The First RIPE Address Policy and the People in the Archive

Summary

• RIPE-65 records that the RIPE NCC began acting as a delegated registry on 1 May 1992 and set out a hierarchical model in which European NICs and NOCs received blocks and reassigned numbers. That was not merely technical housekeeping: it allocated discretion, documentation duties and access paths.
• The public record identifies a visible chain of contribution linking Daniel Karrenberg, Jon Postel, Elise Gerich, global coordination bodies, RIPE entities, the Local Internet Registry Working Group and RIPE NCC staff. Those names and institutions are evidence of work, not a complete roll of everyone affected.
• Mailing-list volume exaggerates participation when repeated posts by one person are counted as separate voices. A more credible reconstruction groups senders by employer, affiliation, operational role and shared drafting activity, then distinguishes speaking, reviewing, deciding and implementing.
• RIPE 23 in January 1996 is a decisive institutional moment because its minutes say the community worked through disputed points in the circulated European registry document so consensus could be reached. The later ripe-136++ notice says the first five sections were approved there while revisions continued.
• The archive supports a strong claim about procedural ancestry and a weaker claim about representation. It shows how rules emerged and who was publicly visible; it cannot by itself establish how many independent networks understood, authorised or accepted every rule.
• Modern legitimacy should not be retroactively demanded from a 1992 coordination community. The constructive use of the history is to disclose the participation map, preserve dissent and explain which inherited powers now require wider review.

A policy appeared before a modern policy system

The phrase first RIPE address policy risks imposing later expectations on an earlier moment. In 1992 there was no mature regional registry association with thousands of contractual members, a familiar proposal template, a settled appeal path and years of published precedent. There was an urgent operational problem. The Internet was growing, classful address space was visibly constrained, routing tables could not expand without consequence, and a registration function centred elsewhere had to become more distributed. The rules emerged while the institution that would administer them was still taking shape.

RIPE-65 captures that compressed birth. It says that from 1 May 1992 the RIPE NCC acted as a delegated registry for IP network numbers to European NICs and NOCs. It did not present the NCC as a retail counter for every organisation. The stated policy was to allocate to service providers and registration organisations, which would then reassign addresses to users. Requests for scarce Class B space required justification. Class C blocks carried reporting and registration duties. Coordination with the US NIC remained explicit. Each provision joined technical conservation to an institutional chain.

That chain deserves political attention even though its entities described their task technically. Whoever decided what counted as adequate justification influenced which networks could grow in the expected manner. Whoever received blocks became an intermediary between a regional coordinator and an end user. Whoever maintained the registry could make a reassignment visible and globally coherent. The first rules therefore did not merely ration numbers; they distributed administrative standing.

The absence of a modern apparatus does not make those choices illegitimate. It changes the question. We should ask whether the record identifies the people and organisations who converted necessity into rules, how disagreement moved, which decisions were provisional, and what constituencies remained outside the visible conversation. That is a historical audit, not a trial conducted under standards invented decades later.

The delegated registry changed where power sat

Before delegation, an applicant might understand address assignment as a direct relationship with a central Internet registry. The European model inserted regional and local layers. RIPE-65 describes the NCC allocating to NICs and NOCs, and those organisations reassigning to applicants. Hierarchy helped administration scale and supported aggregation. It also created several places where judgment could be exercised: the global allocator, the regional coordinator, the local registry and the service provider facing the user.

Technical histories often present this as an obvious response to volume. Operationally it was persuasive. A coordinator near the networks could understand local conditions, communicate in relevant professional circles and distribute work. Blocks aligned with providers could help reduce routing announcements. Registration returned information needed to avoid collision and troubleshoot the network. The policy addressed a real collective-action problem.

Yet decentralising execution did not automatically decentralise rulemaking. A small number of connected operators could design a distributed administrative structure while retaining concentrated influence over its terms. The relevant measure is not how many local registries later existed, but how many independent organisations shaped the conditions under which those registries operated. An institution can have many implementers and few authors.

