A Call for Content Is Not a Selection Record: What NANOG’s Public Programme File Can Show

Summary

• NANOG’s 2026 call for NANOG 98 identifies proposal formats, an initial review, typical shepherd assignment, work on slides and content, and dated milestones. Those disclosures make the front of the programme process visible, but they do not reveal a complete set of proposals or the reason for any outcome.
• Historical materials from 2005, 2009, 2012, 2017 and 2020 illuminate different pieces of earlier selection and appointment arrangements. They must remain dated evidence, not be assembled into an invented description of current practice.
• A proportionate public record would report aggregate movement through defined stages, the agenda version and a general conflict-handling category while protecting unaccepted abstracts, comments, identities and small groups. Its value would be procedural legibility, not proof of representation, technical quality or freedom from influence.

The difference between an open door and a visible route

A call for content performs an important public function. It tells prospective contributors that an opening exists and, at its best, explains what may be submitted, in what form and by when. An agenda performs another important function. It tells prospective attendees what has been scheduled. Put the two documents side by side, however, and a large analytical space remains between them. The call describes an entrance. The agenda describes a result. Neither, by itself, is a selection record.

That distinction matters because public language about openness is easy to overread. A widely visible invitation may be generous in scope, yet it does not tell a reader how many proposals arrived, how they moved, how often development changed their form, or how capacity shaped the eventual programme. An agenda may be rich and useful, yet it does not reveal the population from which it was drawn. Those observations are not criticisms of the process. They are statements about what two kinds of public document can and cannot establish.

The 2026 call for NANOG 98 offers a particularly useful starting point because it says more than “send us a talk.” It accepts proposals for presentations delivered in person or live remotely. It identifies a submission deadline, an agenda date and a final-slide deadline. It says the Program Committee conducts an initial review, typically assigns a shepherd within two weeks, and works with submitters on slides and content before evaluating presentations for the agenda. The disclosed path therefore contains several distinguishable moments: submission, initial review, shepherding, development, agenda evaluation and scheduled release.

That is meaningful process information. It tells a prospective submitter that a proposal may enter a developmental relationship rather than face only an instantaneous yes-or-no judgment. It tells a reader that the slides and content may be worked on before agenda evaluation. It establishes announced timing points against which the public sequence can be understood. It also distinguishes two delivery formats at the point of entry.

Yet the call does not become a complete record merely because it names stages. It does not disclose the full population of proposals. It does not say how many reached each stage. It does not identify reviewers or describe every exchange with a shepherd. It does not explain the outcome of a particular proposal. It does not show whether an invitation to revise ended in selection, withdrawal or another result. Nor does it offer a basis for judging the final agenda’s technical quality. The document is informative because it describes an announced route, not because it resolves every question about traffic along that route.

This is the central discipline for reading a public programme file: treat evidence of a stage as evidence of that stage, and no more. A call can establish that proposals were invited under stated terms. Review language can establish that review was announced. A shepherding statement can establish that development was contemplated. An agenda date can establish when a public result was expected. None of those facts reveals, without additional evidence, why any item appeared or did not appear.

The practical question is therefore narrower than “Is the programme open?” and more useful than “Can the public see everything?” It is this: what is the smallest set of aggregate information that would allow an outside reader to reconstruct the broad proposal-to-agenda path without intruding on legitimate confidentiality or exhausting a volunteer committee? Framing the issue that way avoids turning missing detail into suspicion. It also avoids assuming that maximal disclosure is either possible or desirable.

Six dated windows, not one timeless procedure

The reviewed public materials span more than two decades, but they do not form a single continuous rulebook. Each source answers a different question at a particular time. Their value depends on preserving those boundaries.

A public presentation from 2005 gave a historical account of talk selection. It described a call for presentations, review and rating, a conference call, initial selections and possible accept, reject or conditional outcomes. It also described a conflict-of-interest policy adopted in 2004, disclosure of conflicts during discussion, and recusal when a committee member considered recusal appropriate. That presentation is evidence that such elements were publicly described at that time.

It is not evidence that every later proposal passed through identical steps, that the same conflict arrangements remain current, or that any particular conflict existed.

