Summary

  • Resolution 201110.135 led quickly to a tangible draft: in November 2011, a four-name committee presentation set out proposals on election timing, NomCom composition, voting methods, candidate rules, counting, disclosure and ballot custody.
  • The 2012 bylaws later separated NomCom from a staff Election Committee and made election procedures more explicit, but the surviving record does not establish a clause-by-clause path from committee proposal to constitutional adoption or practical implementation.
  • The correct verdict is partial, not binary: the mandate and draft output were real, and later constitutional change is visible, yet the missing final report, response ledger, adoption decision and evaluation leave institutional accountability materially incomplete.
  • This was governance work inside a private registry company. AFRINIC is only a private bookkeeper and coordinator, with no sovereign, regulatory, police, punishment, confiscation or adjudicatory authority; better internal elections can discipline its agents but cannot enlarge that narrow role.

The slide that prevents an easy dismissal

The most revealing document in this story is not grand. It is a six-page presentation from AFRINIC-15 in November 2011, short enough to be mistaken for a meeting aid and detailed enough to frustrate anyone tempted to say that the election review vanished immediately after it was announced. Its title was “Election Review Process”. It named four people—McTIM, Gift, Adiel and Mark—and described the process as a draft open to community input.

Then it moved briskly through the machinery of an election: a countdown to polling day, the composition of the nominations body, paper and proxy voting, an electronic window, candidate conduct, counting, disclosure and custody of the ballots.

That presentation is the fixed point around which the rest of the account should turn. It proves that the committee created by the Board in October did produce a concrete public output within weeks. It also shows why speed is an inadequate measure of institutional repair. A draft may be thoughtful, and it may influence later rules, without becoming the final position of the company. Unless the record preserves what happened after consultation—what was accepted, changed, rejected or deferred—the historian can see activity but not the complete chain of authority inside the corporation.

This distinction matters because the later constitutional text looks, at first glance, like a satisfying ending. AFRINIC's 2012 bylaws contained separately titled provisions for a Nomination Committee and an Election Committee. They assigned functions to each and recognised paper ballots, electronic voting and proxies. They also provided a way to resolve election-day questions not expressly covered by the rules. It is tempting to set the November draft beside these articles, observe the shared subject matter and conclude that the committee's work was adopted.

Tempting, but not proved. The available documents contain no final committee report, no ledger of comments, no Board decision disposing of each recommendation, no comparison mapping proposals to clauses, and no later evaluation of how the arrangements performed. The record therefore supports neither of the easy stories. It does not support the claim that the review was empty. Nor does it support the claim that the review succeeded as designed. What it supports is more interesting and more useful: follow-through was partly visible, while the institutional memory needed to explain that follow-through remained incomplete.

The narrowness of AFRINIC's role should discipline the analysis from the outset. AFRINIC is only a private bookkeeper and coordinator. It has no sovereign, regulatory, police, punishment, confiscation or adjudicatory authority. Its internal election processes matter because they shape oversight of a company whose registry records are operationally important to networks. They do not constitute public elections, create jurisdiction over outsiders or turn corporate officeholders into governors of Internet resources. Better rules can restrain agents, improve predictable administration and reduce continuity risk.

They cannot manufacture a power the company never possessed.

A disagreement over whether choice still existed

The immediate prehistory lay at AFRINIC-14 on 8 June 2011. The episode concerned the selection of co-chairs for the Policy Development Working Group, not a Board election, and that boundary is essential. The 2011 NomCom had two nominees for two open terms. It initially announced the nominees into the positions without holding a ballot. Participants objected. Their concern was not merely ceremonial. They argued that an election—or at least a show of hands—was still required, because a voter might reject a candidate even when the number of candidates matched the number of vacancies.

The recorded exchange exposes a basic design problem. Counting candidates and vacancies can determine whether there is competition between names; it cannot determine whether the electorate has been offered a meaningful choice. If acceptance is treated as automatic whenever supply equals the number of seats, nomination quietly becomes appointment. A rejection option changes the logic. It asks not only which person should fill a place but whether the person should fill it at all.

After debate, the NomCom asked for objections and proceeded to a show of hands. The record supports Paulos Nyirenda for a two-year term and Timothy McGinnis for a one-year term, with no opposition recorded at that step. A later objection was dismissed by the NomCom chair, Gift Shava. These documents establish the sequence and the positions taken in the room. They do not establish the legal validity of the outcome, the representativeness of the attendance or the motives of those involved. Nor do they justify converting the episode into a dispute about a Board election.