The distinction matters for inherited legitimacy. If a rule was initially written by a narrow group because only that group possessed the necessary expertise and connectivity, later adoption by a growing region may stabilise it. But silent continuation is not the same as a fresh mandate. Some early assumptions may become infrastructure before affected networks acquire practical means to question them. That makes documentary ancestry important: it reveals which choices were necessity, which were compromise and which simply arrived with the architecture.

The first address policy thus changed two maps at once. It reorganised address distribution across Europe, and it mapped authority across institutions. The former is visible in blocks and procedures. The latter appears through authorship, list messages, meeting minutes and the organisations behind entities.

The famous names are evidence, not the whole community

The October 1992 LIR Working Group message preserved by RIPE is unusually valuable because it shows a handoff in motion. Daniel Karrenberg forwarded a message from Jon Postel about guidance written by Elise Gerich and an implementation schedule associated with Claudio Topolcic. The enclosed material said it had been reviewed by the Federal Engineering Task Force on behalf of the Federal Networking Council, the co-chairs of the International Engineering Planning Group and RIPE, with general consensus supporting its recommendations.

This is a dense institutional scene. It links an individual RFC editor, an author at Merit, global and US coordination bodies, an international engineering group, RIPE and the newly operating NCC. The message also describes the texts as something like trial balloons: discussion could produce new editions; limited reaction might allow them to become official policy. That wording reveals both openness and ambiguity. Silence could move a draft toward authority even when no formal vote translated non-response into assent.

A poor history would convert the named entities into heroic founders and stop. A different poor history would treat the presence of US bodies as proof that Europe had no agency. The archive supports neither shortcut. People such as Karrenberg, Postel and Gerich performed visible, consequential work. RIPE entities reviewed and adapted guidance. The regional model was connected to global Internet administration, not isolated from it. Influence travelled through people who occupied more than one professional circle.

Names should therefore be attached to acts. Who authored a text? Who forwarded it? Who reviewed it institutionally? Who argued on the list? Who chaired a meeting? Who declared agreement? Who converted a document into daily registry practice? The same person may appear in several roles, but those appearances should not be counted as several independent constituencies.

The public names are the spine of the account. They are not proof that the surrounding body was representative of European network users, small providers, universities, governments or countries not yet strongly connected to RIPE.

Count organisations before counting messages

Mailing-list archives invite easy arithmetic. A researcher can download a period, count posts, rank senders and announce that a policy attracted extensive participation. That method confuses activity with independence. One engineer answering ten implementation questions does not supply ten separate mandates. Five colleagues from one operator may bring useful operational diversity, yet their institutional incentives may overlap. A message forwarded across lists can appear several times without creating another entity.

The first correction is entity resolution in ordinary historical terms: identify the person, contemporaneous employer or affiliation, operational role and any known drafting relationship. Then group contributions at several levels. The individual level shows labour and expertise. The organisation level shows independent institutional participation. The sector level indicates whether universities, commercial providers, research networks, public bodies and coordination organisations were present. The country level can reveal geographic concentration, although a country is not itself a unified interest.

No grouping is perfect. Early Internet engineers changed employers, spoke personally, held voluntary roles and represented organisations informally. An email domain may identify a network but not the authority under which the sender spoke. A consultant may advise several entities. A meeting contribution might be recorded without an affiliation. The answer is not to abandon counting; it is to publish confidence and avoid pretending that a clean spreadsheet can resolve ambiguous representation.

Message volume remains useful for a different purpose. It can show which issues consumed attention, who performed maintenance work, how quickly objections received answers and whether a text changed after criticism. It cannot establish the number of independent organisations merely by summing posts.

For the first RIPE address policy, the defensible question is therefore not “How many emails were there?” It is “How many distinguishable people and organisations performed which acts, over what period, with what evidence that their contribution affected the text?” That question produces a smaller number and a more honest history.

Participation had several grades

The word entity can hide crucial differences. A person copied on a message was not necessarily a reader. A reader was not necessarily a contributor. A contributor who raised a question was not necessarily a supporter. A meeting attendee may have listened without speaking. A working-group chair could shape the order of discussion without taking a position. RIPE NCC staff could supply operational facts and later implement a rule, giving their contribution a different weight from an outside comment.