The account is nevertheless analytically useful because it shows how much more a process description can reveal than a call alone. It names possible outcomes beyond a simple published/not-published comparison. “Conditional” is especially important as a historical category because it suggests that development and selection need not be cleanly separated. A proposal can be neither finally accepted nor simply rejected at an intermediate moment. But the historical category cannot be imported into the 2026 call as a claimed current status. The 2026 material has its own words; the 2005 presentation has its own historical scope.

A 2009 announcement described the Program Committee as responsible for soliciting and selecting material for meeting programmes. It also said that, after an election, a new committee would be selected by the Steering Committee. This is dated evidence about remit and an appointment path. It is not a basis for asserting how a present committee is selected or how authority is presently distributed.

A 2012 announcement offered another dated view. It described a 16-person Program Committee responsible for soliciting and selecting programme material. It said eligible candidates would serve two-year terms and stated expectations involving attendance, good standing, technical knowledge, familiarity with meetings and recruitment of presentations. Those details show what an appointment notice told potential candidates at that point. They do not prove current committee size, term, eligibility or practice. They also do not show whether every candidate met every expectation or how expectations were weighed.

A 2017 announcement supplied a different kind of fact: 27 candidates sought nine available Program Committee positions, and the Board announced the appointees after completing the selection process. The ratio is a fact about that appointment round, not a measure of the committee’s representativeness, legitimacy or programme choices. The announcement does not provide candidate-by-candidate reasons, deliberations, conflicts or a vote. It links a stated candidate population to a stated number of positions and a Board-announced result; it cannot explain more than that.

A 2020 Board-candidate page said that the Board was responsible for selecting committees, including the Program Committee. That statement identifies a governance responsibility as it was presented on that page. It does not show the Board choosing a proposal, directing a shepherd, deciding an agenda item or influencing a speaker. Committee selection and content selection are different acts, even when they sit in a connected chain of institutional authority.

Finally, the 2026 call describes announced proposal handling for NANOG 98. It is the most recent material in the reviewed set, but recency does not allow the older descriptions to be silently attached to it. The call’s review and shepherding language can be reported as current to that call. The historical presentation’s rating, conference-call and conditional-outcome language cannot be called current without current evidence. The 2009 Steering Committee path, the 2012 size and expectations, the 2017 appointment round and the 2020 responsibility statement all remain dated.

Keeping the windows separate produces a more accurate, if less dramatic, picture. NANOG has repeatedly made some aspects of programme solicitation, committee remit, appointment and selection description public. The public emphasis and level of detail differ by source and era. The reviewed set does not publish a complete contemporary proposal record, and it does not establish that the process is deficient. It shows partial visibility across distinct functions and times.

Who acts, who participates and who bears the result

Programme discussion becomes imprecise when every person around a meeting is collapsed into “the community.” The sources identify roles that must remain analytically separate.

Submitters offer proposals. Some may eventually become accepted speakers, but the two groups are not identical. A person can submit and not appear on an agenda; a speaker can appear only after development changes a proposal; the reviewed material does not enumerate either population. The 2026 call also distinguishes in-person and live-remote presentation proposals at entry, but that format distinction does not describe every characteristic of the people proposing them.

Shepherds, as described in the 2026 call, work with submitters on draft slides and content. That is a developmental role. It is not automatically identical to the role of every reviewer, every Program Committee member or any person with appointment authority. The call says the Program Committee conducts an initial review and typically assigns a shepherd; it does not map every individual to every act.

The Program Committee is described in the 2026 call as reviewing proposals and working through the shepherding and agenda-evaluation sequence. Historical announcements describe solicitation and selection responsibility in their own periods. Those statements do not turn the committee into a proxy for all attendees, all members, all sponsors, all network operators or a continent. A committee may exercise bounded programme judgment without claiming a mandate from everyone who could conceivably care about the meeting.

The Board appears in the 2017 appointment announcement and in the 2020 description of committee-selection responsibility. Those facts concern the selection of committee members, not direct selection of programme proposals. The Steering Committee appears in the 2009 appointment description, again within that historical setting. Treating appointment authority, review authority and final programme judgment as a single undifferentiated power would obscure rather than clarify the record.

