Summary

  • RIPE policy proposal 2014-01, "Abandoning the Minimum Allocation Size for IPv4", appears on the registry's proposal page with an Author field naming Carsten Schiefner, a publication date of 17 March 2014, a state of Accepted, and an acceptance date of 28 July 2014.
  • The same page names no individual as having accepted the proposal, and the policy document it points to — ripe-621, published 28 July 2014 — now displays the state Obsoleted. Documented authorship is present on the record; a personal decision is not.

One page, one name, and a great deal of white space

The page for proposal 2014-01 is a working document of the region's policy archive, and it is short. It opens with the title "Abandoning the Minimum Allocation Size for IPv4" and a longer descriptive line that frames the work as the "Abandonment of the Minimum (Sub-) Allocation Size concept in the 'IPv4 Address Allocation and Assignment Policies for the RIPE NCC Service Region' document". Two sentences follow: "This policy proposal has been accepted" and "The new RIPE Document is: ripe-621". Then come the fields — Author, Publication date, State, Accepted.

The Author field reads Carsten Schiefner. Nothing else on the page is a person. The remaining values are a date of 17 March 2014 for publication, a state of Accepted, a date of 28 July 2014 for acceptance, and a document number.

That distribution is worth pausing on, because it is the entire attributable content of the record. Anyone who wants to say what this individual did in the RIPE policy process can point to one field and be right. Anyone who wants to say what this individual decided has to add something the page does not contain: a chair's ruling, a tally, a minute, an appointment, a signature on a decision instrument. None of those is present. The gap between the two sentences is not an accident of a thin page; it is a difference in record type. An authorship line records who is answerable for a text.

An acceptance state records that a collective body reached a position. Those are different acts by different kinds of actor, and the archive keeps them in adjacent rows.

What the proposal said it would do, in the page's own framing

The page does carry a description of the effect, and its attribution matters. A passage on the page reads: "It is the RIPE NCC's understanding that this proposal would remove the minimum size for IPv4 allocations and sub-allocations. The proposal would allow LIRs to split their existing allocations to any size, and to create more specific sub-allocations of any size."

Three things follow from that quotation, and only three. The proposal's subject was a floor: a minimum size below which an allocation or sub-allocation could not go. Its intended effect was to let a local internet registry divide what it holds to any granularity, and to create more specific sub-allocations of any size. And the sentence is framed as the secretariat's understanding of the proposal, not as the author's own words and not as the text of an adopted rule. A reader who wants the operative policy language is being pointed elsewhere — to ripe-621 — and the page does not reproduce it. The difference is not pedantry.

The secretariat's reading of a proposal and the adopted instrument are separate documents with separate standing, and only one of them governs.

The instrument is dated, and it is now marked obsolete

The document the page names has its own record. The ripe-621 page displays a Publication date of 28 July 2014 and a State of Obsoleted. The publication date of the resulting document and the acceptance date of the proposal are the same day. From 17 March 2014, the proposal's publication date, to 28 July 2014, its acceptance date, is a span of 133 days. The page displays no dated step falling inside that span: no impact analysis date, no working-group milestone, no last-call window, no version history. The interval is real and its contents are unrecorded on the page as it stands.

Obsolescence carries its own lesson about durability. Whatever ripe-621 established in July 2014, the document's own record now marks it as superseded in status. The page as read does not display the name of the document that replaced it. So the policy answer to "what is the rule now" is not on either of these pages, and the honest formulation of the chain is short: an accepted proposal in 2014 produced a document in 2014, and that document is no longer current.

What the record establishes

Four statements survive scrutiny, each tied to a displayed value.

First, a proposal text is recorded under Carsten Schiefner's name, in an archive that assigns proposals individual numbered pages. Second, that proposal bears the title "Abandoning the Minimum Allocation Size for IPv4" and was framed as an amendment to a named IPv4 policy document. Third, the archive records a publication date of 17 March 2014 and an acceptance date of 28 July 2014, with the state Accepted. Fourth, the same page points to ripe-621 as the resulting document, and ripe-621's own record shows publication on 28 July 2014 and a current state of Obsoleted.

That is a real contribution, and it is narrower than a biography. An authored proposal that reached acceptance is a documentable act of participation: somebody put a specific change in front of a community, in writing, with a number and a date, and the community eventually recorded a position on it.

What it does not establish

Six things sit outside the record, and each is the kind of thing readers most readily assume.

The record does not name the person who decided. There is no acceptance signature, no chair's determination, no vote count. The record does not show the procedural steps between publication and acceptance, so no one can say from these documents what arguments were made, by whom, or how the position formed. The record does not state an office — no elected or appointed position, no term, no nominating or appointing instrument, no election result appears anywhere in it. The record does not state motive, employment, affiliation, nationality or relationship to any registry or company.

The record does not identify the document that superseded ripe-621. And the record does not show how many people contributed to drafting the text recorded under one name.

Each of those is an absence in these two records, not proof of absence in the world. There may be minutes, mailing-list threads, an impact analysis, a term of office. None of it is here, and an accountability piece cannot manufacture it by inference from a byline.

Why one byline matters beyond one name

Policy archives everywhere perform the same compression. A collective process, spread over months and many participants, is filed under an individual's name because a document needs an owner. That compression is useful: it gives the public something to hold, and it gives the author a stake in accuracy. It is also unstable, because the byline is the most legible field on the page and the decision fields are the least. Readers reach for the name; the record's own structure puts the consensus somewhere else.

The practical consequence is a pair of symmetric errors. Personalise the byline and you credit one person with a room's judgement — or blame them for it. Ignore the byline and you erase the single attributable act the record preserves. Holding both at once is not a rhetorical compromise; it is what the document set actually supports. The author line is where a person can be held to a text. The acceptance line is where responsibility passes into process.

A reader who keeps them apart can still say something true and specific about what Carsten Schiefner did here: he authored a proposal that the archive records as accepted, and the instrument that followed is now marked obsolete.