https://logs.xmpp.org/council/2021-02-17?p=h#2021-02-17-a7279024de1951f7

1) Roll Call
Present: Daniel, Georg, Zash, Jonas, Dave

2) Agenda Bashing
There are no complaints.

3) Editor's Update
Nothing to report.

4) Items for voting
None.

5) Pending Votes
Dave votes on PR #1032: +1 (doing otherwise will break several implementations, 
including anything based on Smack; think this is the original intent [of 
XEP-0004] and early implementations appear to follow it.)
Georg hasn't yet caught up on the really long thread.
Zash thinks the thread appears to have a rough consensus, and votes +1.
Jonas says that the behaviour of aioxmpp on the client side matches what Drew 
described [1] - Dave doesn't think reordering matters much when consuming, but 
does when producing, and seems entirely wrong when forwarding - Jonas agrees 
that forwarding shouldn't reorder things. Kev doesn't think breaking Smack is a 
motivation to do things, but that creating headers after their content is 
probably broken, and hopes any forwarding entity which changes the order is 
only used within a closed deployment.
Daniel doesn't see how the existing text intended for 'reported' to come first, 
nor why it's desirable - Jonas agrees - Kev thinks it seems like the Right 
Thing, but even though it doesn't appear the original text necessarily mandated 
it, it's clear that was the intention (exemplified in §3.4 "…describing the 
data to follow") - Dave agrees that it's not very clear, but the intent is 
there.
Georg verifies that the decision is to break implementations which don't follow 
this - Dave suggests the alternative would be to break implementations which do 
- Georg says they don't have an explicit right to rely on it, rather a vague 
implication. That said, Georg is okay with changing the vague implication to a 
formal requirement.
Jonas is quite reluctant about the 'MUST' because this is Final (less issue if 
it were Draft or Experimental), and thinks this needs a good rationale and text 
content explaining the change - Georg suggests adding a note about earlier 
versions of the XEP not strictly mandating order. Dave would be okay with a 
'SHOULD', though aioxmpp and ejabberd would still need to be fixed.
Georg votes: -1 (as-is, with the MUST; +1 for a MUST with explanatory textbox, 
or just a SHOULD.)
Daniel could be on board with a SHOULD, since this shouldn't be a big deal in 
practice (ignoring those who need to fix their code) - Dave asks why a SHOULD 
is preferred here - Daniel feels it's less bad when changing a Final XEP.
Georg proposes some text to go in a box. Dave thinks the standards list should 
be consulted - Jonas considers the addition of a box as editorial - Dave says 
that means Council don't get to vote, not that the community is bypassed. Kev 
thinks that adding a box is editorial, but adding text about implementations 
and encouraging behaviour certainly isn't (even if that happens to be inside a 
box.) Jonas will put it to the list, but doesn't see the necessity given that 
this PR wouldn't have been discussed there otherwise - Dave thinks that's a 
problem in itself, and prepares a can of worms for AOB. Zash can't handle any 
worms today.
Jonas pushes a patch onto the PR, and plans to vote on the updated version next 
week after possible list feedback [2].
Jonas votes once again: -1 (the new diff looks better considering the current 
state of the world.)

6) Date of Next
2021-02-24 1600 UTC

7) AOB
Dave notes that the Data Forms discussion on-list was very useful, but only 
started once voting had already begun, which feels awkward; it would be better 
to discuss things first, and only add them to the agenda once they have been 
discussed, so Council can just vote - Jonas agrees.
Jonas thinks it would be great if some CI magic could be used to automate 
announcements of PRs tagged with 'Needs List Discussion' and 'Needs Council'.
Dave suggests advising that things don't get added to the agenda until after 
somebody starts the discussion on-list - Jonas doesn't see any significant 
issue with that; Zash approves of subverting the GitHub flow. Jonas continues 
to fantasize about automation, but likes Dave's proposal as a first iteration. 
Dave admits that it pushes extra work onto the submitter, but encourages them 
to add some rationale and convince people. Daniel is somewhat fine with the 
current flow (Council gets to decide on uncontroversial things and refers 
others back to the list), but is also fine with this new suggestion - Dave 
doesn't think Council deciding what is actually controversial is optimal, given 
potential bias and the lack of an appeals process.
Jonas prepares an email explaining the proposed change [3].

8) Close
Thanks Jonas, Tedd, All, Sundry.
Dave apologises for his contribution to the meeting overrun - Jonas thanks Dave 
for the neat idea.


[1] https://mail.jabber.org/pipermail/standards/2021-February/038175.html
[2] https://mail.jabber.org/pipermail/standards/2021-February/038183.html
[3] https://mail.jabber.org/pipermail/standards/2021-February/038185.html

_______________________________________________
Standards mailing list
Info: https://mail.jabber.org/mailman/listinfo/standards
Unsubscribe: [email protected]
_______________________________________________

Reply via email to