A useful reconstruction separates at least six grades. Observers had access to the forum. Contributors supplied comments, objections or operational evidence. Editors changed wording. Reviewers assessed drafts on behalf of an institution. Deciders participated in the moment when agreement was declared. Implementers converted the agreed text into registry practice. The grades can overlap, but the overlap should remain visible.

This prevents two opposite errors. The first is understating community work by attributing everything to the document's named authors. Policies often absorb dozens of corrections that never appear in a byline. The second is overstating public authority by treating everyone who touched the discussion as a co-author of the final rule. A person may have objected unsuccessfully or commented only on presentation.

The distinction also clarifies the place of RIPE NCC staff. Technical staff had direct knowledge of requests, applicant behaviour, database limitations and coordination with other registries. Excluding that knowledge would have made the rules worse. But operational expertise does not alone answer how burdens should be allocated among users. Staff participation is necessary evidence, not a substitute for the affected community.

An archive can support this graded account if meeting minutes, message threads and document revisions are read together. The result is not a roll call of voters. It is a map of functions. That is more useful because it explains how a small visible group could produce rules that were technically grounded while still leaving representational questions for later generations.

Scarcity gave expertise exceptional force

Address policy in the early 1990s was inseparable from technical constraint. Class B space was scarce and often poorly utilised. Assigning many Class C networks raised routing and management concerns. Supernetting and aggregation were not abstract preferences; they responded to pressure on the network's ability to scale. Registration accuracy protected uniqueness. A person who understood these constraints could identify consequences that many affected organisations could not yet see.

Expertise deserved weight. A rule made by polling every connected organisation without technical explanation could have produced a popular but damaging allocation practice. The early community's credibility rested partly on its capacity to connect address assignment to routing behaviour and to coordinate implementation across institutions. This is why the visible engineers cannot be reduced to self-appointed gatekeepers.

But expertise creates a governance problem precisely because it is valuable. When only a few entities can translate a technical risk, their framing of available choices may define the decision. A conservation measure can be presented as inevitable even when several allocation designs could satisfy it. An administrative convenience can acquire the authority of network survival. End users may be invited to comment only after the range of options has narrowed.

The appropriate historical test is not whether non-specialists had equal influence on every technical detail. It is whether the specialists exposed the choices embedded in their recommendation: who would qualify, what evidence would be demanded, which intermediary would decide, how rejection could be questioned, and how experience would lead to revision. Technical necessity should explain the objective, not conceal the distribution of discretion.

RIPE-65 is valuable because it makes many operational choices explicit. The archive around it can show whether entities treated those choices as provisional and contestable. The stronger the technical case, the easier it should be to publish the reasoning without claiming that expertise itself constituted regional consent.

The end user was governed through an intermediary

One of the earliest policy's most consequential choices was its treatment of the organisation seeking numbers. RIPE NCC policy directed individual organisations toward their service provider rather than allocating directly in the ordinary case. That made the provider or local registry a gate through which need, documentation and reassignment passed. The arrangement supported hierarchy and aggregation, but it also meant the user might experience regional policy as a private service relationship.

This had representational consequences. Providers were more likely than end users to participate in a registry working group because address administration was central to their operations. They possessed staff who understood the forms, routing implications and list culture. Their participation could therefore be both rational and valuable. Yet the interests of an intermediary are not identical to those of the organisations it serves. A provider may prefer administratively efficient blocks, stable customer attachment or rules that reduce support costs. An end user may care about portability, independence, confidentiality or direct recourse.

The archive should be read for evidence of this gap. Were user burdens discussed even when users did not speak? Did entities distinguish the registry's conservation role from a provider's commercial relationship? Could an applicant contest a local registry's interpretation? Were exceptional assignments available, and under what authority? These questions turn a technical hierarchy into an accountability analysis.