It was a PDWG co-chair selection whose procedural awkwardness supplied one visible example of the election-related feedback circulating at AFRINIC-14.

That last qualification matters because the October Board resolution referred broadly to feedback during AFRINIC-14. It did not itemise every complaint, rank them or say that the co-chair episode exhausted the committee's remit. The June minutes therefore illuminate the sort of procedural failure that could prompt review; they do not allow every later proposal to be traced back to that exchange.

Candidate eligibility, proxies, electronic voting, ballot custody and the composition of NomCom all appeared in the November draft, but the Board's one-sentence mandate did not publish a problem register showing which incident gave rise to which workstream.

This is the first gap in the accountability chain. A remediation process is easier to evaluate when the initiating decision names the failure modes it is meant to correct. “Review the election process” is broad enough to permit useful work, but too broad to provide a test of completion. Without a stated inventory of problems, nearly any procedural proposal can be presented as responsive, and nearly any omitted concern can disappear without explanation. The June dispute makes the need for a rejection mechanism intelligible. It cannot, by itself, prove that the later draft captured every significant concern raised at the meeting.

The baseline before the review

The nearest earlier archived constitution in the available record dates from 2007. It contained provisions on member nominations, proxies and director elections, but it did not contain separately titled articles for a Nomination Committee and an Election Committee. That comparison is useful but bounded. It is the nearest earlier constitutional text in the record examined here; it does not prove every amendment made between 2007 and 2011 or identify, by itself, the exact text legally operative on every day of 2011.

A more immediate operational baseline is Election Guideline v2.2. Its revision history ended in September 2010, though the archived mailing-list attachment is dated 28 June 2011. The attachment date must not be allowed to turn an older guideline into a post-review product. Read correctly, it shows what existed before Resolution 201110.135: nomination and election work combined in one volunteer committee, supported by AFRINIC staff, with rules covering timelines, member and candidate eligibility, proxies, paper voting, candidate checks, counting and the announcement of results.

This baseline complicates any claim that the review invented election administration from nothing. Many of the subjects later addressed in the draft already existed in operational guidance. The committee's value, therefore, could lie in revision, clarification, separation of functions or constitutionalisation—not necessarily invention. That is precisely why provenance matters. If an item appears in the old guideline, the committee presentation and the 2012 bylaws, one cannot assign authorship merely by observing the middle document.

The proposal might preserve an existing practice, sharpen it, move it into a higher-order instrument or coincide with work being done elsewhere.

The baseline also illustrates the importance of instrument hierarchy. A guideline can organise practice without having the same status as a constitutional clause. A presentation can recommend without adopting. A Board resolution can create a committee without approving its eventual output. A later bylaw can establish a corporate rule without revealing which drafter or consultation supplied the idea. Good institutional history keeps these layers separate. Poor history lets the presence of similar words substitute for the missing authorisation chain.

NRS's exact-instrument and member-authorisation method is valuable here because it resists self-proving labels. Calling a document a guideline, a resolution, a constitution or a community output does not answer who was authorised to do what, under which instrument, and with what effect. Each link must be read with the others. The method does not prove what this committee did; it provides the discipline for distinguishing documents that record different kinds of act.

A mandate with a large remit and a small sentence

On 12 October 2011, AFRINIC's Board resolved to appoint a committee composed of past NomCom chairs, the CEO and Legal Counsel. Its task was to review AFRINIC's election process and propose improvements as and where needed, following feedback during AFRINIC-14 in Dar es Salaam. That is the mandate. It is clear enough to establish the act of appointment, the categories from which the body was to be drawn and the general purpose of the work.

It is also strikingly spare. The published sentence did not name the individuals. It did not identify a chair, set a deadline, specify deliverables, describe a consultation method, state how recommendations would reach a decision, impose a publication duty or define success. It said nothing about budget, quorum or how disagreement inside the committee would be handled. A flexible remit may have helped the group move quickly. From the perspective of later accountability, however, flexibility became ambiguity: there was no public checklist against which to judge whether the mandate had been completed.

