace reading it.

Volume rewards persistence rather than breadth

Internet policy lists often have a long tail of participation: a few regulars post frequently, a wider group comments occasionally and many subscribers remain silent. This pattern is normal in specialised communities. It becomes distorting when frequent contribution is treated as multiple support.

A persistent entity may answer every objection because they authored the proposal or possess relevant expertise. Their replies can improve the text. They still constitute one visible viewpoint, perhaps informed by one institutional setting. Message count should not multiply their constitutional weight.

Persistence can also be strategic. Repeating a claim, demanding answers to minor points or replying immediately to every critic increases the apparent size and recency of a position. Other entities may leave because the cost of keeping up exceeds the value of another response. The archive then records endurance as dominance.

RFC 2418's warning that dominance is not determined by volume or persistence applies precisely here. Rough consensus permits chairs to conclude that a repeated objection has been adequately answered. It also requires them to recognise that a repeated endorsement adds little if it introduces no new reason.

Moderation should target behaviour, not viewpoint. Chairs can ask contributors to consolidate, use issue-specific subjects, avoid full quotation and pause repetitive exchanges. A summary can state that several messages expressed the same argument rather than counting each as independent evidence.

Breadth is better assessed through distinct contributors, independent contexts and reasons, with uncertainty. Even then, counts remain descriptive. A small number of affected experts may reveal more than a broad set of unreasoned approvals.

One thread can be one disagreement repeated

Long threads often grow because entities disagree about the meaning of one term. Each reply introduces an example, correction or alternative wording. The discussion may be productive, but the number of messages does not indicate the number of policy issues or supporters.

Chairs should identify the underlying proposition. Is the dispute about whether a requirement can be verified, whether it is fair, whether the registry has authority, or whether the benefit justifies cost? Messages can then be mapped to that issue and its subclaims.

Without an issue map, an unresolved point creates apparent breadth through recurrence. A critic raises it in several contexts; the author answers each; staff clarifies implementation; others comment on examples. A superficial count may show equal camps or overwhelming activity. The substantive question may still depend on one factual uncertainty.

Issue tracking also distinguishes movement from motion. Twenty messages can refine text and resolve the concern, or circle without changing either side's understanding. Chairs need to record what new evidence or concession occurred. A high-volume thread with no movement may justify focused facilitation rather than a support claim.

The public summary should preserve minority reasoning even after resolution. Saying “one objection generated forty messages” is less useful than stating the concern, evidence, response, textual change and whether the objector accepted the result. Continued disagreement can remain compatible with rough consensus if reasons were honestly considered.

The unit of deliberation is the proposition and its treatment. The thread is merely the container.

Quotation makes activity look larger than it is

Email convention encourages quoting prior text. Some clients include the entire thread, signatures and list footers. A reply containing one sentence may reproduce thousands of words. Archive page length, storage size and automated sentiment analysis can therefore misrepresent engagement.

Quotation also blurs authorship. A search may count a phrase each time it appears inside replies, making one original claim look widely repeated. A entity may quote an opponent to challenge them, while a simplistic tool classifies the message as support.

Governance reports should avoid word-volume metrics unless quoted material is reliably separated. Even then, verbosity is not intensity or quality. A concise operational example can be decisive; a long legalistic post can restate assumptions.

List guidance can encourage trimmed quoting, clear attribution and one issue per subject. Chairs can model the practice in summaries. Archive software might collapse repeated quoted blocks for readers while preserving the original message.

Machine-generated summaries require caution. They may flatten sarcasm, conditional support and changes of view. Any classification used in a consensus record should be verified by a human and, where practical, by the entity. The public archive remains authoritative.

Formatting seems trivial, but it shapes perceived participation. When a thread page displays repeated blocks and dozens of indented replies, readers infer magnitude. Clean presentation helps the community see the actual exchange rather than the accumulated weight of email syntax.

Staff messages are not community support

Registry staff participates in lists to publish proposals, answer procedural questions, explain current practice, provide data and clarify implementation. These contributions are essential and can make a thread highly active.

