Mirjam Kühne and the Work of Making Experiments Public

Mirjam Kühne, RIPE Labs and the public evidence layer around an RIR

This article treats RIPE Labs as a public evidence layer around RIPE NCC's registry, number-resource and measurement work. The linked record shows provisional tools, address-space and autonomous-system data services, probe participation and infrastructure-mapping methods being exposed under named authorship and bounded claims.

That public record does not make an institution sovereign, and it does not make Kühne the owner or engineer of the systems described through the platform. It shows a registry institution acting as recordkeeper and operator while public evidence lets technical communities inspect running work, attribution, uncertainty and continuity.

Summary

• RIPE Labs began as a collective experiment involving Róbert Kisteleki, Daniel Karrenberg, Mirjam Kühne, RIPE NCC teams and the wider community. Kühne's documented contribution was to make that experiment legible and public as a Community Builder, maintainer, editor, curator, author and presenter, not to serve as sole founder or platform engineer.
• The useful unit of RIPE Labs was not simply an article. It was a public connection between work in progress and a community able to inspect it: launch explanations, comments, serial RIPEstat demos, invitations to host measurement probes, detailed mapping articles, meeting participation and working-group distribution.
• RIPEstat, RIPE Atlas and RIPE IPmap/OpenIPmap demonstrate that public explanation can sit beside technical development without becoming the development itself. Mark Drayton, Daniel Karrenberg, Jasper den Hertog, Massimo Candela, RIPE NCC engineering and operations teams, probe hosts and contributors retain the credit for the work the evidence assigns to them.
• The archive also exposes the costs of public technical knowledge. Feedback did not always work as intended; comments and information architecture needed attention; article and probe counts show scale rather than influence; and the 2020 hand-off to Alun Davies marked both continuity and the maintenance burden created by more than a decade of publication.

A laptop that would not behave

Early in 2009, a prototype was demonstrated in Moscow on a broken laptop. The machine had to be held at a particular angle to keep it working. That awkward detail, preserved in Alun Davies's account of RIPE Labs' beginnings, matters because it places the project before the confidence of an anniversary. The demonstration was not a polished unveiling of an established service. It was a vulnerable showing of something unfinished, carried by equipment that could fail in the room.

The practical question was whether an idea could become useful while its limitations were still visible.

The origin was collective. Davies's retrospective identifies Robert Kisteleki's idea and describes discussions involving Kisteleki, Daniel Karrenberg and Mirjam Kühne. A contemporaneous presentation by Kühne likewise credited Róbert Kisteleki and Daniel Karrenberg. RIPE NCC supplied the institutional setting and resources in which a prototype could become a maintained public platform. That division of credit is not a footnote to the story.

It is the story's first rule: the public face of experimental work must not erase the people who conceived, built, operated or contributed to it.

After the Moscow demonstration, Karrenberg committed to launching RIPE Labs by RIPE 59. The public launch followed in October 2009. The exact day is less important than the sequence the available record can defend: a fragile prototype, a public commitment tied to a community meeting, and a launch within months. That sequence did not prove that the platform would succeed. It established a test.

Could RIPE NCC staff and people beyond the organisation expose ideas, prototypes, analyses and tools before those things had acquired the authority of finished services?

Kühne's launch-era explanation answered by defining the proposed relationship with readers. RIPE Labs would carry new ideas, prototypes, analyses and tools, while comments and suggestions would be part of how the public surface worked. Publication in this conception was not an award presented at the end of development. It was an intermediate operating condition: work became visible enough to be understood, challenged and discussed while change was still possible.

That is why the broken laptop should not be turned into a founder legend. Its value is almost the reverse. It shows a team accepting that unfinished work might have to appear unfinished. Kisteleki's idea, Karrenberg's commitment, Kühne's public explanation, RIPE NCC's resources and the community's prospective participation were different contributions. The platform would depend on keeping those differences legible even as RIPE Labs gave them a common address.

The scene also contains an institutional wager. Technical organisations often have good reasons to delay publication: a prototype may break, an analysis may be revised, a tool may lack guarantees, and an early explanation may invite criticism before the work is stable. RIPE Labs treated those conditions as reasons for careful framing rather than automatic secrecy.

Its premise was that a community concerned with Internet operations and number resources could benefit from seeing the provisional state, provided that the limits and ownership of the work were made clear.