Attendees are another population. A meeting audience experiences the resulting agenda, but attendance does not prove that each attendee endorsed every item or the selection method. Members, where a source refers to them, are not necessarily the same as attendees or submitters. Sponsors may help sustain a meeting, but the reviewed sources do not establish a sponsor role in programme selection. The wider operator sector extends beyond all of these groups, including people who neither submit nor attend. No agenda can be treated as a survey of that sector merely because the meeting concerns network operations.

These separations have a direct bearing on claims about representation. A call can be available to prospective submitters without proving that the resulting submissions represent every relevant constituency. A committee can be appointed through a stated path without its membership serving as a statistical sample of an audience. An agenda can be valuable without expressing the collective will of non-entities. Participation is real, but it is not the same thing as delegation of authority by absent people.

The same caution applies to influence. Public material showing that one body appoints a committee establishes an authority link. It does not establish that the appointing body determined a particular content outcome. Public material showing shepherding establishes a developmental interaction. It does not establish that a shepherd caused acceptance or non-acceptance. A historical conflict policy establishes that conflict handling was publicly described at a point in time. It does not prove that an undisclosed conflict existed in any case.

Role separation is not semantic fussiness. It lets a public record be designed around actual decision points. A useful record might distinguish the body responsible for publishing the call, the committee responsible for review, the developmental stage assigned to shepherding and the date on which an agenda version appeared. It need not name individuals to clarify that the roles differ. Indeed, a role-based aggregate account can improve legibility while protecting people more effectively than a loose narrative about “NANOG” deciding everything.

What the public file establishes — and leaves unresolved

One way to resist overclaiming is to test each public item against a narrow evidentiary question.

The 2026 call establishes an announced opening for in-person and live-remote proposals, several dates and a described path through initial review, typical shepherd assignment, content development and agenda evaluation. It leaves unresolved how many proposals entered, how many moved through each stage, who reviewed them and why any one proposal reached an outcome.

The 2005 presentation establishes that a more detailed process, including rating, a conference call, several possible outcomes and conflict language, was publicly described historically. It leaves unresolved whether the same process continued, how consistently it operated, what any rating meant in a particular case and how any recusal affected an outcome.

The 2009 announcement establishes a historical committee remit and a stated Steering Committee appointment path. The 2012 announcement establishes a historical description of committee size, responsibility, terms and candidate expectations. The 2017 announcement establishes competition for a finite number of committee positions and a Board-announced appointment result. The 2020 page establishes a stated Board responsibility for selecting committees. None supplies the missing proposal population for NANOG 98. None explains a particular programme outcome.

This produces a public chain with identifiable pieces but important open intervals. At the front, there is a call with formats and milestones. In the middle, there is announced review and development language. At the institutional level, dated materials explain aspects of committee remit and appointment. At the public endpoint, the call names an agenda date. What is not published in the reviewed materials is a contemporary aggregate account connecting the number and types of proposals received to their movement through the named stages.

The phrase “not published in the reviewed materials” is deliberately limited. It does not mean the information does not exist. It does not mean anyone is concealing it. It does not mean the committee made poor choices. It does not mean commercial interests prevailed, that a constituency was excluded, or that a conflict was mishandled. It identifies only a boundary on what an outside reader can reconstruct from this set of public documents.

Nor would a complete aggregate path settle every substantive question. Suppose a record showed 100 received proposals, 70 shepherded, 40 selected and 10 withdrawn. Those counts would make movement visible, but they would not show whether the selected presentations were technically superior to the alternatives. They would not prove that review was impartial, or that it was partial. They would not transform submitters into a representative sample. They would not expose the quality of discussion or the judgment needed to fit content into a finite meeting.

Auditability and quality are related only indirectly. A visible path can help readers ask better questions and can help organisers explain constraints. Quality depends on subject knowledge, relevance, preparation, delivery, timing and other judgments that aggregate counts cannot capture. The case for a minimal record is therefore not that it verifies excellence. It is that it reduces ambiguity about the broad administrative path while leaving room for expert evaluation.