They should be classified separately from community support. APNIC's formal policy document states that RIR, ICANN and PTI secretariat staff do not participate in consensus. Other RIR structures differ in wording, but the governance principle is broad: institutional employees have defined roles and should not gain extra policy weight through administrative presence.

A staff clarification can support a factual claim without supporting the proposal. An implementation assessment may identify benefits and risks. A policy officer may post reminders or summaries. Counting all these messages as activity on one side would convert secretariat work into a bloc.

The same distinction applies to chairs when speaking procedurally. A call for comments, deadline notice or issue summary is not a policy position. If a chair speaks personally or for an employer, that capacity should be explicit and facilitation transferred if necessary.

Consensus reports should show staff, chair, author and other community contributions by role. Role classification is not a hierarchy of evidence. Staff facts may be uniquely important. It prevents institutional repetition from being described as popular breadth.

Separating roles also protects staff. Employees can answer openly without fear that every sentence will be counted as lobbying. The public can challenge evidence while recognising that the formal decision rests elsewhere.

Affiliation does not equal a mandate

Entities often use organisational email addresses and disclose employers. This helps readers understand experience and possible interests. It does not automatically mean the organisation authorised the position.

One network may have several employees on a list with different views. A consultant may work across companies. A person may speak as a community member rather than in their executive capacity. Counting organisations without this nuance can invent endorsements.

At the same time, affiliation concentration matters. Ten messages from employees of one large operator do not show the same range of operational contexts as ten messages from independent networks. Chairs should report estimated concentration cautiously and invite speakers to state capacity.

Formal organisational statements can be identified when clearly authorised, but they should not receive vote-like weight. A large member is not necessarily more entitled than a small one in an open policy process. Its evidence may be broader because of scale; that relevance should be explained rather than presumed.

Geographic labels need similar caution. A entity located in one economy does not represent it. A regional spread of contributors increases exposure to different contexts, but it is not a representative sample. National Internet registries or associations may have delegated roles, yet their mandates should be stated.

The record can therefore distinguish individual contributors, declared affiliations, formal statements and independent operational settings. This improves understanding without turning identity into arithmetic authority.

Subscriber count is not a denominator

Support percentages require a denominator. Mailing lists rarely provide a meaningful one. Subscription totals include inactive addresses, duplicates, staff accounts, automated archives, people following only selected topics and former entities. Some readers receive digests or access public archives without subscribing.

Delivery does not show attention. A message can reach a server and remain unread. Opening does not show understanding. Privacy-respecting analytics may reveal broad reach, but they cannot measure assent.

Using all members as a denominator is equally misleading. RIR membership and policy participation are distinct. Some policy communities are explicitly open beyond members. One organisation can have many staff and one membership. A member may delegate policy monitoring to an association or national registry.

Response rates can still diagnose notice when defined carefully. If a high-impact proposal receives comments from very few independent organisations after targeted outreach, chairs should acknowledge limited visible breadth. They should not claim a precise percentage of community support or opposition.

The absence of a denominator is not a fatal flaw in consensus governance. Rough consensus is not a referendum. It relies on open opportunity, relevant expertise, reasoned objection handling and transparent judgment. The mistake is borrowing the language of turnout without a population.

A defensible statement says how many visible contributors and contexts appeared, which channels received notice and what reasons were considered. It does not say “most of the community” unless a representative method actually supports that claim.

Low volume has many meanings

If high activity is not support, low activity is not opposition or consent. Silence is compatible with several states that the archive cannot distinguish.

Subscribers may agree and see no reason to repeat an adequate argument. They may trust known specialists, lack direct interest or wait for staff implementation detail. This rational delegation can make a healthy list quiet.

They may also be unaware, overloaded or unable to follow rapid exchanges. Language, time zone and technical complexity raise participation cost. A small operator may understand the effect but have no employee assigned to public policy. A person may avoid a hostile thread.

Some remain silent because they believe chairs have already decided, or because past contributions received no visible response. Others are uncertain and do not want a tentative view archived permanently. A proposal may be so obscure that affected people recognise the issue only after implementation.

