Africa’s Long Wait for Its Own Registry

Summary

• AFRINIC’s decisive operational moment did not arrive as a clean constitutional switch: on 21 February 2005 it became the primary service desk while predecessor registries still supplied second-opinion review.
• The 1997-2005 formation period ran on three clocks at once: regional coalition building around AfNOG, legal and governance formation through steering and board design, and operational learning through hostmaster training, co-evaluation and phased migration.
• The first public policy meeting cannot be reduced to one mandate number: the available records offer more than 85 entities from 30 countries, about 120 registered entities including more than 30 African LIRs, and about 125 entities from 30 countries.
• The bounded finding is that delay built useful capacity and safer continuity, while also placing predecessor data, routines and personnel paths inside AFRINIC before its own membership could fully govern them; whether those paths later constrained choice remains an empirical question.

The handover that was not a single handover

Begin on 21 February 2005, not on the day of final recognition. On that date AFRINIC assumed the primary service role for the African region, while the existing regional Internet registries continued to provide second-opinion review until final approval. The AFRINIC Progress Update Submitted to ICANN presents the moment as a sign that the emerging registry could handle requests while still benefiting from external review. That detail is central because it breaks the neat myth of a founding day. Authority was divided in practice before it was complete in law.

The divided service desk shows why the question cannot be answered by asking when AFRINIC was born. A registry that receives requests, checks records, evaluates policy, communicates with applicants and relies on predecessor review is already exercising operational judgement. Yet it is not fully independent if its decisions still move through the established registries that previously served the territory. The handover therefore sat between two conditions: African operators had built enough capability to begin serving their own region, but continuity still depended on incumbent institutions.

That arrangement was not necessarily a weakness. Number-resource administration is a continuity service. A failed transfer could damage real networks, delay expansion, confuse resource holders and weaken confidence in policy administration. A cautious sequence protected applicants from a sudden break and allowed new staff to learn against live cases. The strongest defence of the 2005 arrangement is that second-opinion review reduced operational risk while the new registry gained experience.

The same arrangement also exposes the dependency question. If predecessor registries still reviewed requests after the African desk became primary, then inherited data, case habits, documentation expectations and staff judgement mattered before AFRINIC’s own community had complete authority over the service. The record does not prove that this constrained later representation or policy choice. It does show that operational authority moved gradually, and that gradual movement can embed routines before they are openly debated by the new membership.

That is the analytical problem. The eight-year wait from a 1997 registry proposal to the 2005 service transition was not empty time. It produced operator training, personal networks, a steering committee, an initial board, incorporation, a first public policy meeting, hostmaster preparation and a phased transfer. It also left different parts of the continent under APNIC, ARIN and RIPE NCC service relationships until AFRINIC could take over. Capacity and dependency accumulated together.

Three clocks instead of one origin story

AFRINIC’s formation is often easiest to describe as a chronology: 1997 proposal, 2000 AfNOG, 2001 board, 2004 incorporation, 2005 recognition. Chronology is necessary, but it is not enough. The important governance question is what each interval did to authority. Three clocks were running at once. The first was a regional coalition and training clock, beginning with the proposal at an INET workshop in Malaysia in 1997 and gaining practical force through AfNOG from 2000. The second was a legal and governance clock, moving from steering work to an initial board and then to incorporation in Mauritius.

The third was an operational learning clock, moving through training, co-evaluation and phased service transfer.

The three-clock view avoids two bad readings. One bad reading treats the long interval as mere delay, as if nothing happened until ICANN finally approved the registry. That misses the technical and organisational work that made final recognition credible. Another bad reading treats early meetings and training as a settled democratic mandate. That overstates what the evidence can show. Attendance, technical skill and institution building are not the same as full representation by every affected network.

The regional coalition clock matters because a registry cannot be created only by a charter. It needs operators who understand routing, allocation requests, address planning, policy norms and the daily consequences of bad administration. AfNOG’s 2000 workshop in Cape Town and its subsequent archive of operator training show a community building that capability. AfNOG’s stated concern with optimal use of number resources aligned closely with the practical needs of a regional registry.