A minimum viable aggregate record

The strongest case for more visibility begins with restraint. The objective should not be a public archive of unsuccessful proposals or a transcript of committee discussion. It should be a compact, dated account capable of answering a few basic questions: What was invited? How many items entered? Through which broad stages did they move? What format constraints mattered? When did the public agenda version appear? What categories of information were deliberately withheld?

The first field should be the call date and the criteria stated in the call. This anchors the record to the invitation actually made, rather than to a later recollection. The criteria should be reproduced or summarised only as publicly stated. If the call distinguishes in-person and live-remote presentations, the record can preserve that distinction without adding assumptions about the people choosing each format.

The second field should be proposal type. A type should be broad enough to avoid revealing a submitter through an unusual subject in a small pool. Format may be safer than topic when numbers are low: in-person and live-remote are already categories named in the 2026 call. If any category contains too few items, it should be combined or suppressed. The purpose is to understand capacity and movement, not to reverse-engineer confidential content.

The third field should be stage status. A practical set might include received, under review, invited to revise, selected, not selected and withdrawn. These labels are an editorial proposal, not a claim that NANOG currently uses them. Their value lies in separating different paths that a call and a final agenda otherwise compress into a binary result. “Invited to revise” recognises the developmental character suggested by shepherding. “Withdrawn” avoids treating every item absent from the agenda as a committee rejection.

Stage labels would require plain definitions. “Received” might mean administratively complete by the deadline. “Under review” might mean admitted to substantive consideration. “Invited to revise” might mean development was requested without a final programme commitment. “Selected” might mean included in a specified agenda version. “Not selected” might mean no place was offered for that meeting. “Withdrawn” should remain a submitter-side status, not a disguised judgment. Definitions matter because counts are misleading when categories overlap or change meaning from one meeting to another.

The fourth field should be aggregate counts by stage and format. A compact table could show totals at a few dated checkpoints. Counts should never be reported so granularly that an individual becomes obvious. Where numbers are small, formats can be combined, time periods can be widened, or a threshold can replace an exact count. The record should also explain whether a proposal can appear in more than one stage over time, so that readers do not add columns as if they were mutually exclusive final outcomes.

The fifth field should be a general conflict-handling category. The historical 2005 material shows that disclosure and recusal language has appeared publicly before, but it cannot establish current arrangements. A contemporary aggregate field should therefore be framed generically and only if it reflects a current, publicly authorised practice. Possible reporting could be limited to whether a defined conflict-handling process was available and, if safe, whether one or more matters required handling. It should not name a person, identify a proposal or imply misconduct.

With very small numbers, even a count may disclose too much; “process available; case details not disclosed” may be the proportionate boundary.

The sixth field should be agenda publication and version date. A dated version matters because programmes can change for reasons unrelated to selection: speaker availability, travel, format or scheduling may alter what appears. The record should identify which version its selected count describes. If later changes are reported, they should be labelled as programme changes rather than retroactively rewritten as original selection outcomes.

The seventh element should be an explicit non-disclosure boundary. Unaccepted abstracts, reviewer comments, submitter identities, reviewer identities and small groups should remain outside the public aggregate record unless the people concerned have separately authorised release. The boundary is not a caveat tucked away at the bottom. It is part of the record’s integrity because it tells readers which questions the record was never designed to answer.

Taken together, these fields would produce a modest chain:

Public element	What it would clarify	What it would not establish
Call date and stated criteria	The terms and timing of entry	Whether the criteria produced a good agenda
Proposal type	Broad format demand and capacity	The identity or technical merit of a proposal
Stage status	The main routes between receipt and outcome	The reasons for an individual decision
Aggregate stage and format counts	Movement through the process	Representation of every interested population
General conflict-handling category	The presence of a bounded handling approach	Any individual conflict or effect on an outcome
Agenda and version date	The public endpoint described by the counts	The permanence or quality of every scheduled item
Non-disclosure boundary	What has intentionally been protected	That protected information contains evidence of a problem

This is a minimum, not a wish list. Demographic breakdowns, employer categories, proposal topics and detailed reason codes may appear attractive, but each can create privacy and interpretation problems, especially in a small field. A narrow record should be preferred if it answers the core process question. Additional fields should be added only when they serve a defined public purpose and remain safe at realistic group sizes.