For Kühne, this created a person-level assignment that was substantial without being total. She would become closely associated with RIPE Labs, but the defensible importance of that association lies in mediation: explaining what the platform was for, presenting it beyond its home institution, maintaining an editorial surface and helping material travel toward relevant audiences. The broken laptop opens that account because it makes the mediation problem concrete.

Before anyone could debate a prototype, someone had to make the imperfect thing public without pretending it was complete.

The job at the boundary

Later official profiles use strong language about Kühne's relationship to RIPE Labs. The RIPE Chair Team page says that, during her time at RIPE NCC, she created and curated RIPE Labs.

The legacy RIPE Labs author archive describes the collected writing as produced while she was a community builder and calls her maintainer and editor of RIPE Labs. A 2019 conference biography places her Senior Community Builder role at RIPE NCC from 2009. These records support a durable editorial and community remit.

They do not support an unlimited one. “Created and curated” in a later profile cannot cancel the origin evidence naming Kisteleki and Karrenberg. “Maintainer and editor” does not reveal a private log of every article commissioned, rejected, rewritten or accepted. A byline archive does not show that its subject personally edited every item on the site. None of those descriptions makes Kühne the engineer of RIPE Labs software, much less the architect of every tool discussed through it.

The precision of the profile depends on holding the public role and the missing private detail at the same time.

The work that can be seen is nonetheless consequential. Kühne authored the launch explanation. She presented RIPE Labs at IETF 76 in November 2009, where the platform was described as a place to test and evaluate prototypes, contribute ideas and research, and discover and discuss work through public formats. In 2010 she presented RIPE Labs as tools and news for operators, explaining a surface that made RIPE NCC and community technical work visible. Those are not generic biography entries.

They are observable acts of translation between builders, an institution and potential users.

Translation here means more than simplifying technical language. A public presenter has to state what kind of object the audience is seeing. Is it a supported service, a prototype, an analysis, an invitation to contribute, or a demonstration of a feature? The 2009 presentation drew boundaries explicitly: there were no service guarantees, and material could disappear because feedback, legal or abuse concerns, or insufficient resources made continuation inappropriate.

That framing protected the distinction between experimental publication and production commitment.

It also made failure discussable. A platform that publishes only finished success stories gives readers little help in judging uncertainty. RIPE Labs' early terms allowed a prototype to be useful without being durable and allowed public attention without promising permanent operation.

The editorial task was therefore partly one of status. The article, presentation and surrounding context had to tell a technically sophisticated audience what could be relied on, what could be tested and what remained contingent.

Kühne's role sat at that boundary. Engineers and analysts could explain their own work under their own names. Community members could contribute. RIPE NCC could support the platform without claiming that every item was a guaranteed service. Readers could comment or carry the material into meetings and working groups. An editor and community builder helped those parties encounter one another without collapsing their authority into a single institutional voice.

The migrated current RIPE Labs author page preserves continuity in Kühne's public authorship, but it also illustrates why an archive needs interpretation. Site migrations can regroup names and articles; a list of posts cannot reconstruct every action behind them.

What survives most clearly is the visible remit: she repeatedly explained RIPE Labs, wrote through it, presented it and later marked its scale. The invisible remainder should stay invisible rather than being filled with claims about private decisions.

This distinction keeps the profile person-centred. To say that Kühne did not engineer the tools is not to reduce her contribution to publicity. Public technical knowledge needs an operating surface of its own.

Someone has to define the invitation, preserve authorship, expose limitations, connect audiences and keep the archive intelligible. Those tasks shape whether technical work can be encountered outside the team that produced it. They are infrastructure for understanding, even when they are not infrastructure engineering.

An invitation with conditions

The original invitation contained an appealing theory of feedback. Publish work early; let operators, researchers and other community members inspect it; gather comments and suggestions; improve the work through a shorter distance between development and discussion. Kühne's launch article and the 2009 presentation support that design intent. They do not show which comment changed which feature, whether a reader became a contributor, or whether public discussion improved a particular tool.

Intent is evidence of a proposed mechanism, not proof that the mechanism worked.

The distinction became visible within the first year. In June 2010, Kühne announced that RIPE Labs 2.0 was going live. The update addressed the article system, comments, user profiles and community usability. It is fair to treat the announcement as evidence that a public platform required iteration after early use. It is not fair to infer that Kühne personally designed or engineered every change, or that the revised version achieved all of its feedback goals.

This is a recurring cost of making experiments public. A prototype may be the object everyone wants to discuss, but the quality of the discussion also depends on less glamorous systems: registration, identity, navigation, comment presentation, article structure and the ability to find related material.