The early model may still have been the best workable arrangement. The number of connected organisations was growing, and a regional coordinator could not treat every reassignment as a bespoke central decision. The governance lesson is narrower: implementation through many providers does not prove that many independent interests shaped the rule. Distribution of administrative work should not be counted as distribution of authorship.

When modern institutions invoke long continuity, they should remember who stood at the edge of that first chain. The end user was often present as the subject of a request described by someone else.

Geography in the archive is not the service region

RIPE's European origins and later service area cover a broad set of countries, languages, economic conditions and network histories. The visible archive of 1992 cannot be assumed to mirror that breadth. Connectivity itself affected the ability to participate. Organisations already close to research networks, international meetings and English-language technical lists had a lower cost of entry. Regions with fewer connected institutions could be governed by rules developed before they had comparable presence.

A geographic audit should therefore separate three things. The first is where a contributor was physically located. The second is the territory served by the contributor's organisation. The third is the set of users affected by the rule. A entity based in one country may operate internationally. A research network may connect institutions across borders. A coordinator may have a regional duty even if its office sits in one city. Simple flag counting can mislead.

Still, concentration matters. If most visible contributions come from a small cluster of well-connected countries and organisations, the policy's public ancestry should say so. The statement need not imply that every country required a delegate or that technical truth changes at a border. It identifies a risk that costs and practices elsewhere were under-observed.

Language is part of the same problem. English enabled a transnational engineering conversation, but it also privileged entities able to argue policy in professional English. Translation in the early 1990s would have imposed real costs. Informal national discussions may have informed a entity's contribution without leaving a trace in the central archive.

The honest historical conclusion is not that absent countries rejected the policy. Absence proves neither opposition nor assent. It shows that the surviving public record cannot support a strong claim of geographically balanced authorship. That limit should accompany any account of “the community” that approved the first rules.

Institutions spoke through people, often ambiguously

Early Internet governance depended on individuals trusted across several organisations. That social structure enabled speed. A well-known engineer could carry a concern from a network operator to RIPE, from RIPE to a global coordination conversation, and back into an implementable document. Formal letters of authority for every statement would have paralysed the work.

The same informality complicates later claims about representation. An email signature identifies an affiliation but does not reveal whether the sender had instructions, spoke from professional experience, or offered a personal view. A meeting minute may record an organisation beside a name because that was where the person worked, not because the organisation adopted the position. Consensus among present individuals can later be narrated as agreement among institutions.

Historical analysis should retain this ambiguity rather than resolve it in whichever direction supports a preferred story. Where a document says a body reviewed a recommendation, that is institutional evidence. Where a person posts from an organisational address, it is evidence of affiliation and participation, not automatically of organisational endorsement. Where minutes record a working-group decision, it is evidence of the forum's decision under its practice, not a referendum of every network served.

This precision protects the contributors as much as it limits claims made in their name. It avoids attributing a controversial position to an employer without evidence. It also acknowledges the personal labour through which institutions functioned. A policy does not write itself because several organisational acronyms appear in a paragraph.

The people in the archive should therefore be described with verbs and documented capacities: authored, forwarded, reviewed, chaired, commented, implemented. “Represented” requires stronger proof. That vocabulary produces a history in which personal trust and institutional authority can be seen together without being collapsed.

RIPE 23 marks a change from circulation to approval

The January 1996 RIPE 23 meeting provides a clearer decision point than the diffuse exchange of 1992. Its minutes say the draft European Internet Registry Policies and Procedures document had already circulated on the Local IR mailing list and attracted considerable discussion. The working group then aimed to isolate salient points, resolve differences and reach consensus so the document could progress. The minutes preserve subjects such as assignment procedures and provider-independent versus provider-aggregatable space.

That record matters because it identifies a sequence: circulation, continuing disagreement, focused meeting discussion, resolution and progression. It is stronger than inferring approval from a quiet list. Entities were not merely given an opportunity to comment; the forum undertook the work of identifying differences and reaching an outcome.

Even so, “the RIPE community” at a meeting remains a bounded body. Who attended? Which organisations spoke? Which positions were carried from the list? Did remote contributors have an effective way to influence the room? The minutes can document the forum's procedure without answering every question about the underlying constituency.

