Session Date/Time: 10 Jul 2024 14:00

CBOR

Summary

This meeting focused primarily on setting the agenda for the upcoming IETF 120 meeting, reviewing the status of various CBOR working group and related documents, and identifying key discussion points for Vancouver. Technical challenges with slide sharing delayed the start. Several documents were confirmed for discussion at IETF 120, while others were noted as stable, out of the WG's hands, or not yet ready for discussion. A specific technical discussion occurred regarding the ABNF structure for edn-literals. The dates for interim calls between IETF 120 and 121 were also confirmed.

Key Discussion Points

Initial Technical Difficulties: The session started with brief technical issues related to the meetecho and data tracker interface, causing delays in sharing slides.
Document Status Tour (Carsten's Presentation):
- Documents Nearing Completion: cbor-time-tag, cbor-update-8610-grammar, and yang-sid are out of the WG's hands or in the RFC Editor queue, with minor editorial matters for time-tag and yang-sid. No IETF 120 discussion is anticipated for these.
- edn-literals: Waiting for AD review. The AD (Ari) is reviewing the updated draft and Rohan's remaining comment. A tentative discussion at IETF 120 was suggested if the comment is not fully resolved.
- cbor-dot-join: Completed WG Last Call. Authors believe comments are addressed. There was a discussion about whether to conduct another short WG Last Call or proceed directly to IETF Last Call. Concerns were raised about the "dot join" operator, which has no expectation of full implementation.
  - Technical Detail: Marco suggested a detailed review of changes before IETF Last Call, especially concerning the dot-join operator's practical implications.
- cbor-packed: The document is stable, but the WG wants more experience with it, referencing Christian's document on building tables by reference and the dns-cbor draft. Potential input from a conference paper on DNS CBOR use was noted.
- cbor-cde: Has seen significant mailing list traffic. A new "-04" version is expected before IETF 120. The document is considered ready for WG Last Call after further discussion.
- deterministic-encoding-explainer: An informational document providing guidance on deterministic encoding. Needs an editorial pass. The WG needs to decide whether to pursue this as a WG document or an Independent Stream Editor (ISE) document.
- cbor-numbers (draft-b-cbor-numbers): A new draft focused on the general handling of numbers in CBOR, intended to offload detailed number explanations from cbor-cde to streamline the latter.
- Independent Stream Document (related to CDE): Confirmed to be an independent stream document, leveraging the application profile concept from CDE. Tag 2011 registration was cleared. No WG action is needed.
- cbor-tags: A next version is expected before IETF 120 to address an "interesting technical dark corner" concerning importing specifications that add to sockets.
- cbor-edn: Recently adopted, now open for reviews. It is already being used by the gm-admin document in ACE. Marco gave positive initial feedback on its simplicity.
- cbor-draft-numbers: A separate draft (distinct from cbor-numbers) dealing with managing CBOR numbers in Internet Drafts, related to the ef document. Marco highlighted its utility for implementers needing clarity on provisional registration assignments.
- Yang CBOR Documents (yang-cbor-binary-data, yang-cbor-metadata): These documents aim to upgrade Yang CBOR definitions, for instance, by allowing binary data where Yang typically uses text. They require an editorial pass and attention from Netmod experts but are not prioritized for IETF 120.
Detailed Discussion on edn-literals ABNF:
- Ari requested clarification on Rohan's comment regarding "significant surgery" on the edn-literals ABNF.
- Carsten explained the current two-level ABNF approach: a general structure for plugins (application-oriented literals) and separate ABNF pieces for extensions (e.g., hex, Base64). This promotes pluggability.
- Rohan's suggestion was for a single, consolidated ABNF. While more streamlined for "bulk usage" implementations, it makes future plug-ins harder to integrate.
- Carsten noted that implementers (like Christian) are fine with the two-level approach, and it offers better expository clarity by separating concerns. A single-level ABNF can still be generated by implementations. Ari expressed a desire to gauge broader WG sentiment on this.

Decisions and Action Items

IETF 120 Agenda:
- cbor-dot-join: Will be a tentative agenda item for detailed review during document status discussion.
- cbor-packed: Will be an optional/provisional agenda item.
- cbor-cde: Will be discussed, focusing on the -04 version.
- deterministic-encoding-explainer: Will be discussed to decide its WG/ISE path.
- cbor-numbers (draft-b-cbor-numbers): Will be discussed to streamline cbor-cde.
- cbor-tags: Will be discussed if a new draft is available before IETF 120.
- cbor-edn: Will be discussed; Marco to provide a report on implementation experience.
- cbor-draft-numbers: Will be discussed for clarity on provisional registration assignments.
- yang-cbor and other informative documents: Not prioritized for IETF 120 unless a strong need arises.
Chair Action: The Chair will post a draft agenda for IETF 120 to the mailing list.
Interim Calls: The dates for interim calls between IETF 120 and 121 (August 21, September 4 & 18, October 2 & 16) were confirmed. The Chair will formally request these dates during IETF 120.

Next Steps

The Chair will finalize and post the draft agenda for IETF 120 to the WG mailing list.
Working Group members are encouraged to review the cbor-numbers (draft-b-cbor-numbers) document in preparation for its discussion at IETF 120.
Ari will continue reviewing edn-literals and seek broader working group feedback on the ABNF structure if needed.
The Chair will schedule the confirmed interim calls following IETF 120.