The legal and governance clock matters because informal coalition work must become accountable authority. A steering committee and an initial board were not ceremonial details. They converted support into a body that could negotiate recognition, define a governance design and prepare an operating entity. The AFRINIC History dates key formation stages, including the 1997 proposal, the steering committee, the 2000 AfNOG consensus, the 2001 initial board, the 2004 Mauritius incorporation and April 2005 recognition. As a condensed institutional retrospective, that page gives useful dates but not the full minutes, dissent, host-country choices or affected-operator denominator needed for deeper claims.

The operational learning clock matters because registration service is learned through cases. A new staff member can read policy, but the real discipline lies in applying policy to requests, records and exceptions. The final application describes hostmaster training, co-evaluation, transition phases and distributed operations. Those details imply a practical apprenticeship before full autonomy. They also raise the question of whose standards, formats and habits shaped that apprenticeship.

The regional coalition clock

The 1997 proposal at an INET workshop in Malaysia is best understood as an opening claim: Africa should have its own regional registry institution rather than remain divided under external service arrangements. The proposal by itself did not prove continental support, create staff, fund an office or settle policy authority. It did something narrower but important. It gave the community a recognisable entity around which later training, advocacy and recognition work could gather.

AfNOG’s emergence from 2000 gave that claim an operational base. The AfNOG archive records annual workshops beginning in Cape Town in 2000 and describes a mission connected to technical coordination and optimal use of number resources. Training is often treated as background, but in registry governance it is a form of institution building. A registry needs people who can evaluate address requests, maintain accurate records, communicate policy and understand network growth. Operator workshops helped create the human capacity without which a regional registry would be an empty shell.

Training also creates durable personal networks. Engineers who meet repeatedly in workshops, list discussions and regional meetings learn who has technical judgement, who can mediate disputes and who can explain policy to operators with different levels of capacity. Those relationships can speed coordination. They can also create informal hierarchies before formal membership rules are mature. The evidence here shows training and meetings; it does not show whether a small circle later dominated decisions. That distinction must be kept firm.

The coalition clock therefore carries both a capacity claim and a caution. Capacity building was real and valuable. It made it more plausible for African operators to administer their own number resources rather than depend indefinitely on APNIC, ARIN and RIPE NCC. But participation in training cannot be treated as consent by every affected resource holder. A workshop entity is not automatically a member voter, board selector, policy author or representative of absent operators.

This is where the available doctrine of authority becomes practical rather than abstract. Doing work and attending meetings can support legitimacy, but they do not authorize claims over networks that were not present. The challenge for a young regional registry is to convert technical coalition into accountable membership without pretending that the first people in the room speak for everyone. AFRINIC’s long formation should be credited for building capacity and questioned for the evidentiary gap between trained participation and governed representation.

The 2000 starting point also matters because it came while predecessor registries still handled African requests. That means the coalition clock ran in parallel with external service. African operators could build a shared technical identity while many administrative records and request histories remained elsewhere. When AFRINIC later took over, it did not begin with a blank ledger. It inherited a live service environment shaped by established registry practice.

The legal and governance clock

The second clock moved from advocacy toward formal authority. A steering committee and an initial board gave the registry effort a negotiating face. The 2001 ICANN presentation on emerging regional registries listed the kinds of work expected from an applicant: ISP support, activity and transition plans, cooperation with RIPE NCC and ARIN, and self-sustaining funding. That list is not proof that every condition had already been met. It is evidence of the test the emerging registry had to satisfy.

Those requirements shaped behaviour. If recognition depended on visible support, transition planning, cooperation with incumbents and finance, then the emerging registry had to spend years producing documents, meetings, institutional design and evidence of capacity. The need to satisfy external expectations could discipline the project and make it more credible. It could also privilege actors who already had the time, language, travel ability and institutional familiarity to participate in recognition work.

