XMPP Council - 2018-11-21

Still a bunch of votes pending, deadline tomorrow EOB.

or rather EOM?

EOM?

We agreed 'end of Council', I believe.

End Of Meeting, I would assume.

I'm ready to vote on all of mine in the meeting, FWIW.

End Of Meeting, indeed.

Kev: 👍

Right.

1) Is there anybody out there?

Kev, SamWhited, daniel ?

hi

I is definitely here.

OK.

2) Agenda

I am working on the assumption that the only items for a vote are anything outstanding from last week.

(And that nobody is mad enough to try adding new stuff).

So with that in mind:

3) Outstanding votes

I think it's everything from last week, yes.

i.e. just the same agenda again.

I shortly considered adding items to vote on, but abstained.

I believe I'm up to date, as is Ge0rG.

I've done all my on-list votes.

And SamWhited is too.

Alright. We did agree last week to have discussions on them all this week, rather than just treat them as purely on-list, but w/e.

I anticipated objections to my requirement to make stanza @id = origin-id @id

We certainly can discuss.

3) Advance XEP-0357 (Push Notifications) to DRAFT - https://xmpp.org/extensions/xep-0357.html -1 4) Advance XEP-0359 (Unique and Stable Stanza IDs) to DRAFT - https://xmpp.org/extensions/xep-0359.html -1 5) PR #692 - XEP-0060: correct "entity" to "<subscription/>" - https://github.com/xsf/xeps/pull/692 +1 6) PR #693 - XEP-0060: Remove unused 'node' attribute on pubsub#event item - https://github.com/xsf/xeps/pull/693 +1 7) PR #715 - XEP-0045: Add missing disco#info feature to example 4, 9, 78 and 218 - https://github.com/xsf/xeps/pull/715 +1 8) PR #716 - XEP-0030: Clarify 'disco#info' feature in 'disco#info' responses - https://github.com/xsf/xeps/pull/716 -1

Ge0rG, I suspect that's desirable, but also potentially difficult if the library sets the stanza's @id at the point of send.

Kev: that lacks rejection rationale

> I anticipated objections to my requirement to make stanza @id = origin-id @id i think i suggested that a few times but the author always objected

dwd: ideally, the library should be the one adding the origin-id then

daniel: yes, I suggested that as well, and got the same reaction.

There was no rationale, so I'm using my Council hat now.

I can't see a reason not to make stanza id = origin id.

Ge0rG, Well, yes. But often, libraries bake in support after various apps have used them first.

dwd, then it’ll take a while for libraries to adapt.

dwd: I'm sure this is a problem easier to solve than multiple mismatching ids on a message

Ge0rG, That all said, I won't object very strongly whichever way it goes.

wasn’t the primary reason that smack can’t do it

besides, we need to fix Receipts and LMC with origin-id in place.

Ge0rG, Could a server just add in the origin-id, then?

daniel: even the old version of smack I run makes it possible to extract and change stanza IDs

dwd: that sounds like a reasonable proposal to solve some problem. I'm just not sure which one.

dwd: also not sure whether the rules of 0359 allow injection of origin-id

Ge0rG, If a client wishing to add in origin-id has to add the same id twice, then a server could add in that as well.

Ge0rG, Probably not, as written. My question was whether it made sense to do so.

As already stated, I consider origin-id to be a hack to work around bad servers, so I'd rather burn it.

I can see some benefit in <stanza-id/> for MAM purposes, because you have an independent entity defining _that_ id.

Originally, <origin-id> was a payload you were supposed to stuff into messages you send into a MUC or into transports, in the hope to see the element reflected

But then we got MUC changed to discourage that, and transports probably can't retain XML payloads anyway.

So in my eyes, <origin-id/> could go away.

That I could go along with.

(and if you have a transport that can maintain XML payloads internally, you surely can send a reflected message to the sending client with the same @id attribute)

(for example, the transport implementation could inject its own version of the <origin-id/> thing into the "legacy" protocol and convert it back to the @id when the reflection comes back)

Ge0rG, Yes, I was just thinking that. So does that mean there's no point to origin-id excepting some old servers?

the last reason to have <origin-id/> is that it's an indication that the client is using sufficiently-random @id's, and I think that can be signalled in an easier fashion, if needed at all

dwd: that's my understanding, yes.