The later announcement of ripe-136++ adds another important qualification. It says the first five sections concerning distribution of public address space were approved at RIPE 23, while the revised document remained open for comments before official publication. Approval was therefore neither the beginning nor the end of the text's life. Some provisions had a clear meeting mandate; editorial or additional material continued to move.

This is the kind of granularity a legitimate institutional memory needs. Rather than saying a long document “was approved by consensus,” the record should identify which sections, at which meeting, after what circulation, and what remained open. RIPE's surviving notices make that distinction possible. Modern readers should use it.

Document families reveal change that titles conceal

The European registry rules existed across a family of numbered and revised documents. A stable title can make that family look like one continuous policy. In reality, each version may alter criteria, terminology, procedures or explanatory emphasis. The archive needs to be read as a sequence, not as a search for the oldest file bearing a familiar name.

Version comparison can show which concerns generated change. A requirement added after list discussion demonstrates influence more directly than the number of comments. A disputed paragraph preserved unchanged may reveal that a decision-maker rejected the objection, though reasons must be found elsewhere. A provision moved from guidance into mandatory language changes the burden even if the subject remains the same.

Document lineage also distinguishes ancestry from authority. RIPE-65 can illuminate the early delegated-registry practice. RIPE-104 and ripe-136 belong to later consolidation and approval. A 1996 statement should not be projected backward into May 1992 merely because later editors preserved its theme. Conversely, a provisional early instruction may become settled practice before it appears in a formally approved text.

The people map and the version map should be joined. For each material change, record who proposed it, which independent organisations discussed it, where the decision occurred and who implemented it. This is painstaking work, but it prevents historical credit from flowing only to final editors and prevents institutional authority from being attributed to a vague cloud of email.

The resulting account may show that no single “first policy” moment exists. There was an operational start, an early published procedure, global guidance entering discussion, years of registry experience, a circulated consolidated draft, meeting approval of major sections and later publication. Governance matured through those transitions. The archive is most informative when it preserves the transitions instead of compressing them into an origin myth.

Silence in 1992 cannot be decoded from 2026

Jon Postel's forwarded note conveyed an attitude common to early technical coordination: circulate a proposal, revise it if discussion reveals problems, and allow it to settle if reaction is limited. In a small trust network, this could be efficient. Entities knew one another's competence, shared urgency and could often reopen a decision when experience changed. The cost of formal balloting might have exceeded its value.

Three decades later, however, silence in that exchange cannot be assigned one meaning. Some recipients may have agreed. Some may have lacked the time to respond. Some may have regarded the authors as competent and deferred. Some affected organisations may not have received the message, understood its consequence or existed yet. Some views may have travelled privately. The archive does not supply a denominator against which non-response becomes support.

This is why historical legitimacy should be argued from more than silence. Evidence includes affirmative institutional review, recorded meeting agreement, implementation without unresolved operational breakdown, later revision opportunities and the growth of a community that continued to maintain the rules. Each adds a different kind of support. None turns every absent response into a vote.

The distinction is not pedantic. Modern policy forums still use last calls and quiet periods. Origin stories can normalise the idea that limited objection equals broad consent. The early community's practical methods should not become a universal theorem detached from their social scale and context.

The archive permits a modest conclusion. The guidance was publicly circulated among relevant technical institutions; named people and bodies reviewed it; European registry practice developed through RIPE; and major consolidated sections received a recorded meeting approval in 1996. It does not permit a precise claim about how many independent European organisations endorsed each early choice.

Implementation created evidence but not retroactive unanimity

Once the RIPE NCC and local registries operated the rules, experience became a powerful source of legitimacy. Requests were processed, blocks allocated, reassignments registered and operational problems exposed. A procedure that functioned under growing demand could earn authority through performance. Entities could see whether conservation and aggregation goals were met and could amend the text when practice diverged.

Performance matters because governance cannot rest only on participation. A perfectly attended meeting can adopt an unworkable rule. A policy initially developed by a narrow expert group can provide a reliable public service and later attract broader stewardship. Institutional legitimacy is cumulative.

