Summary

  • LACNIC's entry for LAC-2018-14 prints three authors — Jordi Palet Martinez, Edwin Salazar and Edmundo Cázarez — each annotated "Version [1, 2]", and the Version 1 view repeats the identical block, so the author set did not change between the two versions of the text.
  • The register does hold an organisation beside each name, but only on a version-comparison page: The IPv6 Company, Wifitelecom and NIC.MX. The detail page a reader lands on prints names and version tags alone.
  • When the proposal was withdrawn on 13 May 2019, the chairs' result announcement attributed the decision to "THE AUTHOR" in the singular — while LACNIC's own blog post of 30 May 2019 named all three as promoters.

The field that lists three names

The entry for LAC-2018-14, read on its own detail page, prints an Authors block consisting of three names, each followed by the same bracketed annotation. As printed: "Jordi Palet Martinez - Version [1, 2] Edwin Salazar - Version [1, 2] Edmundo Cázarez - Version [1, 2]" (LAC-2018-14 register entry).

That annotation is the register answering a question most readers never think to ask. It is not saying these three people wrote version 2. It is saying all three cover version 1 and version 2. Open the version-specific view and the block is character-for-character the same (Version 1 view). Open the Spanish rendering of version 1 and the same three names appear under "Autores", each marked "Versión [1, 2]", above the line "Última Modificación 24/10/2018" (versión 1, español).

I set out to test whether a name enters or leaves a proposal's Authors field as the text is revised. On this proposal the answer is no, and the register says so itself. The revision changed the policy text; it did not change who is printed as responsible for it.

The affiliation lives on another URL

The register does keep more than a name. Its comparison view, which sets version 2 against version 1, prints each author as a record with an organisation attached: Palet Martinez at The IPv6 Company, Salazar at Wifitelecom, Cázarez at NIC.MX (v2 against v1). Affiliations exist. They are simply not on the page a reader reaches first: the detail entry's Authors block, in the text returned for this piece, prints the three names and the version tags and nothing else.

So the register resolves the same authorship at two levels of detail, and a reader has to know that a comparison URL exists to see the second one. If you want to know who these three were when they filed the text, the landing page will not tell you. The diff view will.

The record says one author decided

On 13 May 2019 the policy development chairs posted the result for LAC-2018-14 v2. The proposal had reached consensus, the announcement says, but because it was mutually exclusive with LAC-2019-1 v1, "THE AUTHOR HAS DECIDED TO ABANDON IT" (chairs' result, 13 May 2019). Singular. Three names in the field, one actor in the sentence. The same announcement also heads its own item "LAC-2019-14 v2" while its subject line concerns LAC-2018-14 v2 — an inconsistency inside the document, quoted here exactly as it was printed.

LACNIC's own blog post of 30 May 2019 goes the other way, in the plural: "Promoted by Jordi Palet Martínez, Edwin Salazar and Edmundo Cázarez, the proposal on authorizing the transfer of IPv4 resources between LACNIC and other Regional Internet Registries generated a fruitful exchange among participants" (LACNIC blog, 30 May 2019).

Three resolutions of one authorship, then. A field that names three people and covers two versions. A comparison page that gives each of them an employer. A procedural record that reduces the decision to a single author and never says which one. None of the three reconciles with the other two, and all three come from the same organisation.

What the register does establish, unambiguously

The surviving text is LAC-2019-1, and its paper trail is complete. The entry prints Status Implemented, a last-call window of 13/05/2019 to 09/06/2019, ratification by the board on 17/06/2019, and ratified 20/06/2019, with the same three names each annotated "Version [1]" (LAC-2019-1 register entry). The chairs' own announcement of 13 May 2019 records the first step in that sequence: "THE PROPOSAL REACHED CONSENSUS AND HAS NOW ENTERED A FOUR-WEEK LAST CALL FOR COMMENTS PERIOD" (chairs' result, LAC-2019-1 v1). The Spanish rendering adds the date the policy actually took effect: "Implementada 21/07/2020" (LAC-2019-1, español). The abandoned twin carries its own ending: "Abandoned 13/05/2019". The policy list holds LAC-2018-14 at version 2.0 and LAC-2019-1 at 1.0, with 30/10/2018 and 28/02/2019 as their published dates (policy list).

Two dates in that record do not line up with each other, and each is printed on its own page. LAC-2018-14's "Last modified 27/02/2019" precedes the abandonment it describes by more than two months (register entry). The list's published date for the same proposal, 30/10/2018, differs from version 1's "Última Modificación 24/10/2018" (policy list, versión 1) — different fields, so not a contradiction, but not a sequence a reader can reconstruct either. For per-version submission dates you have to leave the register altogether and read the author's own table on the policy mailing list, which gives 24/10/18 for version 1, 27/2/19 for version 2, and 27/2/19 for LAC-2019-1 version 1 (author's date table, 21 August 2020). That table is a convenience, not a register field.

The limit, stated plainly

Neither detail entry, in the text returned for this piece, states which author drafted which sentence, any share of authorship, or which named person took the decision to withdraw (entry, version 1, versión 1, LAC-2019-1, LAC-2019-1, español). Scope that carefully: those five views, as delivered. The pages came back as extracted text, several of them truncated, and the version-selection widgets were not inspected.

What follows from the limit is practical. A citation of LAC-2019-1 that names one author is asserting something the register does not print. So is a line in a procurement memo or a conference slide that credits "the author" of a three-name proposal. The register records an outcome, a set of dates, and a field with three names in it; it leaves the assignment of responsibility to whoever writes the next summary.

The register can tell you, precisely and usefully, that Edwin Salazar's name covers both versions of a policy that reshaped address transfers across a continent. It cannot tell you what he decided, or what he drafted, or whether he was in the room when the text was pulled. For a document class that exists to fix the public record, that is the gap worth printing — because the next person to cite this policy will fill it in without noticing.