The serious case against detailed disclosure

Any proposal for a public programme record must contend with the possibility that disclosure can damage the process it seeks to illuminate. That counterargument deserves more than a paragraph of ritual acknowledgement.

First, proposals can be confidential. A prospective speaker may want early feedback on operational experience, a developing incident analysis or work not ready for broad circulation. An unsuccessful abstract may reveal plans, weaknesses or affiliations that the submitter never consented to make public. Even a topic label can identify a person when only one specialist is known to be working on an issue. Protecting unaccepted abstracts and identities is therefore substantive, not cosmetic.

Second, expert judgment does not always reduce to stable public reason codes. Reviewers may balance novelty, operational value, evidence, clarity, overlap, readiness, time available and the composition of the programme as a whole. Two individually strong proposals may compete for one slot because they cover similar ground. A developing proposal may become excellent through shepherding, while a polished submission may fit poorly with the available format. A simplified reason such as “topic overlap” can mislead if readers treat it as the full explanation.

Third, speaker development benefits from candour. Shepherding can require direct criticism, experimentation and multiple drafts. If every comment were expected to become public, reviewers might write defensively and submitters might avoid ambitious ideas. Confidential discussion can create space to improve a presentation without turning early weaknesses into a permanent public label. The 2026 call’s description of work on draft slides and content makes that developmental interest especially relevant.

Fourth, meeting design imposes format constraints. An agenda has finite time. In-person and live-remote delivery may create different scheduling or production considerations. The strongest set of proposals considered independently may not form the strongest coherent programme when placed together. Aggregate reporting can acknowledge format and capacity, but it cannot reproduce the full texture of programme construction.

Fifth, the relevant groups may be small. A category containing one or two proposals is not meaningfully anonymous. Combining small cells reduces detail; suppressing them reduces apparent completeness. Both are appropriate when privacy is at stake. A public record must be designed for the actual distribution of submissions, not for an imaginary large sample in which every cross-tabulation is safe.

Sixth, volunteer capacity is finite. Every field requires definition, collection, checking, release and correction. A record that is too demanding can redirect expert time from reviewing and developing presentations. It can also become inconsistent if maintained under pressure, producing false precision. The smallest sustainable record is more valuable than an elaborate account published irregularly or without clear definitions.

Seventh, disclosure can generate strategic behaviour. If prospective submitters believe a particular stage count or reason category offers a clue to selection, they may tailor labels rather than improve substance. If organisers publish ever more granular categories in response, the process can enter an unproductive cycle of gaming and counter-gaming. Aggregation, delayed release and stable broad definitions can reduce that risk, but they cannot eliminate it.

These objections argue against a detailed public dossier. They do not necessarily argue against every aggregate measure. The proportionate response is to set a high bar for adding fields and a low bar for protecting people. A record can say that 60 proposals were received without saying who sent the 21 that did not reach the agenda. It can say that a development stage existed without releasing comments. It can identify the agenda version without narrating every scheduling exchange.

The relevant balance is not secrecy versus transparency in the abstract. It is public legibility versus the conditions required for candid, expert and workable selection. A minimal record succeeds only if it improves the first without materially weakening the second.

How to test the idea without prejudging the process

A proportionate record should begin as a time-bounded trial. The trial would test administrative usefulness, privacy risk and interpretability. It would not be an investigation into presumed wrongdoing, because the reviewed materials supply no basis for such a presumption.

Before the call opens, the trial should define its fields, stage meanings, small-group rule and release timing. Definitions should not be changed after outcomes are known merely to produce a cleaner-looking result. If a definition must change, the record should state the change and keep earlier and later counts comparable where possible.

During the proposal period, organisers would collect only the minimum information required for the agreed aggregates. Proposal content and identities would remain in the existing confidential environment. The public-facing count file would use broad categories and dated snapshots. A privacy check before release would combine or suppress small groups.

After the agenda appears, the trial would publish one concise table tied to that agenda version. It would state the total received, aggregate movement through the defined stages, broad format counts where safe, the general conflict-handling statement and the non-disclosure boundary. A later note could record aggregate programme changes without assigning motives.