But implementation does not prove that every design choice was accepted. Users may comply because addresses are essential and the registry is the only recognised allocator. Providers may adapt because there is no alternative route to globally unique numbers. Operational success can coexist with burdens that affected parties lack the means to contest. A monopoly coordination function requires especially careful distinction between use and consent.

The evidence should therefore be divided. Implementation data can support claims about functionality, timeliness, conservation and registration quality. Revision records can show responsiveness. Participation records can show who influenced changes. Complaint and exception records can reveal burdens. No one category should do all the work.

For the first European rules, this approach avoids two extremes. It rejects the claim that narrow origins invalidate everything built afterward. It also rejects the claim that decades of reliance erase the need to understand whose assumptions entered the foundation. A durable registry can honour its operational achievement by making that ancestry more, not less, visible.

The archive itself reflects institutional priorities

RIPE's public document store, mailing lists and meeting minutes are a substantial governance asset. They allow a reader decades later to follow drafts, dates, people and decisions with a precision unavailable in many institutions. Preservation lowers the cost of accountability. It also lets later entities recover arguments rather than reinvent them.

Yet an archive is not a neutral window. What was recorded depended on which communications used official lists, who wrote minutes, what files survived migration and which informal exchanges remained elsewhere. Meeting summaries compress discussion. Thread subjects can obscure related debate. Attachments and obsolete links may disappear. People with a habit of writing leave a larger footprint than those who influenced a room through speech or private review.

Institutional priorities shape discoverability too. A final policy document is easier to find than an objection buried in a long thread. A famous contributor's name attracts searches. An organisation that later became central may receive more historical attention than a small network that supplied one decisive operational example. The archive preserves inequality of attention as well as evidence.

Researchers should respond by triangulating. Compare document revisions with list discussion. Read meeting minutes for decision points. Follow announcements that distinguish approval from publication. Record broken links and missing attachments as gaps rather than silently omitting them. Where an organisation's participation is inferred from an email address, label the inference.

The reward is not perfect completeness. It is an auditable account in which another reader can see why a person or institution was counted and what kind of participation the record proves. The existence of the archive makes such discipline possible; it does not remove the need for it.

A contributor census needs a public method

If RIPE or an independent historian conducted a census of participation in the first address policies, the method should be published before the result. Start with a defined period, perhaps from the NCC's delegated-registry start in May 1992 through approval and publication of the consolidated sections in 1996. Identify the relevant lists, meeting records, numbered documents and global guidance explicitly cited by European entities.

Create one record per person, then attach affiliations by date. Do not merge names mechanically when initials or addresses differ. Preserve uncertainty around employer changes and personal contributions. Tag each appearance by function: author, forwarder, reviewer, commenter, chair, meeting speaker, editor, approver or implementer. Record the policy section or issue involved.

Next create an organisation layer. Count an employer once for independent-participation estimates even if several staff contributed, while retaining the individual labour total. Mark parent-subsidiary relationships, shared research consortia and coordination bodies so a reader can choose different independence assumptions. Separate commercial providers, academic networks, public institutions, coordination organisations and user networks without claiming that each sector has one interest.

Finally publish coverage limits. State which private correspondence is unavailable, how meeting attendance was obtained, whether non-speaking attendees are counted, and which countries lack reliable affiliation. Release both the strict count, based on documented active contribution, and a broader count including documented attendance or review.

Such a census would not produce a legitimacy score. It would answer narrower questions: how concentrated visible work was, which organisations bridged global and regional forums, and when participation broadened. Those facts would improve institutional memory and help later communities understand why specific assumptions entered policy.

Most importantly, the census would prevent message totals from masquerading as representation. The people in the archive deserve to be counted accurately, neither multiplied by their productivity nor erased behind the word community.

Dissent should be preserved as part of policy authorship

A consensus account often centres the wording that survived. That can make objections look like failed contributions. In fact, dissent may define exceptions, expose implementation costs or force proponents to explain a principle more clearly. Even when an objection does not change the final text, its reason belongs in the institutional record because later conditions may make it newly relevant.