The 2001 initial board recommendation sits in that context. A board is a bridge between community aspiration and legal capacity. It can sign documents, engage external bodies and guide preparation. It also concentrates authority at a time when membership may not yet be fully formed. That is not automatically illegitimate. Early institutions need interim structures. But later analysis should ask how board selection occurred, which subregions and operator types were represented, and what record was kept of support and dissent.

February 2004 incorporation in Mauritius gave the effort a legal home. The final application identifies the incorporation date and describes an emerging structure with distributed operations. Incorporation matters because it makes the registry an entity that can contract, hire, receive recognition and assume responsibilities. It also moves the project from a community initiative to a legal institution with a specific jurisdiction. This article does not assess Mauritius law, which belongs to a separate question. The point here is simply that legal personality became one of the dependencies in the transition.

The subregional board design deserves its strongest defence. A continent-wide registry needed to avoid being seen as captured by one state, one city or one technical clique. Subregional balance was an attempt to distribute authority across a large and diverse service region. Distributed operations across Mauritius, South Africa, Egypt and Ghana similarly avoided concentrating every function in one place at launch. These design choices were not decorative. They were answers to a legitimacy problem that any African registry would face.

The defence has limits. Geographic balance on a board does not prove that affected LIRs, small operators, academic networks, public-sector networks and commercial providers all had meaningful influence. Distributed offices do not prove distributed control. A legally incorporated entity does not prove that every predecessor record, policy interpretation or applicant concern had been reconciled. The legal clock created necessary capacity, but it did not erase the need for denominators.

The operational learning clock

The third clock is the least theatrical and the most important for daily service. AFRINIC needed hostmasters who could evaluate requests, apply policy, update records and communicate with applicants. The Updated AFRINIC Application for Recognition describes hostmaster training, a first policy meeting, co-evaluation and phased transition. It also describes distributed operations and support evidence. Because it is an applicant narrative, it must be read as a favourable account rather than an independent audit. Still, it gives the clearest view of how operational readiness was presented.

Hostmaster training was a necessary safeguard. Registry service is not only clerical. It requires judgement about policy compliance, address need, documentation quality, registration accuracy and communication with network operators. Training before recognition reduced the chance that applicants would face inconsistent decisions on the first day of independent service. Co-evaluation similarly allowed new staff to learn from live requests while predecessor expertise remained available.

Second-opinion review had the same continuity logic. A new registry serving a large and varied region could not simply announce independence and risk errors across active networks. Keeping incumbent review in place after 21 February 2005 gave applicants a layer of assurance while AFRINIC took the primary role. The arrangement made recognition a practical sequence rather than a ceremony.

Yet the learning clock also transmitted incumbent routines. A staff member trained through predecessor review may learn not only the written policy but also the old institution’s expectations about evidence, discretion, timing and applicant communication. That can be good if the old expectations are sound. It can be constraining if inherited habits fit some markets better than others. The evidence here does not show harmful constraint. It shows a plausible mechanism that should be tested with case data.

The operational clock also affects records. A registry’s data are not neutral furniture. They determine which holders are known, how resources are described, where contact points sit, what historical decisions can be reconstructed and how future requests are evaluated. If APNIC, ARIN and RIPE NCC served different African territories before AFRINIC, then the new registry did not inherit one uniform record culture. It had to integrate service histories that may have differed by incumbent institution and territory.

The fixed record does not contain case-level transfer logs, rejected requests, applicant appeals, migration errors or complete schedules of record movement. Without those files, it is impossible to say whether inherited records created measurable bias, confusion or delay. The right conclusion is not an accusation. It is a research requirement: if one wants to know whether the long wait shaped later representation and service, one must examine records and cases rather than rely on founding symbolism.

The predecessor footprint

Before AFRINIC’s service transition, APNIC, ARIN and RIPE NCC served different parts of the African region. That fact is enough to establish a divided predecessor footprint. It is not enough to build a complete table of territories, case volumes, allocation practices or contracts. The available materials do not provide that level of detail, and this article does not invent it.