If those systems are awkward, the publication layer can obstruct the feedback it was meant to recruit. The work of openness shifts from the initial decision to publish toward the continuing work of making participation usable.

The early platform rules acknowledged another cost: experimental material can disappear. Insufficient resources, legal or abuse issues, or feedback itself might justify removal or change. That condition complicates the idea of an archive.

A public platform can preserve institutional memory, but it must also distinguish memory from guarantee. The presence of an old article shows that a claim, prototype or invitation was made; it does not establish that the underlying system remained available, correct or supported.

Kühne's public contribution can be located in these distinctions. She told audiences that RIPE Labs was a platform and a tool for the community, presented examples of technical work, and made the conditions of experimentation explicit. She did not promise that public exposure would produce adoption. The strongest account of her role is thus not that she converted every experiment into a service.

It is that she helped create a recognisable route by which experiments could be offered for scrutiny without being mislabelled as finished products.

Institutional support mattered throughout. A later RIPE NCC planning document describes the organisation as maintaining RIPE Labs as a platform for community research and ideas. That wording places the platform inside RIPE NCC activity and resources while retaining a community purpose. Kühne could curate and explain the surface; the organisation, participating teams and contributors supplied the capacity and material that kept it from being a one-person publication.

The public invitation also changed the address of technical work. An internal prototype can be known mainly by its builders. A RIPE Labs article gives it a stable public explanation that can be linked from a presentation, a meeting notice or a mailing list. This does not automatically create a community. It does create the possibility of a shared reference.

People who did not attend the same meeting can inspect the same description, and later readers can recover what was being tried at a particular time.

Publicness also changes the order in which accountability can begin. A finished service is usually judged after an institution has already chosen its architecture, language and operating assumptions. A visible experiment gives outsiders an earlier object to question, even if no public response can be shown to have altered it. That distinction keeps the claim modest but useful. RIPE Labs could not guarantee that scrutiny would produce a better tool.

It could reduce the problem of scrutiny arriving only after provisional choices had hardened into an official surface. The value lay in making questions possible at a different stage, not in claiming that every question received an answer.

The cases that followed matter because they show several versions of that address. RIPEstat used repeated demonstrations and a meeting bridge. Active measurement used a participation request and later public scale reporting. Infrastructure mapping used collaborative explanation, technical detail and working-group distribution. None proves that publication caused the eventual outcome. Together they show what Kühne's public-facing remit was capable of connecting.

RIPEstat and the value of a serial demo

RIPEstat is the clearest case of publication organised as a sequence rather than a launch announcement. In May 2011, Kühne published RIPEstat Live Demo #2, exposing functionality to RIPE Labs readers in a feedback-oriented format. By August, Mark Drayton presented Live Demo #5. The change of byline is important. It shows a named colleague using the same public format and prevents a series associated with RIPE Labs from becoming evidence of Kühne's technical ownership.

A serial demo does something different from a product description. It lets a public explanation track a changing tool. Each instalment can focus attention on a feature or question without requiring the audience to treat the whole service as complete. Repetition also creates an expectation that explanation will continue. Readers can see that a tool has a development history rather than appearing suddenly in its final form.

The tenth demonstration connected that rhythm to a community meeting. RIPE NCC's notice for RIPEstat Live Demo #10 placed it during RIPE 63 in Vienna and invited remote participation, input, feedback and suggestions. The public article and the meeting room were not separate channels. The demonstration could reach attendees and remote participants, while the notice gave the event a public reference that survived beyond its scheduled time.

This bridge is where Kühne's community-facing work becomes operationally intelligible. Her role was not to substitute editorial visibility for engineering. It was to help give technical work a route toward the people likely to question it. A RIPE meeting gathered operators and community participants; RIPE Labs could prepare, extend or preserve the encounter. Serial publication made the tool discussable between meetings, and a live demonstration made the discussion less abstract.

The evidence does not reveal the return path. It does not identify a comment that changed RIPEstat, quantify participation, establish data quality or show that the demo series caused adoption. Even the tenth instalment is evidence of continuity in public presentation, not evidence that every invitation produced useful feedback. This gap should remain part of the account. A feedback interface can be real and still leave its effects difficult to observe from public records.

Current RIPE NCC documentation describes RIPEstat as its information service for Internet-related data and analytics concerning address space and autonomous systems, drawing on numerous datasets.

That present service boundary matters only in a limited way here. It establishes RIPE NCC as the operator and shows that the name belongs to a continuing institutional service. It cannot be projected backwards as proof that Kühne caused durability, chose the architecture, maintained the data or secured adoption.

