CELLAR

Summary

The CELLAR working group met to introduce the new Area Director (AD), Charles Eckel, and discuss the status of several documents in the publication pipeline. The group reviewed three errata, provided updates on the FFV1 and Matroska specifications, and held a significant technical discussion regarding the versioning and "Updates" relationship for future Matroska v4 additions and EBML schema changes.

Key Discussion Points

AD Transition and Document Status

Spencer Dawkins introduced Charles Eckel as the new Area Director for the working group.
draft-ietf-cellar-tags (Matroska Tags): Charles Eckel noted that IANA provided feedback regarding language in sections 3.2.1 and 6.1. A small clarification is needed to ensure IANA's interpretation matches the group's intent. Charles Eckel will post a proposal to the mailing list.
draft-ietf-cellar-codec (Matroska Codec): Charles Eckel confirmed that a question regarding a downref was resolved. He will perform a final review of the IANA considerations before moving the document forward to the IESG.

Specification Updates

Chapter Codecs: Steve Lhomme reported no change. Implementation of the Javascript version in video players remains uncertain. Discussion occurred regarding the archival nature of the DVD specification, which is currently only available physically at the Japan National Library, complicating the mapping process.
draft-ietf-cellar-ffv1-v4 (FFV1v4): Jérôme Martinez noted a new proposal for RGB bayer support on the FFmpeg-devel list. Michael Niedermayer has a different proposal for raw LSB (Least Significant Bit) storage. Jérôme Martinez indicated that while floating-point support is available in FFmpeg, the specification text still needs updating. The group aims for a review of the text in late 2024.

Errata Review

Spencer Dawkins led a review of three errata:

RFC 8658: The erratum suggested adding detail regarding recurring elements. Steve Lhomme recommended rejection, as default values already cover the cases where elements are not present. The sense of the room was to recommend Rejected.
RFC 8854: The erratum suggested explicitly adding "unbounded" as a value for MaxOccurs. Steve Lhomme argued that the absence of the attribute already implies "unbounded" per the current text. The sense of the room was to recommend Rejected.
RFC 8810: The erratum corrected a typographical error where "shifted right" was used instead of "shifted left" for pads. The group agreed this was a valid correction. The sense of the room was to recommend Verified.

Matroska v4 Additions and EBML Evolution

Steve Lhomme proposed moving away from a "Matroska v5" designation in favor of "Matroska v4 Additions" to maintain backward compatibility.
A discussion ensued between Steve Lhomme, Robert Sparks, and Spencer Dawkins regarding whether new additions should be published as a "bis" document (obsoleting the base spec) or as "Updates" documents.
Robert Sparks argued that the "Updates" relationship is better for the IETF community to review, as it isolates the new changes rather than requiring a re-review of the entire 150-page base specification.
Steve Lhomme noted that "Updates" also provides visibility at the top of the original RFC on the IETF datatracker, signaling to implementers that additional elements are available.
The group discussed applying this "Updates" logic to draft-ietf-cellar-tags and draft-ietf-cellar-codec as well, even if they do not change the base text of RFC 9559.

Decisions and Action Items

Decision: The group recommends Rejected for errata on RFC 8658 and RFC 8854, and Verified for the erratum on RFC 8810.
Action Item: Charles Eckel to post proposed wording changes for the Tags draft to the mailing list following coordination with IANA.
Action Item: Steve Lhomme to continue the discussion on the mailing list regarding the structure of Matroska v4 Additions and the use of the "Updates" relationship for current and future drafts.

Next Steps

Refine the "Updates" relationship metadata for the Tags and Codec drafts before they proceed further in the publication process.
Jérôme Martinez to review FFV1 proposals (RGB bayer and raw LSB) and update the draft.
Schedule the next interim meeting to continue work on the Matroska and EBML revisions.

Automatic IETF Minutes