Even the resolution number is not cleanly preserved. The Board page labels the October act as Resolution 201110.135. Yet a record dated 21 November says that Resolution 201111.137 approved the minutes of the 12 October teleconference with amendments, including the election-review sentence labelled “Resolution 201111.135”. The same published page therefore presents 201110.135 in one place and 201111.135 in another. Nothing in the available record resolves which label was intended.

The discrepancy should not be dramatised into evidence of substantive failure, but neither should it be normalised away. Numbering is part of institutional memory. When a later reader tries to connect a committee's mandate to minutes, drafts, Board decisions and constitutional clauses, a stable identifier is the thread that keeps the file together. Here the thread already frays at the initiating act. This article follows the webpage's main designation, Resolution 201110.135, while preserving the fact that the amended-minute wording uses 201111.135.

The two labels cannot responsibly be treated as interchangeable without an explanatory record that is not present.

There is a broader lesson in the sparseness. Internal reform often begins with a decision that feels sufficient to contemporaries because everyone in the room knows the background. Time strips away that shared context. A concise minute then becomes the only surviving account of what the Board thought it had commissioned. If the mandate lacks named problems, milestones and a disposition path, future readers cannot reliably distinguish a completed review from an abandoned one that happened to leave a draft behind.

What the committee actually put on paper

The November presentation makes the review tangible. It named McTIM, Gift, Adiel and Mark as the committee. Those names should remain exactly as printed. The document does not authorise expansion of the aliases, and it does not establish how each individual mapped onto the resolution's categories of past NomCom chairs, CEO and Legal Counsel. The AFRINIC-15 meeting report separately records that Mark Elkins gave the election-process update, but that does not justify inferring that every appearance of “Mark” or any other shortened name supplies a complete role map for the committee.

The first substantive proposal was temporal. The draft laid out a countdown containing 30 days for calls for volunteers, 10 days for Board agreement on NomCom, 10 days for agreement on a chair, 30 days for nominations, 20 days for reviewing candidates and 15 days for comments. It also contemplated a five-day electronic-voting window, beginning ten days and ending five days before election day. The figures gave the process a visible cadence: construct the nominations body, open nominations, review candidates, allow comment, then move toward voting.

That cadence was useful because timing is a form of power inside an election. A deadline can widen or narrow participation, allow scrutiny or frustrate it, and determine whether a late procedural dispute can be repaired before a vote. Yet the terse slide does not prove that the countdown became an adopted calendar. It is a proposal expressed compactly, not a versioned guideline carrying an effective date.

The draft also proposed a four-volunteer NomCom: three community volunteers and one Board volunteer, accompanied by the CEO and a staff liaison. It contemplated selection of the chair by consensus, with the Board volunteer chairing if consensus failed. Here, too, the proposal responded to a genuine institutional question: how should the body that manages candidate selection be constituted, and what happens when it cannot agree on its own leadership? The design sought both participation and a fallback.

But a fallback that gives a particular category the chair can shift influence, which is why the eventual disposition and reasons would be valuable. The presentation tells us what was proposed, not whether the balance was accepted.

On voting methods, the draft listed paper, proxy and electronic voting for Board elections. It proposed a “None of the above” option, directly addressing the conceptual weakness exposed in June: two candidates for two places need not eliminate the voter's right to reject. The draft went beyond the presence of a ballot. It addressed whether candidates could vote or carry proxies, who should count, who should observe, what unsuccessful candidates could learn about their own results, and who should retain physical custody of ballots.

The counting proposal contemplated visiting staff from another regional Internet registry, with observation by NomCom and a legal representative. Winners would be announced; unsuccessful candidates would have access to their own results; legal counsel would retain ballot custody. Some slide items may describe a proposal and others a practice or tradition, but the presentation alone does not verify implementation. Its importance lies in showing the committee grappling with control points: the people who touch ballots, the observers who can contest a count, the information disclosed after a result and the location of the physical record.

Those control points are not decorative. Election integrity depends less on invocations of community than on arrangements that make error and manipulation detectable. A “None of the above” option preserves choice. Observed counting provides a check on custody and arithmetic. Candidate access to their own result gives limited visibility into the outcome. Retaining ballots creates the possibility of later review. Separating nomination work from administrative election work can reduce concentration. None of these devices proves legitimacy by itself, but each can lower the cost of challenging or reconstructing a decision.

