CBOR

Summary

The CBOR Working Group held an interim meeting to discuss the path forward for the Extended Diagnostic Notation (EDN) specification following its Working Group Last Call (WGLC), and to receive an status update on the CBOR serialization draft.

A key focus of the meeting was resolving a lack of consensus among the chairs and the working group on technical features in draft-ietf-cbor-edn-literals. The group debated a proposal to publish multiple EDN variants in a single experimental RFC, but ultimately coalesced around an alternative strategy: publishing a minimal, non-controversial baseline specification and deferring contested features to future extension documents.

Additionally, the author of draft-ietf-cbor-serialization reported that the draft is technically complete, with outstanding Security Considerations resolved. The working group plans to transition this document to WGLC shortly after the upcoming face-to-face meeting in Vienna.

Attendees

Joe Hildebrand, Ira McDonald, Paul Hoffman, Barry Leiba, Marco Tiloca, Rohan Mahy, Carsten Bormann, Dave Robin, Mikolai Gütschow, Laurence Lundblade, Вадим Гончаров.

Note: Вадим Гончаров reported in the chat that the MeetEcho link was blocked in Russia.

Key Discussion Points

1. Concise Diagnostic Notation (EDN) Post-WGLC Way Forward

Paul Hoffman presented slides on the EDN WG Last Call summary.

The WGLC for draft-ietf-cbor-edn-literals revealed that while there is universal agreement on the need to publish an EDN document and strong consensus on editorial improvements, there is a deep divide on several key technical features.

To break the deadlock, Paul Hoffman presented a proposal originally suggested by Area Director Andy Newton:

Publish a single Experimental RFC containing multiple EDN variants (each as a separate appendix with its own naming, MIME types, etc.).
The working group would only review the variants for technical correctness (i.e., making sure they can correctly specify CBOR) rather than trying to achieve consensus on a single feature set.

Opposition and Alternative "Minimalist" Proposal

Developer Pushback: Joe Hildebrand strongly opposed the experimental multi-variant proposal, stating that as an open-source library developer, he would refuse to implement multiple competing experimental specifications. He argued that this approach would fail to gather useful implementation experience and stated he would prefer no document over a multi-proposal document.
The Minimalist Path (The "Joe/Barry Proposal"): Joe Hildebrand proposed instead that the working group aggressively strip out all controversial features to produce a clean, minimal baseline specification. Any contested features could be introduced later in backwards-compatible extension documents. Joe Hildebrand expressed willingness to sacrifice any features he had previously championed to achieve this baseline.
Group Support: Barry Leiba, Rohan Mahy, and Laurence Lundblade spoke in support of this minimalist approach. Laurence Lundblade suggested the WG proceed feature-by-feature to establish a clear list of what is in and what is out.
Existing Dependencies: Carsten Bormann expressed concern that some specifications already use EDN features currently in the draft. He noted that developers of ~12 drafts/RFCs currently referencing EDN might have their work disrupted if those features are cut. Paul Hoffman requested that Carsten Bormann compile a specific list of these dependencies for the working group.
Mailing List vs. GitHub: Carsten Bormann suggested leveraging the GitHub repository to offload detailed technical discussions. Paul Hoffman and Barry Leiba strongly disagreed, emphasizing that final technical decisions must occur on the mailing list to ensure broad, open participation and to prevent participant attrition.
List Convergence: Mikolai Gütschow expressed surprise at the reported lack of consensus, noting that from his perspective, recent mailing list discussions had been constructive and quiet, moving toward convergence.

The Naming proposals presentation was deferred to the mailing list and future face-to-face sessions to prioritize the path-to-consensus discussion.

2. Serialization and Determinism Draft Update

Laurence Lundblade provided an update on the status of draft-ietf-cbor-serialization (v07):

Technical Updates: Added sample source code for encoding floats in Preferred plus and Deterministic formats. Replaced CDDL control operators .prefP and .dtrm with .cbor (feedback is requested on this change). Reworked explanatory text on general/special serialization.
Security Considerations: The draft's only remaining open GitHub issue has been resolved. Rohan Mahy drafted the Security Considerations text, and a pull request has been agreed upon by both Rohan Mahy and Laurence Lundblade.
Bundle Protocol: Laurence Lundblade noted that byte string wrapping issues related to the Bundle Protocol are being discussed on the list, but are not critical for the serialization draft. Rohan Mahy recommended keeping any reference to the Bundle Protocol minimal (one or two sentences max).
WGLC Readiness: Laurence Lundblade considers the draft functionally complete and ready for WGLC.

Decisions and Action Items

Decisions

EDN Strategy: The working group abandoned the experimental multi-appendix RFC proposal. Instead, they will pursue a two-step consensus process to publish a single, minimalist baseline specification for draft-ietf-cbor-edn-literals.
Mailing List Priority: All final technical consensus decisions must remain on the IETF mailing list, not GitHub.

Action Items

Chairs: Draft and post a formal proposal to the mailing list by Friday (or earlier) to launch the "minimalist baseline" consensus call.
Joe Hildebrand: Draft and submit a proposed rubric to the mailing list outlining the criteria for aggressively pruning controversial features from the EDN draft.
Carsten Bormann:
1. Compile a list of RFCs and active drafts that currently rely on specific EDN features.
2. Perform an editorial-only reorganization of draft-ietf-cbor-edn-literals (merging agreed-upon editorial PRs) and submit a clean revision before the Vienna publication deadline (target: Sunday). Crucially, this revision must not contain any technical changes.
Laurence Lundblade:
1. Share the proposed Security Considerations text for draft-ietf-cbor-serialization on the mailing list immediately.
2. Publish a revised version of the draft containing this text as soon as the IETF draft submission window opens.

Next Steps

IETF 125 (Vienna) Planning:
- The working group will hold its face-to-face meeting on Thursday of the Vienna meeting week.
- The revised draft-ietf-cbor-serialization will be discussed in Vienna. The chairs intend to launch a standard two-week WGLC immediately after the Vienna session.
- The EDN naming proposals (Naming proposals) will be addressed during face-to-face or subsequent interim sessions.
Cadence: The working group will maintain its biweekly interim schedule. The next interim will be held approximately two weeks after the Vienna meeting to review serialization WGLC feedback and advance the EDN baseline consensus.

Automatic IETF Minutes