Minutes interim-2023-cbor-11: Wed 14:00
minutes-interim-2023-cbor-11-202306281400-01

CBOR working group conference call, 2023-06-28

Other things we need to talk about this call

time-tag: Leave as Informational? Change to Standards Track? Does it matter?

Change to standards track was discussed at the 2023-01-25 interim, we
agreed that the document should be standards-track as opposed to
informational.

This is now implemented in draft-ietf-cbor-time-tag -07, which claims:

The present version (-07) picks up a few loose ends, completing
the processing of the WGLC input.

See also:
https://mailarchive.ietf.org/arch/msg/cbor/EGDpQs7A--NWqwqOgMZ2RzxHJQg

Closed issues:
https://github.com/cbor-wg/time-tag/issues?q=is%3Aissue+is%3Aclosed+

Are we ready to submit to IESG?
(no voice raised)
BL: Will move it ahead to the IESG

CB: Maybe we need a new WG Last Call.
BL: I don't think so, I think we have consensus. Do you think we need to
have one anyway?
CB: We can do it in parallel with other things needed to before sending
to the IESG. 1-week call should be sufficient.
BL: Sounds good; I'll start it after this call.

Packed?

Pretty well-ripened now in WGLC completed status.
Needs processing of recent discussions

Table reference pretty stable.
(Issue: loop detection.)

Basic "batteries included" table setup should be moved to
single-parameter (setting up both shared and argument table to the same
value).

Submit before cutoff.

CB: We are waiting for implementations and haven't got many feedback.
Many might be hesitant because of pending changes.
CB: A few things to do first, e.g., further simplifying the table setup.
That makes it easier to write packers. Text hopefully ready before this
weekend, to be looked at before the cut-off.
CB: This may really need a new WG Last Call.
BL: Certainly. Let me know when it's ready.
CB: I expect one new version first, then that should be ready for WG
Last Call.

CDDL ...

BL: Any discussion needed today?
CB: We now have WG documents on this. The small corrections (errata plus
clarifications) can happen soon. Control and module are independent, can
benefit from some discussion at IETF117.
BL: any discussion now?
CB: Interested parties not here. Some implemented features don't need
standardization (e.g., flattening), but are in cddlc-tool. Feedback
welcome anyway.
CA via chat: I'm here but the one I'm rooting for is back-annotation :-D

IETF 117 agenda

Requested session is 60 minutes.

• Introduction, agenda. Chairs, 3'.

• Brief overview of documents not discussed today. Chairs, 2'.

  • time-tag: Done (??).
  • packed: Please suggest alternative table setups. (We'll
    come back to a use case later in dns-cbor).

• WG documents

  • edn-literals: Discuss ABNF proposal in -01 (??). CB, 5'.
  • CDDL: CB, 20'.
    • update-8610-grammar
    • cddl-more-control
    • cddl-modules

• Individual documents

  • dns-cbor. ML, 10'.
  • dCBOR: Deterministically encoding numerically typeless CBOR, and
    standardizing a deterministic encoding. CAl, 15'.
    (possibly advertising larger side meeting)

• Flex time. 5'.

CA's notes preparing the agenda:

• (??) indicates items where it is questionable whether they will
  really be in that state at the time of the meeting.
• On packed we had the thread of "items standing in for others" on the
  conceptual side. Is that closed? [for minutes on this, see below]

BL: Keeping time-tag for now; most likely, it will be sent to the IESG
soon (see above) and we will not need it in the final agenda.
BL: Does this agend need tweaking?
CB: LGTM. On edn-literals, it's on my list right after packed. Grammar
needs translation, but I'm writing tools for that.

AOB

BL: More input?

CA: About -packed, we discussed CBOR tags standing for concrete values.
Was the discussion concluded? Still open points to address? Can't find
the original discussion.
CB: We should write this up.
CB: we've had the discussion, and examined space between purely
syntactic validity (7049) and totally semantic validity (hard to get
interoperable); decision was to go a little bit beyond syntactic, but
with the awareness that too much of it is detrimental for
interoperability. Need to write it up, more like memo-to-future-selves,
that we want to be careful following along the lines of packed.
Informational probably right.
CA: Sounds good. Would that be informationally referenced by -packed?
CB: We can't finish this before -packed. packed charges ahead, and is
example in the new doc. This is more for infrastructural documens, less
for allowing any tag to say it's now also a number.
BL: I like the idea of a memo to our future selves.

Note taking: Marco Tiloca