RIPE 23's minutes describe an effort to isolate salient points and resolve differences. A richer account would connect each disputed point to its disposition. Was the concern accepted through an amendment? Rejected because evidence showed the risk was small? Deferred to operational guidance? Left open for a later document? The answer reveals what “resolved” meant.

Preserving dissent also limits retrospective overclaiming. A later policy document may say it was developed by consensus, while the decision record shows substantial concern around one section. That does not negate the outcome. Rough consensus is compatible with remaining objections. The legitimacy comes partly from demonstrating that the objection was heard, understood and answered rather than pretending unanimity.

For early documents, complete dispositions may not exist. The proper response is to identify the gap. Historians can quote the issue, show the textual change and avoid inventing a reason. Modern institutions can learn from the absence by maintaining concise decision tables alongside future policies.

The first address policy was authored not only by those whose preferred language survived. It was shaped by the questions that editors had to answer and by operational constraints that narrowed the choices. A people-centred archive should credit that negative contribution. Otherwise the history rewards only agreement and teaches new entities that objections vanish unless they win.

Founding competence is not permanent authority

The engineers who built European address administration possessed scarce knowledge and accepted practical responsibility. Their competence helped create a functioning registry at a moment when delay carried network costs. That achievement can justify substantial initial discretion. It cannot confer permanent authority on the same organisations, professions or inherited assumptions after the constituency changes.

Institutions often convert founding success into a cultural test. New entities are expected to learn the old language, attend the same forums and accept practices whose original reasons are no longer stated. Those who question a settled rule can appear ignorant of history. The archive should have the opposite effect: it should expose the contingencies that made a rule sensible and allow present entities to ask whether they remain.

For example, a provider-centred hierarchy designed for classful allocation and routing aggregation may not answer later questions about portability, transfers, contractual membership or address scarcity markets. The principle of accurate registration may endure while the appropriate intermediary changes. A confidentiality practice may require revision as public accountability expectations grow. Historical continuity should attach to objectives, not freeze every mechanism.

This is where a contributor census meets modern membership accountability. If a rule's visible authorship was narrow, that fact does not demand immediate reversal. It creates a reason for deliberate contemporary review, especially where the rule allocates discretion or burdens groups that were weakly represented at origin. The review can affirm the rule with stronger evidence.

Founders earn respect by solving the problems of their time. A living community earns authority by showing that inherited solutions remain justified to the people governed now.

Modern openness should be measured by effective entry

The current RIPE Address Policy Working Group states that anyone with an interest may observe, participate and contribute. That open door is an important development from the small coordination circles of the early 1990s. Public archives, remote access and mature documentation make entry possible from far beyond the original network.

Formal openness, however, is only the first measure. Effective entry depends on notice, language, time, technical context and confidence that a contribution can alter the result. A small operator may be free to join a list yet lack staff to follow months of discussion. An end user may not realise that an address-policy change affects its future options. A entity outside established professional circles may struggle to distinguish a question welcomed by the group from one treated as already settled.

The institution can measure these barriers without abandoning technical quality. Publish concise issue explanations beside full proposals. Identify which assumptions are open. Summarise positions without erasing minority reasoning. Report independent organisations contributing, not only message totals. Invite evidence from affected users through channels they already use. Provide remote discussion that does not depend on travel to a meeting.

None of this requires representational quotas or treating every opinion as technically equal. It ensures that expertise can be challenged with relevant evidence and that the term open describes more than the absence of a locked door. The early archive shows how much policy labour a small set of people can perform. Modern accountability asks whether the institution also makes room for those without the same time, status or historical familiarity.

The comparison should be fair. Today's region and stakes are vastly larger. A method adequate in 1992 may be inadequate now precisely because the institution succeeded.

The first rules need a constitutional annotation

A practical way to use this history would be a constitutional annotation attached to major inherited policies. For each principle, identify the earliest documented form, the operational problem it addressed, the forum that discussed it, visible contributors, approval point, later revisions and current justification. The annotation would not alter the rule. It would make its authority legible.