Evaluation should ask practical questions. Did readers understand the difference between received, revised, selected and withdrawn? Did the fields answer a real question that the call and agenda could not? Did any field risk identifying a person? How much volunteer time did collection and checking consume? Did submitters report that the prospect of aggregate disclosure changed their willingness to propose? Did the record invite misleading claims that could not be corrected by clearer definitions?

The trial should also have stopping rules. A field should be removed if it regularly creates a small-group disclosure risk, is too burdensome to maintain, or is consistently interpreted as a judgment it cannot support. The entire record should be paused if it materially chills submissions or compromises developmental candour. Conversely, a field should not be expanded simply because readers want individual explanations. The stated purpose is aggregate reconstruction, not case-by-case appeal.

Success would be modest. A reader could trace the broad transition from call to agenda and distinguish withdrawal, development and non-selection at aggregate level. Organisers could explain capacity without discussing individuals. Submitters could see that their confidential material would not be exposed. No claim of broad representation or guaranteed quality would follow.

Failure would also be informative. If the categories prove unstable, the groups too small or the burden too high, the responsible conclusion may be that the public call and agenda should remain the principal documents, perhaps with only one or two headline counts. Proportionality includes the possibility that less disclosure is the better design.

What stronger visibility would still not prove

The temptation after building a better process record is to ask it to carry substantive conclusions it cannot bear. Several limits should be explicit.

It would not prove capture or its absence. Aggregate movement can reveal concentration at stages, but it cannot identify motives or hidden influence. The reviewed public materials provide no evidence that a particular constituency controls the programme, and a new table would not automatically answer that question.

It would not prove bias or perfect neutrality. Counts can show differences in outcomes only when categories are both available and safe, but they do not explain why differences arise. The proposed minimum intentionally avoids sensitive or identifying breakdowns. It is a map of stages, not an evaluation of fairness.

It would not prove exclusion. A low number in a format or topic category might reflect the submission population, meeting constraints, category definition or chance. People who never submitted are absent from the proposal path altogether. The record could not tell whether non-submitters felt invited, had relevant content or chose other venues.

It would not prove representation. Submitters are self-selected. Accepted speakers are selected from that self-selected group. Attendees choose whether to attend. Members, sponsors and the wider operator sector each have different boundaries. A transparent programme path cannot turn one of these groups into the authorised voice of the others.

It would not prove technical superiority. Expert review can aim for a strong programme, but aggregate stages do not compare the substance of selected and unselected proposals. Public release of all abstracts would not solve that problem and could cause serious harm. Quality remains a judgment informed by material that a privacy-preserving record deliberately withholds.

It would not prove individual influence. A Board responsibility for selecting committees does not show a Board decision on a talk. A shepherding relationship does not show that a shepherd determined an outcome. Historical conflict language does not identify a conflict. Authority links tell readers where formal responsibilities have been described; they do not establish causation in individual choices.

These limits do not make the record pointless. On the contrary, a clearly bounded record can improve public discussion by replacing speculation about basic counts with defined facts. Its credibility depends on refusing the next leap. Better evidence about stages should produce narrower, better questions, not grander claims.

Comparability is a choice, not an automatic benefit

If an aggregate record were repeated, its apparent precision would create another challenge: people would naturally compare one meeting with the next. Comparison can be useful, but only when the underlying terms, formats and checkpoints remain comparable. A count that looks identical may describe a different administrative reality if one cycle records a proposal at first receipt and another records it only after an initial completeness check.

The safest first record would therefore stand on its own. It could explain one call-to-agenda path without promising a time series. Only after the stage definitions had survived practical use should organisers consider a repeated comparison. Even then, every release would need to state whether the call used the same proposal formats, whether stage definitions changed and whether a privacy rule caused categories to be combined.

This caution is particularly important for ratios. An apparent selection rate compresses withdrawal, revision, format limits and timing into one number. It can be arithmetically correct and still convey the wrong story. A denominator based on everything received may differ from one based on administratively complete proposals. A numerator based on an initial agenda may differ from one based on the programme delivered after withdrawals. There is no universally correct choice; there is only a clearly defined choice suited to the question being asked.