The proper distribution of credit is therefore specific. RIPE NCC teams built, operated and documented the service. Drayton and other named staff presented their work. Meeting organisers and remote participants created the setting for exchange. Kühne demonstrably authored at least one demo article and worked through RIPE Labs' editorial and community surface. She helped make the tool visible and discussable; the record does not make her its engineer.

That boundary reveals why public explanation is more than promotion. Promotion tends to compress uncertainty and direct attention toward a conclusion. A live-demo series can instead expose change over time and invite questions. It can show that a service has people behind it, features under development and a relationship with a community. When it works, the publication surface does not borrow credit from engineering. It makes engineering easier to encounter on its own terms.

RIPEstat also demonstrates how institutional memory is assembled. A current service page tells readers what the service is now. The older demo articles tell them that its public explanation was once incremental and that multiple people carried it. The meeting notice shows an attempt to connect online and live participation. Together those records preserve a history of presentation without proving a history of impact. That is a useful, disciplined form of accountability.

RIPE Atlas and participation before scale

The active-measurement case begins with an invitation to do something, not merely to read. In May 2010, Daniel Karrenberg published “Active Measurements: Hosting a Probe” on RIPE Labs. It described early active-measurement probes and proposed that people host them in their networks. The article is evidence of a technical author using the platform to expose a participation mechanism. It is not evidence that Kühne designed the probes or the measurement system.

That difference matters because a distributed measurement network depends on action beyond the institution operating it. A public explanation has to make the requested participation legible: what is being proposed, what a host would do, and why distributed vantage points matter. RIPE Labs could place that invitation where an operator community might find and discuss it. The platform did not itself install a probe, maintain a host's network or operate the resulting infrastructure.

Seven years later, RIPE Labs recorded 10,000 active RIPE Atlas probes. The number establishes scale at a dated point. It does not establish geographic representativeness, measurement quality or the causal force of any article. It cannot tell us how many hosts first learned of the project through RIPE Labs, which outreach mattered most, or whether Kühne influenced a person's decision to participate.

The temptation to draw a straight line is strong: an early hosting invitation appeared on a platform Kühne maintained; a later measurement network reached a large probe count; therefore the publication or the editor caused the growth. The public record does not permit that conclusion. Between invitation and scale sit engineering, operations, funding, host decisions, hardware, software, support and years of community contribution.

Daniel Karrenberg, RIPE NCC engineering and operations teams, probe hosts and other contributors own the work the record assigns to them.

What the case does support is a narrower institutional mechanism. RIPE Labs made a request around experimental measurement public at an early stage. Later, the same publication environment could document scale and explain what had developed. The archive lets readers compare an early call for participation with a later organisational milestone. That before-and-after visibility is valuable even when causality remains unknown.

Current documentation describes RIPE Atlas as a RIPE NCC-operated global network of probes measuring Internet connectivity and reachability, with most collected data made freely available. As with RIPEstat, the present page establishes an operator boundary and a durable public-data service. It should not be used to turn the 2009–2020 editorial story into a personal success claim.

The Atlas example also sharpens the meaning of network-resource evidence. Measurements become useful to a wider public only when people can understand where they came from, what infrastructure produced them and what limits attach to them. The publishing surface cannot guarantee the evidence. It can provide the explanations, participation history and institutional context needed to interrogate it.

Kühne's role belongs around that context. As RIPE Labs editor, maintainer, curator, author and presenter, she helped sustain a place where a technical invitation could be published and later revisited. The evidence does not show her operating measurements or directing probe growth. It shows work at the interface between an institution with technical capacity and a community whose distributed participation could not be assumed.

That interface is a form of accountability because it preserves the names and stages that a success story might otherwise flatten. The early article belongs to Karrenberg. The operation belongs to RIPE NCC. The network's reach depends on probe hosts and contributors.

The public platform belongs to a larger institutional and community effort in which Kühne held a documented editorial role. Ten thousand is not a personal score. It is a reminder of how many separate contributions a public technical system can contain.

Mapping infrastructure in public

RIPE IPmap and OpenIPmap offer a less durable but more revealing test of public technical explanation. In 2015, Jasper den Hertog described RIPE IPmap as a collaborative approach to mapping Internet infrastructure through measurement and community contribution. The byline assigns the explanation to den Hertog. The collaborative framing makes the public important, but it does not transfer engineering ownership to the platform's editor.

