– DRAFT –
DXWG CNEG Telecon

https://www.w3.org/2017/dxwg/wiki/Meetings:Telecon2023.09.20

<pchampin> will have to leave in 45m

Admin

Approve last meeting's minutes

<roba> https://www.w3.org/2023/09/06-dxwg-minutes.html

Approve last meeting's minutes

<YoucTagh> +1

<roba> +1

+1

<pchampin> +0

RESOLUTION: Minutes approved

Open actions from last meeting

https://www.w3.org/2023/09/06-dxwg-minutes#ActionSummary

roba: first action (access for Nick) has been resolved

<roba> w3c/dx-connegp#49

roba: Nick has made some pull requests

roba: https://www.w3.org/2023/09/06-dxwg-minutes#a02 (reactive, proactive cneg)

YoucTagh: Has updated issue

w3c/dx-connegp#39

roba: behaviour is acceptable. If you ask for a profile
… but the media type is noc available, you should
… answer 406 (never return a media type the client
… didn't ask for)

YoucTagh: media type is one of the dimensions in http
… if the default representation does not have the
… requested media type, we can specify that in the
… link header, so the client would know that anyway.
… If the server doesn't want to send a default representation
… it would return 406

YoucTagh: In reacztive CNEG, the I-D specifies using HEAD to query for the
… supported profiles. That should be 300.

ProfileNegotiation/I-D-Profile-Negotiation#28

roba: If a client uses HEAD without specifying a profile,
… the server should answer with the default representation
… including all headers

<YoucTagh> https://www.rfc-editor.org/rfc/rfc9110.html#section-9.3.2

YoucTagh: according to the spec, the response headers for GET
… and HEAD need not be the same.

pchampin: http semantics says that the answer to HEAD
… request SHOULD be the same as for a GET (i. e. the headers)
… that should go for the status code, too

roba: The question is if the server returns 200 or 406 for
… a request for a header/profile combination it cannot
… support
… we need to make decision on that

pchampin: it would be important to discuss this with the IETF group

ACTION: roba to word-smith a proposal for 200 vs 406

IEtF

pchampin: At TPAC there was a meeting with IETF people,
… concerns that the ProfNeg spec is more IETF tha W3C
… there are companion documents in IETF space
… their advice would be send this to the http WG at IETF

roba: sees the benefit of having the protocl side done in one place
… W3C work includes an ontology with semantics
… and the QSA side doesn't fit with IETF, either
… the W3C work should refer to IETF docs, not duplicate them

<roba> LarsG: have been working on this for 5 years with Rueben Verbourgh and Herbert van Soempel

<roba> ... were rejected by the IETF protocol working group - would be happy to recontact

roba: main contacts Martin Thompson and Mark Nottingham from http WG

pchampin: large part of the spec is in W3C space, but the
… protocol part should be IETF

LarsG: The proposed I-D is at https://profilenegotiation.github.io/I-D-Profile-Negotiation/I-D-Profile-Negotiation.html

ACTION: pchampin to contact Martin Tompson and Mark Nottingham in order to get a liaison

rest of actions...

roba: Action to prepare PR for #45 (https://www.w3.org/2023/09/06-dxwg-minutes#a04) is still open

YoucTagh: There is a branch for that
… questions about some examples

roba: A 4.2 is just an example, shouldn't be in the same PR

YoucTagh: format(s) should be consistent with link hint spec (https://datatracker.ietf.org/doc/html/draft-nottingham-link-hint-03#name-formats)
… has updated references (how do we reference RFCs?)
… will try just using RFC...

Open Pull Requests

roba: Nick has done a change to use JSON-LD at some places, too
… we should be able to merge that
… <assigning YoucTagh as reviewer>
… merging that directly
… PR #48
… here Lars needs to review
… and Nick make to changes requested by YoucTagh

roba: Won't attend next meeting