WG documents status and issues
draft-ietf-cbor-cddl-control-05: In Last Call (ends 2021-09-21)
CB: the AD review came today; one comment is about use of quotes.
BL: RFC editor can fix that.
CB: Can’t, the RFC is defined by it’s XML and can only handle the typewrite font.
FP: Maybe worth to get the rendering working, but not a big issue.
CB: The discussion is ongoing, some say double quotes should always be removed. But that causes issues also. This is a known bug, we need to wait for RFC editor to fix it.
draft-ietf-cbor-network-addresses-07: AD review completed, 1 issue to talk about
FP: Document is in last call now, will end in 2 weeks. Carsten, if you are quick to reply to reviews I could schedule this for the next telechat.
CB: Will be on vacation and I might not be quick.
FP: Then I will not put it on the agenda for the text tele chat. I just sent out CBOR network addresses review, have not asked for last call. Wanted to discuss a last point. Why do we have tags 260 and 261, why not obsolete them? Michael said he did not care on mailing list. Carsten said he wanted to avoid restructuring. We can add a note to say “obsoleted in favor of 52 and 54”, the document already states the intention is to replace 260 & 261.
CB: Agree with approach, maybe “obsolete” is not the best thing. Some documents use those tags. It’s more about not recommending them for new things.
FP: Yes, people can still use them, but obsolete gives stronger impression
BL: “Deprecated” would work here, still usable but discouraged.
FP: I am happy with “deprecated”. Also, I prefer the “deprecated” note to be by the tag number.
MCR: So deprecating the old tags in favour of the new ones?
FP: Yes, see chat. "IANA is requested to add the note “OBSOLETED in favor of 52 and 54” to registrations 260 and 261””
MCR: Ok.
CA: (In chat) See datatracker issuse here
FP: Then I will request last call. We can have it on the telechat in 2 weeks, let me know.
BL: I will be around to answer questions.
CB: Shall we resubmit it before you request last call?
FP: I’ll request Last Call, you can submit when you want in the next day or two.
MCR: I have some edits pending, will submit immediately.
Extended Diagnostic Notation (CB)
CB: We had diagnostic notation for a long time, have changed it once. By adding extended diagnostic notation. We may want to think about additions to support specific applications, like CoRAL. It is a hypermedia format and contains links, forms, and metadata. A simple CBOR format coming from discussions in CoRE.
CB: Currently there are 2 representation formats in CoRAL: one that is CBOR-based and a text-based format (good for whiteboards/examples/other documents). We want to avoid designing things that look good in the text-based format but does not work well in the CBOR-format. People are wondering if there should be a JSON-version, that is up for discussion.
CB: This brought up the question, can we use the CBOR diagnostic notation somehow? That is difficult.
CB: (slide 4) Most of a CoRAL document is CRIs. The CBOR diagnostic notation is compact as encoding but it ends up inconvenient to use in examples and the like.
CB: (slide 5) A possible approach is about thinking on a per-application basis and use prefixes - like macros, but without defining a macro language. Applications would define a specific keyword (e.g. “cri”) preceding the specific content (e.g. a URI produced out of a CRI) in between apostrophes.
MCR: This is something used on the input side, and not emitted?
CB: Yes, any tool aware its looking at a document with CRI, and can do the transformation, would benefit.
MCR: So we need a registry? I have no problem. But it may confuse people, if not clear that “cri” does not go on the wire, this is just syntax for diagnostic notation.
MCR: Is the lack of space between “cri” and the apostrophe intentional?
CB: Yes.
MCR: Sounds good to me.
CB: Any further comments? I got mail comments that are positive. Let’s make more examples, but this can be a good direction.
CB: (slide 7) More considerations about CoRAL. It has CURIEs in CRI. It gets difficult with prefixes, especially thinking of CBOR-packed. Maybe it’s worth extending the CURIEs.
CB: I wrote a previous draft on CBOR templates which can solve some of this. We may need to resurrect those.
MT: Can you write a link the the draft in chat?
CB: Yes. (in chat) draft-bormann-lpwan-cbor-template
CBOR use in other SDOs
AOB
BL: Any other business? (none) Thanks everyone
Notes (please volunteer to help take notes): (Marco Tiloca and Rikard Höglund)
[CB]: Carsten Bormann
[BL]: Barry Leiba
[FP]: Francesca Palombini
[MCR]: Michael Richardson
[CA]: Christian Amsüss
[MT]: Marco Tilcoa