Note well reminders
Note taker(s): Rich
ADs
Thanks for your work Francesca, welcome Mike.
Brief document updates
-- api-catalog in RPC queue
-- deprecation-header in RPC queue (one of our first docs!)
link-hint
Please review and provide feedback
rest-api-mediatypes
It's really now registering OpenAPI media types.
Two open issues, hope for WGLC before next IETF meeting.
ratelimit-headers
Much feedback last time, very appreciated.
More explicit and obvious use of structured fields.
New problem types introduced to explain why client was rate-limited. Would like feedback on them.
Darrel has an implementation on GitHub to check this is implementable; URL in the slides.
Mike: maybe add a response field that server can send to indicate the true cost
WG FOLKS Please think about adding such as response field, if we get feedback we can move to WGLC by IETF 123.
A major rationale: is hard to make tooling for JSON Schema, Avro is also better suited for data-binding.
You can think of this as a refactoring of JSON Schema; some features removed, more data types (that can have attributes). They also have serialization rules to/from JSON string.
Restrictions on $defs/$ref for reusable types.
Documents, etc., at https://github.com/clemensv/json-cs/tree/main/jsc (also posted to list)
Andy Newton as individual: I like this, hope it moves forward
AN as ART Ad: Need to be careful about the name "Json Schema"; might be question of venue. Please also send to art@ietf.org
Clemens: Not wedded to the name, but do want to leverage the existing JSON schema drafts for mindshare.
Darrel as individual: agree JSON Schema is hard to do code-generation and it seems to be stuck right now.
Andy: I think the confusion about JSON is reason enough to look at this.
Chairs and ADs to determine a proposed way forward.
Darrel talked about politics of personal views of various JSON authors, IETF relationship, etc.
httpapi-privacy
Mike Bishop (as co-author): need to do some work, not ready for WGLC.
digest-fields-problem-types
Chairs to request SECDIR early review
Question about adding Message Signature (RFC 9421) problem types to this doc?
Justin (author of 9421): I think it makes sense to keep them separate; signatures are pretty complex and have oracle considerations, etc. Consensus to keep them separate
If anyone wants to work on either one, contact the Chairs
patch-byterange
authentication-link