In 2018, a public archive of the Measurement, Analysis and Tools Working Group recorded Kühne announcing an OpenIPmap article to that audience. This is a modest but unusually concrete act. It shows material moving from RIPE Labs toward a working group concerned with measurement and analysis. No private editorial history is required to see the bridge: an article existed, and Kühne publicly carried notice of it to a relevant community channel.

In 2019, Massimo Candela published an under-the-hood account of RIPE IPmap. The progression from collaborative framing to technical detail shows how a public platform can hold different levels of explanation under named authorship. One article can invite contribution; another can expose mechanisms. Readers can inspect not only what a tool is supposed to do but how a technical author says it works.

The available evidence is weaker on what happened next than it is for RIPEstat or RIPE Atlas. There is no comparable current-service boundary here that should be used to claim durable success.

The sources do not establish production outcomes, geolocation accuracy, adoption or which community contributions changed an implementation. Open publication and working-group distribution are the supported result. Anything larger would replace evidence with the attractive assumption that visibility guarantees use.

That limitation makes the case valuable rather than dispensable. Experimental publishing should be judged partly by whether it can hold work whose future is uncertain. If only durable services remain worthy of discussion, the archive becomes a retrospective winners' list.

RIPE Labs' original promise was broader: prototypes, analyses, ideas and tools could be made public while their status was unresolved. The mapping articles fit that promise precisely because their public value does not depend on a triumphal ending.

Credit remains distributed. Den Hertog authored the collaborative explanation. Candela authored the technical account. RIPE NCC teams and tool maintainers held the engineering surface. Working-group participants formed the prospective discussion community. Kühne's evidenced action was editorial and distributive: she worked through RIPE Labs and carried an item into a related public forum. The platform made the encounter possible without making her the author of everyone else's work.

This case also shows why information architecture and attribution are connected. When articles, authors, demonstrations and community notices can be found together, a reader can reconstruct who said what and when. When they cannot, the institution is tempted to substitute a simplified anniversary narrative. Public technical knowledge remains accountable only if the archive preserves the seams between platform, author, operator, contributor and audience.

What a thousand articles costs

By January 2018, Kühne marked the 1,000th RIPE Labs article. Her reflection described a broad range of technical material from RIPE NCC staff and external contributors. By the 2020 retrospective, RIPE Labs held more than 1,200 articles after roughly eleven years. Those figures show that an experimental publication had become a substantial archive. They do not show that every article was influential, accurate, widely read or personally selected and edited by Kühne.

Scale changes the nature of the editorial problem. A new platform needs material and contributors. A thousand-article platform needs routes through accumulated material, consistent author identities, usable categories, functioning comments, links that retain meaning and explanations of whether an old prototype became a service or disappeared. The larger the archive becomes, the easier it is for public memory to turn into public clutter.

The redesign record is candid about that pressure. Davies's account says RIPE Labs had not always worked as the feedback mechanism intended and describes the hand-off and a new look after more than a decade. An external design case study likewise frames the later work around community engagement, feedback, comments, discussion and renewed information architecture. The portfolio has its own promotional perspective, so it cannot serve as an independent verdict on the institution.

Paired with the official retrospective, it supports the narrower conclusion that openness required maintenance and redesign.

This is not a story of a failed platform corrected by a perfect successor. The public evidence does not provide comparative engagement results or show that later changes solved the earlier limits. It establishes that the feedback ideal remained unfinished. A public comment feature can exist without producing sustained discussion. A large archive can preserve memory while making discovery harder. A redesign can respond to those problems without proving that they have disappeared.

The maintenance cost is partly editorial. Old articles need enough context to prevent a prototype from being mistaken for a current service. Migrated author pages need continuity without false precision. Technical posts need persistent attribution. The platform has to serve newcomers who do not remember the original discussion as well as participants who do. Each requirement adds labour after the initial act of publication.

It is also institutional. The RIPE NCC activity plan's maintenance language makes clear that the platform depended on organisational resources. Teams produced and explained tools; contributors supplied research and operational experience; engineers and operators maintained services; community members hosted probes and joined discussions. An editor could coordinate a public surface, but the surface had value only because a wider institution and community continued to place work into it.

The article count can therefore be read in two directions. It is evidence that the invitation to publish produced a large body of material. It is also a measure of future obligation. Every additional item creates another claim on search, navigation, migration, link maintenance and historical interpretation. Public knowledge is not made durable merely by leaving pages online. It has to remain findable and correctly framed.

