**Session Date/Time:** 09 Feb 2022 15:00

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

## Summary

The CBOR working group met to discuss the status of the `file-magic` draft and a new proposal for "illusion of values" in CBOR diagnostic notation. The `file-magic` draft (`draft-ietf-cbor-file-magic-08`) was deemed ready for a shortened Working Group Last Call (WGLC). A new mechanism using three dots (`...`) for indicating omitted data in diagnostic notation, backed by a new CBOR tag (T888) for machine processing, was presented and received positive initial feedback.

## Key Discussion Points

*   **Audio Issues:** The meeting began with a brief discussion of a participant's audio setup causing an echo, which was subsequently resolved.
*   **IETF Vienna Attendance:** Participants briefly discussed their travel plans for the upcoming IETF meeting in Vienna, with some awaiting travel approvals.
*   **`draft-ietf-cbor-file-magic-08` Status:**
    *   `draft-ietf-cbor-file-magic-08` was submitted, primarily containing document cleanup (normative/informative distinctions) and tutorial changes.
    *   Carsten, as editor, confirmed that previously discussed historical statement changes were already incorporated in earlier versions (`0.14` or `0.15`).
    *   A sense of those present indicated no objections to progressing the document.
*   **Illusion of Values in CBOR Diagnostic Notation:**
    *   Carsten presented a proposal for adding an "illusion mechanism" to CBOR diagnostic notation. The motivation is to simplify examples in RFCs and other documents by omitting less important parts.
    *   **Proposed Syntax:** Three dots (`...`) would represent one or more omitted data items.
        *   Examples: `[1, 2, ..., 5, 6]` for arrays, `{1: "a", ...}` for maps.
    *   **String Illusion:** The `...` could also be used *within* text or byte strings (e.g., `"part1" ... "part2"`), leveraging RFC 8610 Appendix G.4 for concatenated strings. This would require diagnostic parsers to implement G.4.
    *   **Machine Processing:** To allow machine generation and parsing of examples with illusions, a new CBOR tag, T888, was proposed.
        *   When a diagnostic notation tool is configured to process illusions, it would convert `...` to `T888(null)` for simple omissions or `T888(["part1", null, "part2"])` for illusions within strings.
        *   This tag would signal that the data is intentionally incomplete.
    *   **Parser Configuration:** It was emphasized that a special flag would be needed when invoking diagnostic notation tools to enable illusion processing, preventing unintended interpretation of `...` or T888.
    *   **CDDL Interaction:** Carsten raised a question about potential interactions with CDDL, such as CDDL guiding or verifying illusions, but no concrete use cases were identified at this time. A participant noted the resemblance between the `T888` array structure and CDDL's `cat` operator.
    *   **Initial Feedback:** The proposal received positive feedback, particularly for its utility in machine-generated examples and ensuring that intentionally incomplete examples can be processed and verified without errors.

## Decisions and Action Items

*   **Decision:** The working group will proceed with a **shortened, one-week Working Group Last Call (WGLC)** for `draft-ietf-cbor-file-magic-08`.
    *   **Action Item:** Michael (Chair) will initiate the WGLC and prepare the IESG shepherd's write-up if no issues are raised during the WGLC.
*   **Decision:** Carsten received encouragement to proceed with developing the "illusion of values" proposal.
    *   **Action Item:** Carsten will implement the proposed illusion mechanism in his diagnostic package.
    *   **Action Item:** Carsten will draft a specification for the illusion mechanism, potentially as an addition to an existing extended diagnostic notation draft or a new standalone document.

## Next Steps

*   Monitor the `file-magic` WGLC.
*   Carsten will work on the implementation and draft for the illusion mechanism.
*   The next WG meeting is scheduled for two weeks from now.