Chairs should therefore treat silence as absence of submitted evidence, not a positive position. Where prior discussion establishes a strong case and notice is adequate, a quiet interval can permit closure. The conclusion is that no new material objection appeared, not that non-posters endorsed the proposal.

Low volume should trigger proportionate questions about reach and stakes. Targeted notice may be appropriate for consequential changes. A minor clarification need not mobilise the region. Accuracy requires matching the claim to the evidence.

Endorsement messages vary in evidentiary value

Messages saying “support,” “+1” or “I agree” provide some evidence that a person favours progression. They do not explain why or whether the person examined material objections. A hundred brief endorsements can show mobilisation while adding little deliberative substance.

Reasoned support is more useful. A entity can identify operational experience, explain which benefit matters, address a known cost and state whether revised text resolves concern. This helps chairs evaluate understanding and exposes assumptions to challenge.

Processes should not prohibit concise support. Not everyone has time or language confidence to write an essay. A short message may truthfully indicate direction. The chair should classify it as an expression of preference rather than an independent reason.

Campaign-style calls can inflate endorsements. Authors and organisations may ask colleagues to post identical text. Mobilisation is not illegitimate; affected groups should organise. The summary should note identical or coordinated messages without dismissing the underlying interest. A petition or formal vote has rules for counting; rough consensus does not become one accidentally.

Conditional support deserves precise treatment. “I support if the exception remains” is not support for a version that removes it. “I prefer this to current policy but share the implementation concern” contains both direction and objection. Binary coding loses the information chairs need.

The report should group endorsements by reasons and conditions, not publish a leaderboard. Support gains legitimacy when it demonstrates that entities understood consequences, not when the inbox fills fastest.

Objections are not weighted by frequency

An objection may appear once because the evidence is clear. It may recur because authors did not answer, because the text changed or because the objector refuses to accept a reasoned disposition. Frequency alone cannot determine which.

Chairs need an objection ledger. For each material concern, record the claim, evidence, affected text, response, revisions, status and whether the proponent believes it resolved. The chair then judges whether the community has adequately considered it.

One technical incompatibility can block progression despite broad support if the policy would fail its objective or cause serious harm. This is not minority veto; it is reasoned deliberation. Conversely, a repeated preference against a trade-off may remain after adequate consideration without preventing rough consensus.

The phrase “no new arguments” should be used carefully. A new example can materially strengthen an old concern. A text revision can make a previously answered objection relevant again. Chairs should explain why repetition adds no decisive evidence rather than using age as dismissal.

Objectors also have duties. They should consolidate, answer clarifying questions and state what change or evidence would address the concern. Flooding a list to create procedural doubt undermines others' access. Moderation can limit repetition while preserving the core issue.

Frequency helps locate contention. It does not determine merit. The ledger converts volume into an inspectable history of reasoning.

Thread subjects can hide the real distribution

Email subject lines shape archives. Entities may change the subject, remove a proposal identifier or begin a new thread. One policy discussion fragments across pages. Conversely, an old subject may continue after the topic shifts entirely.

Counts based on one thread therefore miss contributions or include irrelevant material. Chairs should maintain a canonical proposal page linking all known threads, meeting records, versions and assessments. Contributors can be reminded to include the identifier without having messages rejected solely for formatting.

Cross-posting creates duplication. A message sent to several lists may appear multiple times and reach overlapping subscribers. The public record should identify cross-posts rather than treating each archive copy as independent support.

Language-specific lists or translations pose another issue. Equivalent contributions may appear in separate archives, and summaries can privilege the dominant-language thread. Chairs need a combined issue map with links and translated summaries. A concern should not disappear because it arrived under another subject or language.

Private forwards should not be silently added to counts. If an organisation submits a formal statement for publication, post it with permission. Otherwise the consensus record relies on the open channels entities were told were authoritative.

Archive completeness is administrative work with constitutional effects. A chair cannot accurately assess reasons if the institution's own navigation fragments them.