The divided footprint matters because each predecessor relationship could leave a different trace. One network may have dealt with one registry’s request forms, another with another registry’s hostmasters, another with a different documentation expectation. Even if policies converged in broad terms, service experience could differ in timing, style, contacts and record format. When AFRINIC began serving the whole region, it had to move from this mosaic toward a single African service desk.

That is a continuity achievement. It is easy to underestimate the administrative value of a smooth migration. The absence of a public crisis in the provided records is not proof that every case went well, but the planned use of co-evaluation and second-opinion review shows an effort to avoid a hard break. In infrastructure governance, avoiding disruption is often a substantive success.

The same predecessor footprint complicates later legitimacy analysis. If resource holders entered AFRINIC with different histories, then early membership and service experience may have been uneven before any African policy debate began. Some networks may have known the registry environment well; others may have been less connected to the emerging forums. Some may have had staff who attended AfNOG or policy meetings; others may have remained ordinary applicants whose voice is harder to see in meeting reports.

The footprint also affects staff capacity. A new hostmaster team learning under multiple incumbent registries may gain broad comparative skill. It may also inherit a strong sense of how predecessor institutions framed acceptable documentation and risk. That tension is not a scandal. It is the ordinary cost of building a new institution from an old service environment. The governance question is whether the inherited habits stayed adaptable once AFRINIC’s own membership matured.

Because the records do not include case-level predecessor data, the footprint remains a map of dependency rather than a measured effect. A serious assessment would need historical APNIC, ARIN and RIPE NCC allocation and service records for African networks, matched to the transition sequence and later AFRINIC service outcomes. Without that, one can identify the mechanism but not quantify its distributional impact.

The first policy meeting and the denominator problem

May 2004 is a key moment because AFRINIC held its first public policy meeting. The meeting shows that public policy activity began before final recognition. It also shows the danger of treating meeting numbers as consent. The records offer three counts that must remain separate. The application chronology says more than 85 entities from 30 countries. The policy section of the same application says about 120 registered entities, including representatives of more than 30 African LIRs. The AfriNIC-1 meeting report says about 125 entities from 30 countries.

Those three figures may be compatible once registration, attendance and counting method are known. They may also reflect different moments or definitions. The available records do not reconcile them. Therefore the responsible approach is to preserve the versions, not select the most convenient number. More than 85 entities, about 120 registered entities and about 125 entities are not interchangeable.

The denominator problem is deeper than the numerical difference. A entity count is not an affected-operator denominator. A country count is not a member roll. Registered entities are not the same as resource holders. Representatives of more than 30 African LIRs are not the same as all African LIRs. A room can be broad, serious and valuable without being equivalent to continental authorization.

The meeting still matters. It created a public policy venue, brought operators and institutions together, and helped show that the emerging registry could convene regional discussion. The attendee categories reported in the meeting account included LIRs, founding members, supporting organisations, regional registries, ICANN and governments. That range suggests a serious institutional setting rather than a private drafting exercise.

But the meeting cannot carry the entire burden of legitimacy. To know whether it represented affected operators, one would need the registration list, attendance list, organisational affiliations, resource-holder status, speaking records, policy submissions and later membership data. One would also need to know which networks were absent and why. Without that, the May 2004 meeting is evidence of convening capacity and public engagement, not proof of universal support.

This distinction is especially important for later claims about path dependence. If the same people appear as trainers, founders, board members, policy speakers and early staff, that could create continuity. Continuity may be beneficial because experienced people keep the institution coherent. It may also narrow choice if new members face an established inner circle. The current materials do not provide the network analysis needed to decide which effect prevailed.

The staged-transition defence

Any fair reading must give the staged-transition defence full weight. AFRINIC did not rush from proposal to recognition with no preparation. It built AfNOG capacity, developed steering and board structures, incorporated, held a public policy meeting, trained hostmasters, planned transition phases and used second-opinion review. These steps are exactly what one would expect if the aim were to protect continuity while transferring authority.