Consider the hierarchical registry model. The annotation could explain routing aggregation, administrative scale and the 1992 delegated-registry start; identify the early document and related global guidance; state that participation was concentrated among technical and registry institutions; then show how later RIPE decisions adapted the principle. A present reader could distinguish enduring coordination needs from historical implementation choices.

Annotations should include uncertainty. If meeting attendance is incomplete, say so. If institutional endorsement is documented but individual reviewers are not, preserve that distinction. If a provision appears in practice before formal approval, identify both dates. If later revisions substantially changed meaning, do not describe the current text as simply adopted in 1992.

This form of institutional memory would improve debate. Proponents of continuity could point to the original problem and accumulated performance. Reformers could identify mechanisms whose context changed. New entities would not need to master decades of scattered records before asking an informed question.

The annotation would also honour the people in the archive more accurately. Their work would appear in context, with specific contributions rather than ceremonial founder status. Institutions that reviewed, disagreed and implemented would be visible. The many affected organisations not documented as entities would not be silently enrolled as authors.

What the archive can prove

The surviving evidence supports several firm findings. The RIPE NCC assumed a delegated European registry role on 1 May 1992. Early procedures allocated blocks through NICs and NOCs, demanded justification for scarce address categories and required registration of reassignments. Global guidance entered RIPE discussion through named people and institutions. The Local IR forum became the venue for developing European registry policy. By RIPE 23, a circulated consolidated document had generated enough disagreement to require focused resolution, and major sections were approved before later publication work continued.

The evidence also supports a finding of concentrated visible labour. A limited circle of technically connected people and organisations authored, forwarded, reviewed, debated and implemented the rules. Their contributions crossed institutional boundaries. The record is rich enough to reconstruct many acts, but not broad enough to claim that every affected country, sector or end-user interest participated independently.

Several stronger claims remain unsupported. The archive does not provide a complete denominator of recipients or readers. It does not turn silence into affirmative endorsement. An organisational email address does not prove a formal mandate. Repeated messages do not create repeated independent voices. Later reliance does not show that every initial choice was freely accepted.

These limits do not diminish the archive. They make it more useful. A record that distinguishes proof from inference can sustain institutional trust because it does not ask readers to accept a founding legend. It shows a real community solving urgent problems with the tools and relationships available, then gradually building more durable forms of decision.

The strongest history is therefore neither celebratory nor prosecutorial. It recognises competence, concentration, adaptation and absence at the same time.

From origin story to accountable inheritance

The first RIPE address policy mattered because it made distributed Internet growth administratively possible. It translated scarcity, aggregation and uniqueness into duties for a new regional coordinator and local registries. People did that work: authors who framed the problem, engineers who connected it to routing, entities who argued over procedures, chairs who found a decision and staff who made the rules operate.

Their public record is unusually durable. It lets us see a delegated-registry start, a global guidance exchange, a working-group conversation, a multi-year document lineage and a meeting that approved key sections. It also shows why “the community decided” is too coarse. Community can mean named contributors, institutions that reviewed guidance, people present in a room, subscribers who did not entity, registries that implemented the result or the much larger population eventually governed by it. Those groups overlap but are not identical.

An accountable inheritance begins by naming the relevant group for each claim. It counts independent organisations before messages. It separates affiliation from mandate and participation from support. It preserves dissent and version changes. It admits where informal conversations or absent records prevent a full census. Then it asks the present community to justify inherited discretion under present conditions.

This approach does not burden a 1992 institution with a 2026 constitution. It recognises that legitimacy can mature. Urgent expert coordination can create a valid beginning. Public preservation, operational performance, wider participation and periodic reconsideration can strengthen it. Mythologising the beginning weakens that process because it turns later questions into challenges to identity.

The people in the archive deserve a better legacy than being invoked as an undifferentiated consensus. They should be remembered for the specific work they performed, the constraints they confronted and the institutions they helped create. The people outside the archive deserve equal honesty: their silence cannot be counted as a voice. Between those truths lies a credible account of how European address policy began and how its authority should continue.