Timing produces bursts that resemble momentum

Mailing-list activity is uneven. Publication, meeting deadlines, chair calls and revised drafts generate bursts. A high-volume week may reflect urgency rather than stable support.

Early responders often come from the established core, while people who need internal consultation reply later. If chairs infer direction too quickly, the first burst frames subsequent participation. A deadline immediately after a meeting favours attendees who already processed the debate.

Time-zone effects matter for short calls. Entities across a region receive and read messages at different times. Weekends, holidays and local events reduce usable response periods. Translation delays can create formally equal but practically unequal time.

The report should show activity over the full interval and identify major triggers: initial publication, staff assessment, revised text, meeting and last call. This reveals whether apparent support attached to the same version. Messages endorsing an earlier draft should not automatically carry to a materially changed one.

Chairs can use intermediate summaries to slow momentum without freezing it. A summary states what has been heard and invites missing perspectives. It should not imply a provisional winner unless the process calls for one.

Consensus is temporal but not instantaneous. The question is whether reasons remained stable after relevant evidence and exact text were available. A burst shows attention; sustained, answered reasoning supports a finding.

Civility affects who remains visible

A formally open list can become substantively narrow if conversation is hostile, repetitive or personal. Volume may rise while participation diversity falls. The most active contributors then appear to embody the community they have helped drive away.

RIPE's mailing-list guidance gives working-group chairs responsibility for guiding discussion and applying the Code of Conduct. It describes graduated intervention, from general reminders and private messages to moderation for continued or serious breaches. Importantly, list moderation exists to preserve participation, not decide policy.

Chairs should monitor whose participation stops after hostile exchanges, without inferring motive. A climate survey or private feedback route can reveal whether people avoid posting. Aggregate findings can inform improvements while protecting reporters.

Moderation actions relevant to a policy thread should be transparent enough that readers know whether message absence resulted from a technical or conduct restriction. The underlying substantive point should be preserved where it can be separated from harmful behaviour. A moderated person should continue receiving public messages where the rules provide.

Civility does not require mild disagreement. Entities can challenge evidence and institutional power directly. The line concerns personal attack, intimidation, discrimination and disruption. Overbroad moderation can suppress dissent as easily as under-moderation can.

A busy hostile list is not evidence of engagement success. Participation quality includes whether new and less powerful contributors can enter, be answered and remain.

Argument maps are better than sentiment tallies

The most useful summary is a map of reasons. It begins with the proposal's objective and exact version. It lists claimed benefits, costs, implementation findings, legal questions, affected groups, alternatives and material objections. Each item links to supporting messages and shows its status.

An argument map can include directional expressions without converting them into votes. It might say that several contributors favoured a change because it reduced ambiguity, while two distinct operational contexts raised transition concerns. The chair can then explain how revised text addresses the concern.

The map should identify evidentiary quality. A public dataset, staff statistic, documented case and personal prediction are different, though each may be relevant. Uncertainty belongs in the summary.

Competing causal accounts should remain visible. Supporters may agree on text for different reasons. If one expects conservation and another market liquidity, implementation review should test both rather than assume a single mandate.

Draft maps should be open to correction. Entities can say their conditional view was misclassified or a concern omitted. The chair retains responsibility for the final synthesis; public correction improves accuracy.

This approach scales better than reading every message anew at each phase. New contributions update issues, and repeated text does not multiply weight. Future reviewers can see why the decision was made without relying on thread volume.

Quantitative measures can reveal concentration

Numbers remain useful when matched to appropriate questions. Chairs can report total messages, unique contributors, median posts per contributor, share from the most active contributors, estimated independent organisations, first-time posters, staff and chair messages, language channels and response timing.

These measures diagnose concentration and access. If two people generated most traffic, the final report should not describe high volume as broad support. If many first-time contributors offered distinct operational evidence, confidence in reach may increase.

Metrics need methodological notes. Alias handling, affiliation uncertainty, cross-posts, automated messages and quoted text affect counts. Personal data should be minimised. Public archives already expose senders, but analysis should not create unnecessary profiles.