The defence is strongest at the operational level. Address registration errors can have real consequences. A wrong record, delayed request or inconsistent decision can affect expansion plans, routing confidence and relations among operators. Training and co-evaluation reduced that risk. The transition was designed to make the new registry competent before it was fully alone.

The defence is also strong at the legitimacy-design level. The first policy meeting crossed 30 countries under all three available count versions. The subregional governance design attempted geographic balance. Distributed operations avoided placing every function in a single city. These choices speak to a recognition that an African registry needed visible breadth.

The defence should not be weakened by hindsight from later institutional controversies. This article does not use later disputes as proof that the formation period caused failure. That would be a retrospective error. The question here is narrower: which dependencies were visible by 2005, and which ones would need later measurement? A cautious transition can be both prudent and dependency-forming.

The defence also shows why a simple anti-incumbent story fails. APNIC, ARIN and RIPE NCC were not only obstacles to African autonomy. They were predecessor service providers whose cooperation helped make the transition safe. The 2001 expectations for emerging registries included cooperation with RIPE NCC and ARIN, and the 2005 transition update reflects incumbent support during second-opinion review. Without that cooperation, operational risk would have been higher.

The fair conclusion is therefore not that delay was harmful by definition. Delay bought time for training, institutional design, incorporation and testing. It helped produce a registry capable of being recognized. The harder question is whether the same delay also allowed incumbent data, routines and personnel relationships to become the default architecture before full African membership could evaluate alternatives.

What dependency could mean

Dependency in this setting should be defined carefully. It does not mean foreign control. It does not mean capture. It does not mean that the new registry lacked capable African leadership. It means that choices made during formation can narrow later options by making some routines, records, people and expectations easier to keep than to replace.

One dependency is data dependency. If records arrive from several predecessor registries, the new institution must trust, clean, reconcile and use them. Early service decisions may depend on the quality and structure of inherited records. Later membership contact, billing, policy communication and dispute handling may also depend on the inherited data. Without record-level evidence, the size of this dependency cannot be measured.

Another dependency is personnel dependency. The people trained earliest often become the people trusted most. They may be best qualified for staff roles, committee work and public explanation. That is rational. It can also concentrate institutional memory among those who had access to training, travel and recognition work during the formation years. The remedy is not to distrust experienced people. It is to measure how roles rotated, how new entities entered and how decision records were preserved.

A third dependency is procedural dependency. Co-evaluation and second-opinion review teach not only outcomes but ways of working. Staff may learn how to ask for evidence, how to read an application, how to escalate a difficult case and how much discretion is normal. These habits help avoid arbitrary decisions. They can also carry forward assumptions from predecessor registries whose historical service areas and applicant communities were not identical to the African region as a whole.

A fourth dependency is narrative dependency. If an institution presents its long formation as a heroic wait, it may understate the unresolved questions left by the wait. If critics present the same period as pure delay, they may erase the real capacity work that made autonomy possible. A better narrative treats the period as institutional accumulation: useful capacity, necessary compromise and measurable unknowns.

The available records support that middle position. They do not show later capture. They do not show that inherited routines dictated later governance. They do show enough to reject a blank-slate account. By February and April 2005, AFRINIC had already been shaped by training networks, board choices, incorporation, meeting participation, predecessor cooperation and live service review.

Representation before full membership power

Representation is the most delicate part of the founding period. A regional registry must speak for a region before all affected operators have joined, voted, objected or learned the new institution’s procedures. That is not a unique African problem. It is a general problem of institution formation. Someone must apply, negotiate and build before the final membership body exists in a mature form.

The AFRINIC record shows several attempts to manage this problem. There was regional advocacy from 1997, technical capacity through AfNOG, a steering committee, an initial board, a public policy meeting and subregional balance. These are meaningful forms of representation-building. They are not the same as a complete electorate.

The distinction among trainers, founders, board members, staff, attendees and resource holders is therefore essential. A trainer may have deep technical legitimacy but no formal mandate from a given operator. A founder may have institutional memory but not speak for every absent network. A board member may hold a regional seat but owe duties to the institution rather than to one constituency. A staff member may be expert but not an elected voice. An attendee may be present without carrying organizational authority.