The AFRINIC-15 report confirms that an update was given on the same broad subjects: committee composition, steps to election day, NomCom constituents, voting methods, counting and announcements. It does not contain a full record of audience comments, a vote or an act of Board acceptance. “Welcomed input” is evidence of an invitation, not evidence that everyone affected participated or consented. The meeting was a forum inside a private association's governance process, not a sovereign assembly.

The constitutional echo in 2012

The 2012 bylaws are materially more explicit than the 2007 archived baseline on election administration. They introduced separately titled provisions for a Nomination Committee and an Election Committee. NomCom was to consist of a chair and three members appointed by the Board. Its functions included seeking candidates, supervising polling on election day, setting eligibility criteria and finalising candidate lists. A staff Election Committee, designated by the CEO, was assigned pre-election processes and liaison with the NomCom chair.

The separation is notable. The pre-review guideline had combined nomination and election work in one volunteer committee supported by staff. The later constitution distinguished the nominations organ from a staff administrative committee and specified liaison between them. At the level of institutional architecture, that is real change. It clarifies who seeks and screens candidates, who handles preparatory administration and how the two sides connect.

The bylaws also addressed unresolved election-day questions. Issues arising during an election that were not expressly covered were to be resolved collectively and by consensus by members present. The resulting resolution would become precedent and form part of future election guidelines. This provision attempted to convert an improvised answer into institutional memory. Instead of letting each gap be solved anew, the company would carry the solution forward.

There is an evident attraction to that design, particularly after a meeting at which participants disputed whether a vote was needed in an uncompetitive field. Yet the rule has boundaries. “Members present” does not mean every member, and internal consensus does not create law for states, courts, non-members or non-consenting network operators. It is an internal corporate mechanism for filling procedural gaps. Its precedent effect operates within future AFRINIC election guidance; it does not confer adjudicatory authority in the public-law sense.

Article 12 of the 2012 text recognised paper ballots, electronic voting and proxies. It limited a voting member to carrying five proxies. These subjects overlap with the committee's November agenda, especially its attention to voting methods and proxies. The overlap makes comparison rational. It does not supply the missing provenance. The five-proxy limit, for example, appears as later constitutional text, but the available documents do not show that it originated in the committee draft, was debated by the committee, or was adopted because of Resolution 201110.135.

The distinction between echo and lineage is the central analytical problem. A later rule may echo an earlier proposal. Lineage requires more: a final recommendation, a decision by the competent corporate body, reasons for acceptance or modification, and ideally a mapping into the final instrument. Without those links, textual similarity can support a hypothesis but not a causal conclusion. The 2012 bylaws prove what the constitution then said. They do not prove that every relevant clause came from the election-review committee rather than a parallel bylaws exercise, staff drafting, legal review, member comment or pre-existing practice.

The charitable verdict—and its limit

The strongest case for treating the review as a success deserves to be stated plainly. The Board acted after visible procedural feedback. A four-name committee produced a detailed public draft within weeks. The draft invited input and dealt with practical weaknesses rather than offering slogans. Soon afterward, the 2012 constitution gave election administration a more explicit architecture, including distinct NomCom and staff Election Committee functions, multiple voting methods and a rule for turning unforeseen election-day decisions into future guidance.

Institutions often learn through precisely this combination of quick drafting and later codification.

On this reading, demanding a bureaucratic paper trail for every clause risks mistaking documentation for substance. The committee's job was to improve elections, not to satisfy a future historian. If the later text became clearer and operations became more predictable, the reform might have delivered value even if some intermediate documents were never published or did not survive.

That argument earns the draft proper credit. It prevents the absence of a final report from being treated as proof that nothing happened. But it cannot answer the questions that matter most for accountability. Which recommendations survived consultation? Which were rejected, and why? Did the Board approve the package, or did another corporate process absorb only parts of it? Which clauses preserved older rules? When did operational changes take effect? Did the next election exhibit fewer disputes, clearer choices, better disclosure or stronger custody controls?

These are not archival luxuries. They determine whether a remediation mandate can be evaluated on its own terms. A private company that relies on voluntary cooperation has particular reason to preserve the chain. It cannot fall back on sovereign authority when trust weakens. Its influence depends on accurate, predictable coordination and on the willingness of networks and members to treat its records and processes as dependable. Documentation is one of the instruments through which that dependability is earned.