No threshold should automatically determine consensus. A concentration index cannot judge whether an objection is technically decisive. Unique organisations cannot vote by proxy. First-time posting can be a useful signal without making regular expertise suspect.

Longitudinal comparison may show whether participation is narrowing or whether outreach works. Similar proposal types can be compared cautiously. A controversial proposal will naturally generate more traffic than an editorial clarification.

The best quantitative statement is modest: “The discussion was active but concentrated among a small number of contributors,” or “Visible participation broadened after targeted notice.” Numbers discipline rhetoric. They do not supply legitimacy by themselves.

Chairs must account for what is missing

An argument map records visible evidence. A legitimate consensus assessment also asks which relevant perspectives are absent. This is not an invitation to invent views for silent groups.

The proposal's effects suggest who may need notice: small and large operators, national registries, legacy holders, new entrants, public-sector networks, researchers or specific language communities. Chairs can compare visible contributions with this impact map and invite missing experience.

Targeted outreach should be neutral. The notice explains the proposed change, known issues, deadline and public channel. It does not ask recipients to support a preferred outcome. Responses return to the common record.

If outreach produces no comment, the summary can state that fact. It still cannot infer agreement. If affected groups cannot be reached in time, chairs should calibrate confidence or recommend review after implementation.

Openness of subscription is necessary but limited public evidence. People cannot comment on a process they do not know matters to them. Regular policy entities often understand consequences earlier than ordinary members. The institution bears some responsibility for translating stakes.

Missingness is most important when a proposal shifts costs. A rule that simplifies registry handling but adds applicant burden should not be judged solely by staff and expert authors. The absent perspective may be the policy's central effect.

Meeting and list evidence should correct each other

Mailing lists favour asynchronous precision but can become slow, technical and repetitive. Meetings favour clarification and compromise but privilege attendance, fluency and schedule. Neither channel should be treated as inherently superior.

APNIC's published process explicitly asks chairs to consider list, meeting and remote opinions. IETF procedure requires materially new meeting decisions to receive list review. RIPE policy development occurs through meetings and working-group lists with public archives. These designs understand channel complementarity.

A busy list may contain a dispute that a short live exchange clarifies. The meeting result should return to the list with exact reasoning. A room gauge may suggest direction that written evidence later qualifies. Chairs must explain how the channels changed their understanding.

Counts should not be merged blindly. A person may speak in the room and post online. An online endorsement may refer to an earlier version. A remote entity may be represented in both meeting chat and list archive. The unit remains reasons, not appearances.

Channel conflict is informative. If the room strongly favours a proposal while the list contains substantial unresolved concerns, the answer is not to choose the louder venue. Examine attendance, timing, text and arguments. The conflict may reveal geographic, organisational or expertise differences that the policy must address.

A final finding gains strength when each channel tests the other's blind spots. It loses strength when institutions select whichever venue offers the preferred impression.

Consensus announcements need evidentiary grammar

Words such as “strong support,” “broad support,” “little opposition” and “community consensus” carry empirical claims. Chairs should use them only with stated evidence.

“Strong” may refer to reason quality, depth of operational evidence or directional intensity. “Broad” should refer to independent contexts, not message count. “Little opposition” can mean few objectors, few objections or weak unresolved reasons; the announcement should specify.

A good statement begins with the policy version and question. It identifies participation and concentration, summarises major reasons, records material objections and dispositions, notes staff and legal findings, describes missing perspectives and explains why the result satisfies the community's consensus standard.

Where the visible contributor base is narrow, say so. Consensus can still exist in an open expert process if notice was adequate and objections were handled. Modesty prevents the policy decision from being inflated into a regional plebiscite.

The statement should avoid treating non-response as assent or repeated posting as extra support. It can say that no new material concern appeared during a final interval, that earlier evidence remained controlling and that exact text was confirmed.

Evidentiary grammar makes appeals and future review possible. A critic can challenge whether an objection was actually resolved or whether breadth was overstated. Supporters can defend the decision through the record rather than inbox mythology.