The same applies to feedback. Inviting comments creates an expectation that someone will notice, interpret and route them. The public sources do not expose who handled individual comments, what was rejected or which suggestions became changes. That absence blocks claims about Kühne's private editorial decisions. It also highlights a structural cost: a feedback mechanism requires attention behind the visible form, and that attention is difficult to prove from the form alone.

In 2020, Alun Davies took over the editor role from Kühne. The hand-off coincided with her move to the RIPE Chair role, but that later period is not needed to explain the RIPE Labs work. The relevant fact is succession. A platform closely associated with one community builder passed to a named editor who could assess its accumulated strengths and limits. Continuity depended on the work becoming transferable rather than remaining an extension of one person's presence.

That endpoint clarifies Kühne's contribution. She did not leave behind proof that public feedback always worked. She helped leave behind a public object large and important enough to require succession, redesign and more explicit attention to usability. The archive's problems do not cancel the work. They show that making experiments public is not a single launch decision. It is a long obligation to keep uncertainty, attribution and participation legible.

What can, and cannot, be made personal

The person-level case for Kühne rests on visible acts and bounded roles. She took a Community Builder remit in the period when RIPE Labs launched.

She authored an explanation of its purpose, presented it to operator and standards-adjacent audiences, maintained and edited the platform, published demo material, distributed a mapping article to a working-group audience, marked an archive milestone and handed the editor role to Davies. Later profiles reasonably compress that record into creator and curator. The underlying evidence remains collective.

What cannot be made personal is just as important. The public material does not show Kühne writing platform software, designing probes, operating RIPEstat or RIPE Atlas, developing mapping algorithms, assuring data quality or causing adoption. It does not show her editing every article or deciding every acceptance. It does not establish which comments changed a tool. It cannot convert the continued existence of current services into proof of her influence over their durability.

Those limits are not ceremonial disclaimers added after a flattering story. They define the kind of contribution being examined. Internet infrastructure institutions need people who build and operate systems.

They also need people who make the state of those systems intelligible to communities whose cooperation, criticism or evidence may matter. Confusing the two roles gives an editor technical credit she did not claim in the available record. Treating the second role as mere publicity fails to recognise the work required for public scrutiny.

RIPE Labs made several institutional surfaces meet. A developer or analyst could publish under a name. An operator could encounter a prototype before it became a supported service. A meeting participant could continue a discussion through a public reference.

A remote reader could inspect a demonstration. A working group could receive notice of a detailed article. A later editor could recover the history of the platform and identify the places where its feedback and information architecture had fallen short.

The resulting public value is practical. Early explanation can expose limitations before institutional memory smooths them away. Serial demonstrations can make change visible. Participation requests can show that infrastructure depends on hosts and contributors beyond the operating organisation. Technical detail can make a tool open to inspection. Archives can preserve how a service or experiment was once described, giving later readers evidence against which to judge change.

None of these benefits is automatic. Publicity can pass without response. Comments can fail to become discussion. Archives can grow opaque. Counts can be mistaken for impact. Current service pages can tempt readers to rewrite a contingent experiment as an inevitable success. The RIPE Labs record is valuable partly because it preserves enough friction to resist that rewrite: the broken laptop, the disclaimers, the different bylines, the uncertain mapping outcome, the redesign and the hand-off.

Institutional legitimacy is often discussed as if it were produced only through formal authority. In technical communities it also depends on whether people can see how knowledge was made available, who authored it, which institution operates the service and what uncertainty remains.

RIPE Labs did not settle those questions. It gave them a public place to be asked. RIPE NCC's resources made that place possible; engineers, analysts, hosts and contributors gave it substance; the wider community gave the invitation meaning.

Kühne's documented work was to help maintain the crossing point. Her importance does not require a claim that she built the systems passing through it. It lies in the sustained effort to make prototypes, measurements, demos and operator analysis readable as public work while preserving their unfinished status and distributed authorship. The strongest evidence of that effort is not a single successful tool.

It is the repeated appearance of named technical work before a community, with enough context to invite attention and enough boundaries to resist false ownership.

The lesson of the Moscow laptop therefore survives the archive that followed. An experiment can be shown before it is stable, but only if the showing carries honest information about who made it, what it can do and why it might fail. Making that encounter possible is not the same as making the experiment.

Between 2009 and the 2020 hand-off, it was the work by which Kühne made a distinctive contribution: giving unfinished Internet-infrastructure knowledge a public life without claiming that its outcomes belonged to her alone.