I am amazed. will we really solve this battle in the last sitting of council? probably not because SW won’t be able to react in time...

Ge0rG, I'm not sure that *is* an indication of more than some vague intent.

dwd: excepting some old MUC implementations, to be precise

I think this is a mailing list or after-Council discussion really. We've reached -1 point, I think.

Kev, I think you're right.

I think.

aw pity

how about "apply these changes and then +1"?

where "these changes" points to a PR?

jonas’: not gonna happen in the next 15 mins

Right.

how about "take authorship away and do it right"?

no, I don't volunteer.

I do.

Anything else ... XEP-0357? Kev, you'll owe rejection reasons, but I guess those are in your mail to list?

We don't need to take authorship away to do it right, incidentally.

I was about to write "challenge accepted!" and do a PR right away, but then I figured that the security considerations of stripping existing stanza-ids probably really can’t be figured out in 15 mins

jonas’: my -1 was already conditioned on "make @id = origin-id", but now it looks like it's just -1

I think my mails today about both 357 and 359 are feedback enough on those.

Especially as I'm the person that needs to make the 357 changes.

True.

Congrats on rejecting your own XEP.

Speaking of taking ownership away.

So, PR #716?

I was in favour of dropping the need to signal disco#info over disco#info.

dwd: is #715 fully voted on already?

dwd: re #716, last Council it was brought up that caps hash calculation would be inconsistent then

Ge0rG, Oh, no, daniel hasn't I don't think.

also it's a MUST from a Final XEP

I think the normative change here is wrong at this stage. I think noting that it'll be elided from many examples is reasonable, and probably even that some implementations may elide it (although what the implications of that are isn't entirely clear. I think it doesn't affect 115, though).

Ge0rG, I think a client that includes it in disco#info includes it in XEP-0115, surely?

I'm not quite sure what Florian was trying to make on the list responding to my -1.

at the very least, doing it inconsistently will defeat some 115/390 caches

dwd: I think that having it being defined implicitly leads to corner cases. Also Final XEP.

We /are/ allowed to modify Final XEPs.

Ge0rG, I'd be more convinced about the Final XEP argument if this affected interop, or indeed was actually followed.

"Every effort" not to, though.

dwd: I think the argument is that an implicit feature there might make future 115 bugs.

dwd, I think most implementations do follow it

Ge0rG, The fact that many implementations *don't* include disco#info, and everything still works, really does suggest it's wrong.

we do have the luck to have a huge stash of disco#info replies from both servers and (a bit outdated) clients if you want to make a survey

My inclination would be to leave the normative language, but note a) that examples don't include it, and b) some implementations elide it.

I can modify the muclumubs ( https://search.jabber.network ) bot to make a stat on that

+1 to what Kev suggested.

I can see I'm on the losing side here, and can go along with Kev's suggestion.

Otherwise I'm not strictly opposed to making it optional if we're very sure that the language we introduce couldn't cause caps weirdness with it being implicitly added.

Kev: in my eyes, it's not about the language but about implementations that add it, then do caps, vs. implementations that just do caps.

Right - anyone want to discuss anything else?

I'd like to thank everyone for their work this term :)

Ge0rG, I think if implementations were inconsistent then caps would be failing for them and people would have noticed, FWIW.

I think dwd is right.

Kev, Yes, I was going to do this:

4) Thanks, All.

I'm worried that we might inadvertently make matters worse by adding language about optionalness. If we're sure we won't, I'm not strictly opposed.

Thanks to everyone for your efforts - not only Council, but jonas’s work on Editor stuff, too.

Indeed.

Thanks to you too, Dave.

thanks to council :-)

this was a fun ride

And to the Chair.

let’s see what the numbers tomorrow bring for next year

Also, this year, we've had a number of useful contributions from the floor, so thanks to anyone who's chipped in.

I've heard we didn't pass that many XEPs to Draft, which is maybe a bit sad.

exactly one.

and it was the compliance suites

three months late or something :)

but whatever

jonas’: I've heard you are already preparing Compliance Suite 2019.

yeah, that cp sure takes a while...

It's not Council's job to make XEPs ready for advancement, mind, so I don't feel too bad about that.

Kev, I didn’t mean to make anyone feel bad anyways

just for the record :)

👍

We have five minutes left. Any technical topics?

