Summary
- Five DISCUSS positions filed by Éric Vyncke are printed in full on IETF Datatracker history pages, with the date each was entered and, in four cases, the date his own position changed.
- Not one of the pages reviewed on 22 November 2026 prints the text change that answered those four objections, and the fifth hold stood until its draft expired.
Five holds, four clearances
On 5 November 2026, at revision 23 of draft-ietf-oauth-sd-jwt-vc, a document history table prints a line that has become ordinary in IETF review: "New position, Discuss, has been recorded for Éric Vyncke." Twelve days later the same table prints the mirror image: "Position for Éric Vyncke has been changed to No Objection from Discuss" (history page). Between those two lines sits his objection, in full, in two numbered sections asking why the type-matching sentence and the status-check sentence are SHOULDs rather than MUSTs.
The ballot page tells the same story from the other end and with less detail. It prints "Éric Vyncke (was Discuss) No Objection", then "Comment (2026-11-17) Sent" (ballot page). Start at the ballot and you learn that a hold existed and is gone. Start at the history and you learn what the hold said. Neither page shows what in the document changed between revision 23 and revision 27.
The clock is printed; the edit is not
The same shape repeats three more times, each with its own dates. On draft-ietf-tvr-schedule-yang the hold is entered on 13 April 2026 at revision 08 and changed on 11 June 2026 at revision 12, six days before the IESG approves the document on 17 June 2026 (history, ballot). On draft-ietf-nvo3-rfc7348bis: entered 2 September 2026 at revision 05, changed 9 October 2026 at revision 11 (history, ballot). On draft-ietf-teas-yang-te: entered 4 February 2026 at revision 41, changed 30 March 2026 at revision 44 — and the record prints that revision 44 was uploaded on 26 March, four days before the position moved (history, ballot).
No interval is printed anywhere on those pages. Subtracting the printed filing date from the printed clearance date is my own arithmetic, not the record's: 12, 59, 37 and 54 days.
The fifth record is where the arithmetic stops. draft-ietf-6man-eh-limits carries his DISCUSS from 11 April 2025 at revision 19, and the ballot summary prints "Needs a YES. Has 5 DISCUSSes" above his entry, "Discuss (2025-04-11) Sent for earlier" (ballot). No later position change appears on the reviewed pages. The document expired on 2 November 2025 and all action holders were removed (history). Again mine, not the record's: that gap is 205 days.
"Blocking" and "can be convinced otherwise" in one instrument
The statement his objections cite prints two sentences that a reader should hold side by side. The first: "These are blocking, and you do need to address them to move the document forward. It’s your responsibility to follow up, etc. Consider this a MUST." The second: "Once the AD (or ADs!) holding the DISCUSS position(s) are satisfied, they will clear their DISCUSS position" (Handling Ballot Positions).
His own objection bodies show how those two sentences sit together. Each is headed "DISCUSS (blocking)", and each opens by quoting the same definition back at the reader: a DISCUSS ballot is "a request to have a discussion on the points below; I really think that the document would be improved with a change here, but can be convinced otherwise." A MUST to address, and a mind that can change. Both sentences belong to the same instrument.
A second statement, dated 22 May 2025, prints the threshold and the exception: approval requires one Yes, at least two thirds of non-recused ADs voting Yes or No Objection, and "no 'Discuss' ballot positions" — with a Single Discuss procedure under which a lone unsupported hold, in a document that otherwise has the votes, leaves "the document simply approved, essentially overriding the single Discuss-holding AD" (IESG Ballot Procedures). The hold is real and not absolute. No reviewed page states whether that override was ever used on these five documents.
The diff that sits one row away
On the draft-ietf-tvr-schedule-yang ballot page, another area director's comment carries an author-tools iddiff link comparing revision 08 with revision 10. On the nvo3 ballot page, another director's comment compares revision 05 with revision 10. Vyncke's entries on both pages carry no diff link. The affordance exists on the very page where his entry lacks it. Nothing in the retrieved text explains the difference, and this piece does not supply a reason; it records what each entry prints.
What he actually asked for
His objections name their targets with unusual precision. In the SD-JWT text he asks "Why not a MUST" for the type-matching rule, then adds: "please provide guidance for the implementers about when to bypass the SHOULD or what are the consequences of bypassing it." On the TVR model he quotes the delay leaf's range "0..16777215" and answers his own question — "16777215 microseconds is just 16 seconds" — before writing, "I find the limit very limiting and lacking a rationale." On the TE YANG model he sets the abstract's claim to be "independent of any technology or dataplane encapsulation" against a model that is about LSPs, objects to IPv4-only text for extended-tunnel-id with the line that the draft "should not be published in 2026 as it does not support IPv6", and points out that the IPv6 addresses named in an appendix example do not appear on figure 10. On the VXLAN revision he asks for the correct BCP14 template, a reference for the UDP checksum guidance, and one more thing: "the implementers and operators are lost in the absence of guidance about the 3 'SHOULD'."
Each of those is a change a stranger could look for in a diff. What the record does not do is tell the stranger where to look.
The reviewer who says what he did not check
The SD-JWT objection ends with a sentence worth keeping beside the rest: "I did not check/mention the other SHOULD/RECOMMEND in the document and I am trusting the authors and responsible AD to address them." That is a reviewer writing his own delegation into the record. It leaves two satisfaction judgments in play — his and the authors' — and only his leaves a dated line on the ballot.
Where a stranger's reading has to stop
The AD-status page prints a table headed "Blocking positions held by Éric Vyncke", with columns for Document, Status, Responsible AD and Discusses (AD page). It lists current positions. It prints no objection text and no original filing date, and the table came back only partially in the material reviewed, so it is not a census of his DISCUSS history.
The limit of this piece is the limit of that record. No reviewed page prints the answering change; a revision uploaded before a clearance is a sequence and not proof of what answered what; the mail-archive threads linked from the ballot comments were not opened here; the 6man ballot names none of the other holders behind its five; and no page says whether the Single Discuss override was ever invoked. Every quotation above is a provider-reported excerpt from thirteen Datatracker pages retrieved on 22 November 2026, not a rendered page inspection, and every absence claim is confined to those thirteen.
The record is superb at what was decided and silent about what changed. Anyone who treats a cleared ballot as evidence that a document improved is relying on the half that has an author and a date. The other half has neither.
Member Briefing
Read the full trend briefing
Members read the full briefing with the source notes behind every finding.
Strategic Circle
A daily briefing in your language, built from changes we have verified.
Join for USD 20/monthCancel anytime.
Members also get the full research library, a daily briefing on the networks they follow and the private forum.






