CBOR working group conference call, 2022-03-09

Agenda and notes:

• WG documents status and issues
  • file-magic shepherd write-up

CA: Just finished the latest shepherd write-up. Anything to add on the document?
CB: “CBOR objects” can be confusing for JSON people. Better “CBOR data items”.
CA: Plan to change already or later in IESG review?
CB: Might be relevant also for the shepherd write-up.
CA: “object” is used in the draft abstract, though it’s the only one.
CB: We may go into AD review anyway and fix it later.
CA: Ok, I’ll point it out in the write-up.
MCR: Always do it last…
CB: I tell students they have to write it twice – once at start and rewrite at end.
CA: Fine then, we’ll move it on after the little update to the write-up.
CB: Ben will probably not review this due to timing.

FP: Not sure I can promise it happens before IETF 113. Once finished the AD review, I expect a longer IETF Last Call if it starts before or during IETF 113.

• Packed: CB: is it being implemented? I’m a bit slow there, but it would be good to get feedback.

CA: Plan to do something for CoRAL, still have to start. SUIT people may also want to start even if that early.
CB: Packed is part of CBOR diagnostic. Maybe to be done in cbor.me …

CA: Also side topic about diagnostic notation on packing. Considered it for CoRAL, readability is lacking.
CA: Ideas on how to make more readable?
CB: Have special notation for references (no big difference for simple, but for the tags). Diagnostic tool might include more information, but not designed yet (but essentially you’d see unpacked form but with breadcrumbs in place).

(CA using the minutes pad as a white board)

[6, ["http", "example.com", "dict", "example"]]    / uncompressed /
[6, cri'http://example.com/dict/example']          / uncompressed but with EDN /
[6, p'example:example']                            / wishful thinking /
[6, p'50' / ["http", "example.com", ...] / ]       / what a decompressor can do easily /
[6, p'50: example:example']                        / what a decompressor can do with additional information /
[6, p'50' / example:example /]                     / (either this or the above; the former is nice for slashes ) /

Dictionary could contain additional information (maybe in different representation, one for unpacking and one for annotation).

CB: could concoct names such as hexdiex

• CBOR use in IETF and other SDOs

• AOB

MCR: Envisioned activities in the WG during next period?
CA: cddl-csv, not sure if CB serious.
CB: No joke, just showing how something could be used in an area where it isn’t used yet. We do have lack of formal descriptions.
CA: More a template for future things when relevant?
CB: Prompted by a spec about CSV so it’s concrete.
MCR: So not talking about CDDL describing CSV format, but column contents?
CB: Yes, matrix-based generic data model, then turned into CSV as representational model. CDDL describes the data model.
MCR: Open Data community could use something like this. Often you get data and a Word document describing its structure. Not sure if relevant in the CBOR WG, not objecting either.

CA: Also topic on CDDL 2. No draft yet.
CB: The freezer document has some parts. But in particular the combination of CDDL docs (module mechanism) is still wobbly.

CA: So one around, one coming up and one brewing. Also notable-tags, but no hurry. CB: Language/direction? CA: based on feedback, some implementations don’t check the length of the tag.

See you all in hybrid Vienna!

Notes: Marco Tiloca

[CB]: Carsten Bormann
[BL]: Barry Leiba
[FP]: Francesca Palombini
[IMD]: Ira McDonald
[MCR]: Michael Richardson
[CA]: Christian Amsüss
[MT]: Marco Tiloca
[PP]: Philip Prindeville
[ST]: Sean Turner