Happy for the 359 discussion to continue now I don't have to pay as much attention :)

Kev: I'm not sure what's left to discuss there. I'd rather discuss Push.

what are my chances that someone updates the spreadsheet of doom right now so that I don’t have to go over the emails?

359-2 then :)

Besides, should we propose a meeting time for the new Council?

new council will have to figure that out tomorrow night I think

Well, not tomorrow night, but yes.

although we could agree on a time; in any case, we know the maximum set of people who’ll be on it anyways

oops, sorry, had guests and wasn't paying attention to the time

it's probably overly optimistic to assume that all five candidates will actually receive a majority of votes.

Ge0rG, sarcasm?

SamWhited: are you missing any votes?

Ge0rG, Thankfully it needs people to vote against to actually lose.

jonas’: no, just saing it's rather inappropriate to proclaim "I'll probably get reelected anyway"

in any case, the wednesday 16:00Z slot wfm

dwd: Simple needs people to not vote for.

for now.

No, I emailed mine and should be all caught up now

SamWhited, Ooops, soryr - didn't think, it's Turkey Day tomorrow, isn't it.

dwd: yup, don't worry, I forgot too and I live here…

Kev, Well, yes, but if they do not vote at all, that's not counted.

I'm not sure that's true.

Someone can submit their vote, for no candidates.

Let's postpone this until after the election.

If a majority of people did that, we'd have no Council.

Kev, True. But that counts as voting.

> Third, the individuals elected shall be those receiving the highest percentage of votes cast, up to the limit set by the Members and with the proviso that no individual receiving less than a majority of votes cast shall be elected.

Does memberbot allow that?

Ge0rG, yes

Ge0rG, Yes.

you can abstain for all five

dwd: Yes,t hat's what I said. It just needs people to not vote 'for'.

Anyway. I think we may be done.

thanks again y’all

Kev, Yeah, I just think it's more likely they won't vote at all.

So, for the final time this Council:

Our voting rules are actually broken, because all we need is for enough (non-joke) candidates to apply and we can't form a Council.

5) Ite, Meeting Est

Diolch, pawb.

Thanks all

Kev, as long as this doesn’t happen to board (or whoever has power to fix bylaws), we’re good!

Kev: so we need to define a second ballot process with at most 10 candidates?

Kev, I suspect that it's memberbot breaking. The way the bylaws are written, one could argue that each council candidate is voted for/against individually.

jonas’: I think that for Board we're probably already covered.

dwd: True.

dwd, I agree

I was a tad surprised to see the memberbot process after reading the bylaws more carefully this time

Meeting's tomorrow night at 1900Z, right?

yes

also, I don’t think that we need to have an elected board to fix the bylaws... we could have an all-member-vote about the change :)

ISTR (without checking) that in the case that we somehow had no Board elected, Peter could pick his own.

jonas’, We did that one time in a meeting when we realised we'd catastrophically f**ked up. Ah, it was fun.

jonas’, Strangely, XSF Meetings were very well attended for the next few.

:D

storytime?

There was a thing. Hopefully there isn't a thing again.

Kev: I can't do any such thing because I'm no longer the executive director.

Ah, ok. I thought you still were until they found a new one :)

Seems to me we're doing fine with just the council. And as noted the members can always call a special meeting to fix things.

er, board

I prefered your first version :)

heh

Kev: accidentally candidated for the wrong panel?

I hope not, otherwise Council's short.

even if the Council were short, it could presumably nominate another member after the election, no?

Yeah.

I think we've entered silly territory now.

"now".

I thought the Council is allowed to be short.

heightism! ✏

:-)

xnyhps for new ED? ✏

Re the discussion about XEP-0359 here, what you actually want is to revive https://github.com/xsf/xeps/pull/689 right?

Without reading and just going on the title, maybe :)

Discussion on the editor issue tracker.

nooo

> XEP-0359: Replace tabs with spaces. WHY?!

I like #689, except the implementation note says "to be set by the emitting client on every message to a MUC" - I don't see <origin-id/> as a MUC-only thing.

Link Mauve: you could have confused the Council by bringing up #689 right in time for today's Agenda.

Council don't need to be more connfused than usual :p

next week, he might be able to confuse himself with that