The balanced conclusion is therefore asymmetrical. There is positive evidence of mandate, activity and draft output. There is positive evidence of later constitutional change. There is no positive evidence establishing a complete line of acceptance, implementation and evaluation between them. The missing links do not erase the acts that are proved. Nor can the proved acts be stretched across the gaps.

Why a registry election has economic weight without public power

It may seem excessive to analyse an internal election procedure in such detail. AFRINIC is not a state. Its officeholders are not public officials by virtue of running the registry company, and its members do not become a sovereign electorate by attending a meeting. Yet the quality of the company's governance can still have economic consequences because networks depend on the continuity and accuracy of number-resource records.

The relevant mechanism is operational, not governmental. Candidate rules affect who can reach office. Proxy rules affect how voting power is exercised. Counting and observation rules affect whether results can be checked. Disclosure affects whether candidates and members can understand an outcome. Ballot custody affects whether a dispute can be revisited. The division between NomCom and an administrative election committee affects concentration of control. Together these arrangements influence confidence in those supervising the narrow bookkeeping function.

Opaque governance raises costs. Members may need to spend more time monitoring procedures, obtaining advice, attending meetings or contesting outcomes. Operators may face greater continuity risk if internal conflict disrupts dependable registry administration. Unclear records make each new dispute more expensive because participants must reconstruct the history before they can argue about the present. LARUS's operational context is important on this point: RIR governance decisions can become infrastructure risk for dependent networks even though the registry lacks public-law authority.

The solution is not to inflate the company's status. Calling registry coordination “stewardship”, “community mandate” or “consensus” does not prove legitimacy or jurisdiction. Official documents establish their words, acts and operative corporate text, nothing more. The discipline is to care about governance precisely because the function is narrow and important. A reliable bookkeeper needs effective controls. An effective control system does not make the bookkeeper the owner of the assets described in its ledger.

That boundary also helps explain why election remediation is worth preserving. Better elections can constrain the agents who supervise the register, reduce opportunities for arbitrary administration and strengthen continuity. They cannot give AFRINIC sovereignty over Internet number resources, regulatory authority over networks, police power, punishment power, confiscatory power or the right to adjudicate outsiders' interests. Corporate legitimacy is relevant to the company's internal operation. It is not a ladder to public power.

The 2011 review should therefore be judged by a practical standard: did it make the internal accountability of a private coordinator more predictable, reviewable and durable? The surviving material permits only a partial answer. It shows serious attention to the machinery of choice and custody. It shows later constitutional structure. It does not show enough of the decision and performance record to determine which remedies became real or how well they worked.

The cost of a broken institutional memory

When a reform record stops at a draft, future disagreements tend to collapse distinct questions into one another. One side points to the presentation and says the problem was addressed. Another points to the missing adoption decision and says the committee achieved nothing. Both claims overreach. The first mistakes proposal for effect; the second mistakes incomplete preservation for proof of inactivity.

A durable institutional memory prevents this binary argument. It can show the problem observed, the evidence considered, the remedy proposed, the comments received, the competent body's decision, the instrument changed, the implementation owner and the later result. Each stage answers a different question. Removing one stage does not necessarily invalidate the others, but it reduces what later readers can know.

The numbering discrepancy is a small example of the same problem. Resolution 201110.135 and the reproduced phrase “Resolution 201111.135” may refer to the same underlying election-review sentence, but the record does not explain the difference. A disposition matrix or final report carrying a stable reference might have made the intended link obvious. Without it, even the name of the initiating instrument demands a caveat.

The absence of a clause crosswalk is the larger example. The committee discussed multiple voting methods; the later bylaws recognised them. The committee discussed NomCom; the later bylaws defined it. These parallels matter, but they do not answer whether the later language was copied, modified, independently developed or inherited. The analyst is left with resemblance where the institution could have supplied provenance.

BTW's reality-layer approach to bylaw research is useful because corporate text, adoption and implementation are separate realities. A rule printed in a constitution is more than a proposal, but it is not evidence of its own origin or performance. A properly adopted rule can still be poorly implemented. An operational practice can exist before it is constitutionalised. Keeping these layers distinct does not weaken the record; it prevents each document from being asked to prove more than it can.

The result is a verdict of material incompleteness, not institutional nihilism. Resolution 201110.135 was not empty. Its committee left a substantive draft. The later constitution changed. But the record that should connect those achievements is too thin to support precise attribution or a confident assessment of success.