This is why the May 2004 counts should not be converted into a mandate. More than 85 entities from 30 countries, about 120 registered entities including representatives of more than 30 African LIRs, and about 125 entities from 30 countries all show breadth. None shows the denominator of affected operators. None shows support and dissent by resource holder. None shows which people later held staff or governance roles.

Representation also intersects with predecessor service. Operators who had long dealt with an incumbent registry may have approached AFRINIC with different levels of trust, knowledge and administrative readiness. Some may have been close to the emerging community. Others may have been peripheral. The evidence does not identify those groups, so the article cannot rank them. It can only identify the question.

The key governance issue is not whether early leaders were sincere. Sincerity is not the metric. The issue is how a young institution demonstrates that early authority remains open to later members, new operators and absent voices once the service becomes real. That requires records of elections, policy participation, staff hiring, board selection and membership growth beyond the initial formation narrative.

Records and staff capacity as public-sector continuity

The overview topic includes public-sector continuity because regional registry service has a public-interest character even when performed by a private or membership entity. Number-resource administration supports network growth, routing stability and public confidence in the Internet’s addressing system. A new registry must therefore preserve continuity while changing authority.

AFRINIC’s formation shows this continuity problem clearly. The registry had to replace external service with regional service without breaking records, delaying applicants or creating confusion over who held authority. Training, co-evaluation and second-opinion review were continuity devices. So were distributed operations and staged recognition. They reduced the risk that African autonomy would arrive as administrative disorder.

Continuity has a cost. It often favours existing records, existing expertise and existing relationships. The people who know the old environment are the people most able to keep service steady. The documents that already exist become the base for new records. The habits that work during co-evaluation become the habits staff trust during independent service. That is not inherently bad. It is the ordinary bargain of institutional succession.

The public-interest question is whether continuity remains accountable. Did the new registry have enough transparency about inherited records? Were early service decisions auditable? Could operators challenge errors? Were transition cases tracked? Did the community know where predecessor judgement ended and AFRINIC judgement began? The selected materials do not answer those questions. They make them visible.

The 6 April 2005 IANA assessment and the 8 April 2005 ICANN Board resolutions gave final approval after provisional recognition and transition work. The Board action recorded completed transition and approval to provide registration and other services for the African region. That is the formal endpoint for recognition. It is not an endpoint for the institutional questions created by the transition.

Recognition confirms that the global coordination system accepted AFRINIC as the African registry. It does not measure whether the eight-year formation period created durable advantages for early entities, whether predecessor records carried hidden friction, or whether later members could revise inherited routines. Those are different evidentiary questions.

What the evidence does not show

The evidence boundary should be explicit. The records establish formation stages, training, meeting counts, incorporation, phased service transfer, second-opinion review and final recognition. They do not provide a complete list of affected LIRs from 1997 to 2005. They do not provide all support and dissent. They do not provide case-level allocation data from APNIC, ARIN and RIPE NCC for African networks. They do not provide migration-error files, rejection logs, appeals or operator complaints.

The records also do not provide the 1997 proposal text, steering committee minutes, the 2001 board-selection record or a host-country bid matrix. Without those, one cannot say how alternatives were weighed, who objected, or how host and board choices were justified. The absence of those records in this analysis is not proof that the choices were flawed. It is a limit on what can be claimed.

The first policy meeting records need the same caution. The three published counts should not be collapsed. More than 85 entities from 30 countries, about 120 registered entities including representatives of more than 30 African LIRs, and about 125 entities from 30 countries each describe a version of the meeting. Without registration and attendance rosters, the figures cannot be reconciled into one electorate.

Nor can later AFRINIC disputes be projected backward as proof that the founding delay caused governance problems. Later events may make historical questions more urgent, but they do not supply evidence for 1997-2005. A valid causal claim would need a trace from formation choices to later outcomes, with intermediate records showing how a routine, person, data field or rule constrained a decision.