Longitudinal figures would also tempt readers to interpret a change as institutional performance. A higher share selected might reflect more agenda capacity, fewer submissions, different formats or a narrower call. A lower share might reflect stronger demand, overlapping proposals, more withdrawals or another constraint. Without evidence connecting the change to a cause, the direction alone should not be praised or criticised.

For that reason, the best monitoring note would privilege levels and movement over a single headline rate. It would report the received count, stage transitions and the agenda version under fixed definitions. If organisers eventually calculate a ratio, they should display its numerator and denominator beside it. A reader should be able to reconstruct the arithmetic and see which populations were excluded.

Comparisons across roles require the same discipline. The number of committee candidates in the 2017 appointment announcement cannot be compared directly with the number of programme proposals in a meeting cycle. One concerns people seeking committee positions; the other concerns content offered for a programme. The 16-person committee described in 2012 cannot be treated as the denominator for a current workload estimate. The 2020 statement of Board responsibility cannot be converted into a measure of Board intervention in content. Numbers become misleading when they are detached from the act and date they describe.

A repeated record should also resist retroactive tidying. If a stage definition turns out to be unclear, the correction should be documented rather than silently applied to earlier figures. If a small-group rule suppresses a category in one cycle but not another, the absence should be labelled as protection, not zero. If an agenda version changes, the record should retain the earlier checkpoint so that the original selection figure and later programme composition are not confused.

The leadership benefit of this restraint is credibility. A small number of well-defined, non-comparable first-cycle facts is more trustworthy than a polished trend line assembled from changing categories. Over time, consistent definitions may permit careful comparison. Until then, the record’s primary job is to describe a path, not to create a performance index.

The public value of a bounded answer

NANOG’s programme materials already demonstrate that visibility is not all-or-nothing. The 2026 call names formats, stages and dates. Historical materials describe earlier selection, conflict and appointment arrangements at varying levels of detail. The resulting file is neither opaque nor complete. It is a set of dated windows onto different parts of programme formation.

The right response to that partial view is not to infer a defect. There is no evidentiary basis here for alleging capture, retaliation, commercial direction, hidden conflict or deficient programme quality. Nor is the right response to treat a public call as if it answered every question a selection record would answer.

A minimum aggregate record offers a middle course. It would preserve the confidentiality of unsuccessful work, the candour of reviewer and shepherd interactions, the flexibility of expert judgment and the practical limits of volunteer effort. At the same time, it could tell the public how many proposals moved through a few defined stages, which broad formats were involved, what agenda version the figures describe and where disclosure deliberately ends.

The measure of such a record would not be whether it satisfied every demand for explanation. It would be whether it made one institutional path easier to reconstruct at acceptable cost and risk. If it did, readers would gain a clearer view of the distance between invitation and agenda. If it did not, the trial could contract or stop without implying that the underlying programme had failed.

A call for content is evidence of an opening. An agenda is evidence of a scheduled result. A selection record, even a minimal one, is evidence about movement between the two. Keeping those three propositions separate is less exciting than announcing that the programme is either fully open or fundamentally closed. It is also the more responsible reading of what NANOG’s public file can show.

Sources

1. NANOG, 2026 NANOG 98 call for proposals and announced programme milestones: https://lists.nanog.org/archives/list/[email protected]/thread/N5OMYFQCCQQUTRTJH6KXHBN2YYJMFC6G/
2. NANOG announcement archive, 2012 Program Committee description and appointment expectations: https://lists.nanog.org/archives/list/[email protected]/2012/9/
3. NANOG announcement archive, 2017 Program Committee appointment round: https://lists.nanog.org/archives/list/[email protected]/2017/2/
4. NANOG mailing-list archive, 2005 public presentation on historical talk selection and conflict handling: https://lists.nanog.org/archives/list/[email protected]/2005/1/
5. NANOG announcement archive, 2009 Program Committee remit and stated appointment path: https://lists.nanog.org/archives/list/[email protected]/2009/7/
6. NANOG, 2020 Board-candidate page describing committee-selection responsibility: https://nanog.org/participate/elections-nominations/2020-board-candidates/