interim-2021-cbor-04
2021-02-24 15:00 - 16:00 UTC
Chairs: Francesca Palombini, Christian Amsüss

Attendees:

• Francesca Palombini, Ericsson
• Christian Amsüss
• Marco Tiloca, RISE
• Peter Yee, AKAYLA
• Carsten Bormann, TZI
• Henk Birkholz
• Michael Richardson

Minutes

• Intro

ordered maps and other map types

CA: Discussion on the list; the WG is interested in having this. Anyone who wants to really work on something that can become a WG document?

CB: I have some early text on maps; it feels natural to build on it, we need to find contributors and add them.

FP: There's some text from Kio and Joe on github: https://github.com/Sekenre/cbor-ordered-map-spec/blob/master/CBOR_Ordered_Map.md

CB: There was a Pull Request from Joe to Kio's text, so they have merged together. I have formatted as a draft to start from.

FP: Better to have a draft for registration.

CB: We can have it well before IETF 110, and there decide how to do this.

CB: One things is being able to tag a map; the other is to work with platforms using two different map types (which may require different tag properties).

CA: Different languages and different defaults make things tricky already anyway.

CB: How many of these alternatives do we want to cover?

CA: Are there requests from anywhere in constrained areas for anything different than from IETF?

CB: Alex has talked about that.

CA: Good to ask directly what's the constraind use case for this, on top of this being needed. (→ CA)

→ CB to put in notable tags.

cddl-control

CA: Short summary of the document?

CB: We have implementations for .cat and .plus and IETF document using .feature. Do we wait for ABNF tag requiring more work?

CA: what's the timeline for the IETF document using this?

CB: We want to complete it this year.

CA: ABNF would be candidate for postponing.

→ CB to split the draft

CA: Yes, though not urgent; it can happen sometimes around April. That gives time to the ABNF use case.

CB: It's more about a meta use-case and someone implenting it to understand all the pitfalls. Preference for having implementations before standardizing. Hard to predict when they're ready.

CA: Are there any use cases urgent enough that they need implementation for it?

CB: The alternative is to convert ABNF to REGEX. ABNF is very used anyway.

CB: The documents has implementation status as recent changes.

FP: Overall, waiting for implementations and to be discussed again at future interim meetings.

CB: Or we split the document now and we move forward those we need more urgently.

CB: September 2021 is the goal for requesting publication of the document using this.

CA: We should keep a similar timeline. Any more complication expected?

CB: No complications expected.

CA: Let's wait for the next interim to come for more implementation feedback.

cbor-file-magic

[wg adoption ending Tue Mar 2]

CA: WG adoption on this looks good. Some question on the list for clarifications, and they got answer.

MCR: We need to agree on using a CBOR sequence and on the number to use, which seems free and is never used in Unicode.

CB: 55800.

MCR: Many in that range are fine. Should the document describe this better?

CA: It's good practice to explain how we come up with the exact numbers.
CB: The text is already saying how we thought through.

network addresses?

[wg adoption ending Tue Mar 2]

CA: WG adoption on this looks good.

CA: Are there any application for an array of prefixes? It can be a sub-thread on the list.

MCR: Some applications may have prefix overlap and want to handle them with a better encoding, avoiding repetitions.

CB: It can become overoptimizing; it's supposed to remain simple.

MCR: Compressed CBOR may already do a good job and be enough.

draft-bormann-cbor-time-tag-04

CB: Getting more input on how to put information in this data structure. Need to address also points from other RFCs, e.g. RFC 3161. Need to check this is also addressing problems from RATS. It's likely to be used relatively soon, so better to put it on the front.

MR: What's RATS's problem?

CB: Need more information about quality of time tag, like uncertainty.

→ everyone: read; WGA questions will be asked.

CB: feels like approaching an adoption call. Still to discuss about registries.

CA: which can be discussed and solved after adoption.

AOB

CB: Items at IETF 110?

CA: Still preparing those. Surely asking about adoption of time-tag.

FP: We'll draft and send out an agenda asking for input.

adjourned 15:38Z