Appeals should review synthesis, not recount messages

An appeal body asked to review consensus should not perform a fresh popularity count. It should examine whether chairs accurately identified issues, considered relevant evidence, handled concentration and missing perspectives, and explained unresolved objections.

The archive makes this review possible if summaries link to messages and versions. Appellants should identify a material misclassification, omitted argument, procedural barrier or unsupported breadth claim. Mere disagreement with the policy outcome is limited public evidence.

The reviewing body should recognise chair judgment. Rough consensus cannot be reduced to an algorithm, and reasonable people may weigh arguments differently. Review asks whether the judgment was reasoned and within the published standard, not whether another chair would write the same paragraph.

Where volume distorted synthesis—perhaps repeated endorsements were described as broad support or a concentrated thread was treated as representative—the remedy can be a corrected statement, targeted consultation or renewed consideration. The policy need not always be void.

Transparent review improves future practice. Chairs learn to preserve role and concentration data, entities learn to make objections specific, and institutions avoid overclaim.

Counting is tempting in appeals because it appears objective. It would replace one error with another. The question remains whether relevant reasons were adequately considered and the authority claimed matches the evidence.

Better list design lowers the volume incentive

Technical and procedural design can make public discussion easier to interpret. A proposal identifier in subject lines, canonical archive page, version links, issue tracker and concise chair summaries reduce repetitive clarification.

Entities should be encouraged to state whether a message introduces evidence, supports for a stated reason, raises an objection, proposes text or asks a question. Labels can help but should not become rigid submission barriers. Natural discussion remains important.

Digest summaries can welcome people who cannot follow every post. They must be neutral, linked to originals and open to correction. Translation of key summaries can widen access. The authoritative record should remain public and durable.

Rate limits or posting pauses may be justified against flooding, but they should be transparent and viewpoint neutral. A better first response is facilitation: ask frequent contributors to consolidate and give others time. Conduct moderation remains separate from policy merit.

Authors can maintain a living response document that links each material issue without rewriting the public archive. Chairs should ensure it does not become the sole account. Staff assessments and meeting records belong on the canonical page.

Good design shifts incentives from posting often to contributing something traceable. It lowers the cost of entry for readers and makes persistence less powerful.

The archive is evidence, not electorate

Mailing lists deserve defence. They allow people outside meeting rooms to participate, preserve institutional memory and force public reasons. In an era of closed chats and disappearing platforms, a durable open archive is a democratic asset.

Its value is undermined when institutions claim more than it can show. Message totals cannot identify a regional majority. Thread length cannot prove importance or consensus. Subscriber totals cannot serve as turnout. Silence cannot become approval.

The appropriate method is harder and better: read the reasons, map the issues, separate roles, identify concentration, test missing perspectives, track exact versions and explain objection dispositions. Use numbers to diagnose the forum, not decide the policy.

Entities share responsibility. Supporters should add reasons rather than duplicate slogans. Objectors should consolidate and state conditions for resolution. Frequent posters should make room. Staff and chairs should identify capacity. Organisations should not manufacture the appearance of independent voices.

The final consensus statement should be able to survive without a message count. It should show why the proposed action addresses an accepted problem, what material concerns were raised, how they were answered and why remaining disagreement does not require a different result.

A list is where the community leaves its evidence. It is not the community itself. Treating it as an electorate rewards the people with the most time, confidence and institutional support to keep posting. Treating it as a record rewards the people who contribute reasons capable of changing a decision. Rough consensus depends on the second.

This distinction also gives institutions a more durable public language. They do not need to promise that a specialised list mirrors every operator, member or economy. They need to show that participation was genuinely open, notice was proportionate to the stakes, material evidence was found, and disagreement received an answer that another reader can inspect. Where visible breadth was narrow, targeted outreach and later review can compensate without inventing assent. Where participation was concentrated but technically deep, the conclusion can credit that expertise while acknowledging its limits.

The resulting decision may sound less triumphant than a claim of overwhelming support, but it is more resistant to challenge because each word corresponds to evidence the archive actually contains. A policy community protects its authority when it refuses to turn traffic into a mandate.