**Session Date/Time:** 23 Jul 2026 09:30

# [CBOR](../wg/cbor.html)

## Summary

The CBOR Working Group met at IETF 120 to discuss the status and next steps for two primary documents: `draft-ietf-cbor-serialization` (CBOR Serialization and Determinism) and `draft-ietf-cbor-edn-literals` (Concise Diagnostic Notation). 

`draft-ietf-cbor-serialization` is approaching readiness for Working Group Last Call (WGLC) following an upcoming -08 revision. For `draft-ietf-cbor-edn-literals`, the working group agreed to adopt a structured feature-winnowing process on the mailing list to establish a stable technical baseline, postponing non-technical debates (such as document naming) to the end of the cycle. Both documents will continue to progress in parallel.

---

## Key Discussion Points

### CBOR Serialization (`draft-ietf-cbor-serialization`)

Laurence Lundblade presented a status update using the slides "[CBOR Serialization Status Update](https://datatracker.ietf.org/meeting/126/materials/slides-126-cbor-cbor-serialization-status-update-00)".

*   **Changes in draft-07**:
    *   Added sample source code for float encoding (preferred plus and deterministic serialization) adapted from Carsten Bormann’s work.
    *   Replaced specific CDDL control operators (`.pref-p`, `.or-b`, `.dtrm`) with a single `.serial` control operator applicable to any CBOR item.
    *   Reworked and streamlined sections on general and special serialization, including when to use preferred plus.
*   **Open Issues & Discussion**:
    *   **Signing and Hashing**: Laurence Lundblade noted inconsistencies in how the bundle protocol, c509, COSE, and other signing designs handle byte string wrapping. An appendix discussing these trade-offs is prepared in a pending GitHub PR and will be shared on the mailing list.
    *   **Nontrivial NaNs**: Addressed feedback from Fay Amacker regarding whether libraries should normalize or reject nontrivial NaNs. The working group consensus is that this decision should be left to the individual library or application. 
    *   **Language-Specific Mapping**: Carsten Bormann suggested that platform-specific mapping issues (such as NaN handling or missing data types) might be best addressed in a separate document targeting library implementers and protocol designers. Paul Hoffman agreed that an independent document tracking language-by-language and epoch-by-epoch behaviors would be useful. Laurence Lundblade and Paul Hoffman agreed that the current PR resolves the immediate issue for this specification and should conclude the NaN discussion for this draft.
    *   **CDDL Big Numbers**: Laurence Lundblade noted a minor upcoming CDDL representation tweak for big numbers stemming from a review of the c509 protocol.

### EDN Literals / CDN (`draft-ietf-cbor-edn-literals`)

Christian Amsüss introduced the status of `draft-ietf-cbor-edn-literals`, noting that version -27 (submitted after the deadline) incorporates significant editorial restructuring requested by reviewers.

*   **Triage and Feature-Winnowing Process**:
    *   The chairs proposed using a structured triage workflow on the mailing list (based on suggestions from Joe Hildebrand) to determine which features to keep, modify, or remove. The goal is to start from a baseline matching RFC 8949 and RFC 8610 and selectively remove features lacking consensus.
    *   Carsten Bormann supported the proposed procedure, noting that while there may be minor undefined corner cases (such as the proposed two-person threshold), the framework is reasonable.
    *   Christian Amsüss emphasized that the process is a collaborative guideline to help the working group reach consensus without dragging on indefinitely.
    *   In the session chat, Martine Lenders requested that the path forward for the `e''` extension (as described in `draft-ietf-cbor-edn-e-ref`) and general extensibility remain intact to support human-readable CDN representations for DNS records.
*   **Sequence of Discussion**:
    *   Paul Hoffman proposed that non-technical debates, such as what to name the final notation, be explicitly deferred to the very end of the process to ensure the group maintains technical focus.

---

## Decisions and Action Items

*   **Decision**: Process both `draft-ietf-cbor-serialization` and `draft-ietf-cbor-edn-literals` in parallel.
*   **Decision**: WGLC for `draft-ietf-cbor-serialization` can begin while the `draft-ietf-cbor-edn-literals` triage process is active on the list.
*   **Decision**: Defer non-technical discussions (e.g., final naming of CDN/EDN) to the end of the `draft-ietf-cbor-edn-literals` development cycle.
*   **Action Item**: Laurence Lundblade to share the signing and hashing trade-offs PR on the mailing list for review.
*   **Action Item**: Laurence Lundblade to publish `draft-ietf-cbor-serialization-08` by July 27, incorporating the covert channels security section (provided by Rohan Mahy), Fay Amacker’s minor editorial corrections, the signing/hashing appendix, NaN text clarifications, and the CDDL big numbers tweak.
*   **Action Item**: Chairs to initiate the structured winnowing process for `draft-ietf-cbor-edn-literals` on the mailing list.
*   **Admonition**: Participants are reminded to use correct and descriptive subject lines on the mailing list to ensure discussions remain easy to follow (supported by Laurence Lundblade and Martine Lenders).

---

## Next Steps

1.  **`draft-ietf-cbor-serialization`**: Initiate Working Group Last Call (WGLC) once version -08 is published.
2.  **`draft-ietf-cbor-edn-literals`**: Kick off the feature triage process on the mailing list to establish the baseline draft.