Minutes interim-2023-core-12: Wed 14:00
minutes-interim-2023-core-12-202308301400-00

CB presenting

Presented slides (from p7):
https://datatracker.ietf.org/meeting/interim-2023-core-12/materials/slides-interim-2023-core-12-sessa-carstens-slides-coreconf-href-corrclar-00.pdf

CB (p7): This document has been revived and its GitHub repository is
active. The content is intermixed with implementation issues.
CB: One specific issue is Californium 2088: Eclipse Californium
used to reuse the CoAP token value throughout the same block-wise
transfer; Zephyr relied on it on the server side. Californium reverted
and provided an option to accommodate Zephyr.
CB: Two aspects -- people are running stuff outside (so some emergency
fix is needed), and people perceiving this as "Californium does things
right again now" (dangerous situation).
CA on chat: facepalms.
CB (p9): The more generic issue is that there are more issues around
FETCH and block-wise. "Taking the 3rd block of a POST on its own" rarely
makes sense (i.e., POST is rarely stateless), but for FETCH it is
different, as it is very likely to be wanted as stateless.
CB: We want to do this right, but also to through a minimal change. I'm
not suggesting that we design this all in this meeting; it is an example
of a corr-clar issue that requires our timely attention.

CB (p10) on process: MT wrote how the process for this document can be
(summarized here).
CA: Do we have to delay WG adoption until we categorize the issues?
Maybe it's better to adopt soon, once we're happy with the process that
then we follow. The document can still change through the process once
it is set up (as "Confirm each issue" is also sufficient to do a change
on the WG document). At least we can point people to a work-in-progress
happening as a WG document. This also serves to discourage people from
rushing implementing things that have to be revised anyway. New sequence
would be 4.1, 5, 1, ..., 3, 4b, 6. That is:

1. Reshape the -corr-clar document in order to include the revised
   process
2. Consider WG adoption.
3. Go through available material (issues, FAQ) and revise/create
   Github issues as needed
4. Categorize the Github issues into topics (dedicated team)
5. Confirm each issue with the CoRE WG, get feedback from protocol
   designers/implementors:
   — include and cover in -corr-clar, as is or revised; or
   — simply omit in -corr-clar; or
   — leave for a possible -bis document (e.g., for some specific
   points
6. Reshape the -corr-clar document in order to reflect a sequence of
   pairs (Diagnosis, Therapy)
7. WG document work can focus on improving the therapy parts

MT: The process can also be transferred into the main corr-clar repo for
finalization (corrclar-process is still private I think).
CB: I can dump the text over the draft too, let's do that quickly.

CA: Nothing fundamentally wrong with Californium I think, the behavior
is correct. It's more about what happened on the other side.
CB: Don't have a link, we could ask.

AOB

Francesca and Christian are happy with their kids.

MT: We will cancel the next interim meeting in 2 weeks, due to Chairs
unavailability. The next one we have is on September 27; most likely,
the topics of today will come back.