The evidence also does not show the opposite: it does not prove that the staged transition had no later effects. Lack of measurement is not vindication. If inherited data, staff paths or predecessor routines mattered, they would be visible in case files, role histories, minutes and policy outcomes. Those files are the next evidentiary layer.

This balanced boundary is important because it prevents the article from becoming either a celebration or an indictment. The long wait had benefits. It created African technical capacity, a legal institution, public policy activity and a safer service transfer. It also created dependencies worth measuring. Both statements are compatible.

A falsifiable research agenda

The first research task is predecessor case data. For each African network served by APNIC, ARIN or RIPE NCC before AFRINIC’s takeover, the file would need service region, resource type, request history, contact accuracy, pending issues at transfer and subsequent AFRINIC handling. That dataset would show whether some predecessor paths produced more friction than others.

The second task is migration-error and appeal evidence. If the transition produced errors, delays, disputes or rejected requests, those should be tracked by date, predecessor relationship, resource type and resolution. If there were few errors, that would strengthen the staged-transition defence. If errors clustered in particular predecessor paths, that would sharpen the dependency thesis.

The third task is appointment and selection records. Steering committee minutes, board-selection materials, host-country bid documents and incorporation records would show how early authority was allocated. The aim would not be to relitigate the founding. It would be to distinguish transparent regional choice from undocumented continuity.

The fourth task is a network analysis of recurring entities. Names and affiliations from AfNOG workshops, steering bodies, board records, policy meetings, staff lists and recognition materials could be compared across time. The question would be whether the same people provided necessary continuity, whether new entities entered at meaningful points, and whether roles became too concentrated. Such analysis must distinguish individuals from organisations and organisations from whole subregions.

The fifth task is a representation file for the first membership years. That file would include LIR counts, voting eligibility, meeting participation, policy authorship, objections, board elections and staff recruitment after AFRINIC became operational. It would allow analysts to see whether early capacity opened into wider membership authority or remained concentrated among formation-era actors.

The sixth task is a document trail for operational judgement. Co-evaluation notes, second-opinion cases and early hostmaster decisions would show where predecessor advice influenced outcomes. A few well-chosen cases could reveal whether AFRINIC staff merely learned technique or adopted substantive interpretations that later shaped regional service. Without such cases, path dependence remains a hypothesis rather than a demonstrated effect.

This research agenda is falsifiable. It could show that the long wait mostly improved capacity and left little harmful residue. It could show that predecessor paths produced uneven administrative burdens. It could show that early entities remained dominant because they were the only people with relevant expertise. It could show that new members quickly gained influence. Each outcome is possible. The current public materials do not decide among them.

The bounded finding

AFRINIC’s long formation should be treated as institutional accumulation. From 1997 to 2005, African operators and allies did the work needed to create a regional registry: they built technical capacity through AfNOG, formed steering and board structures, incorporated an entity, convened a first public policy meeting, trained hostmasters and negotiated a staged handover from APNIC, ARIN and RIPE NCC service relationships.

That accumulation made the final transition safer. A registry that had trained staff, co-evaluation experience, distributed operations and incumbent second-opinion review was less likely to damage continuity. The 21 February 2005 primary-service moment and the 8 April 2005 final recognition were possible because years of preparation had already turned a regional claim into an operating institution.

The same accumulation means AFRINIC did not begin as a blank democratic page. It entered recognition with inherited records, predecessor habits, early leadership structures, trained personnel networks and unresolved denominator questions about who had participated, who had consented and who remained outside the room. Those facts do not prove capture or later failure. They do show why the founding period deserves scrutiny as a transition of authority, not only as a celebration of autonomy.

The most defensible conclusion is therefore double-sided. The wait was useful because it built capacity and protected continuity. The wait also embedded dependencies before AFRINIC’s own membership could fully govern them. Whether those dependencies later shaped representation, records or staff judgement is not answered by the recognition record. It is the next question, and it can only be answered with predecessor case data, migration evidence, appointment records and a careful map of recurring entities across the first years of African registry governance.