XMPP Council - 2019-07-24

Afternoon folks. Been a bit busy, but I'll knock out an agenda now - Ge0rG - I've spotted a couple of messages about Compliance 2020? Anyone else got anything?

dwd: those messages are redundant to each other. I've got no further topics to add

<reaction>👍</reaction>

🙈🙉🙊

🤣

Let’s do the entire meeting only with emoji!

Anyway. I don't see anything new to vote on, just your chat on Compliance Suites.

Link Mauve, From https://www.economist.com/books-and-arts/2019/07/18/the-internet-is-changing-language-less-than-curmudgeons-fear - "They are not a language (try telling a complex story in emoji to someone who doesn’t know it already)."

oh snap, I forgot an AOB

dwd: it's from the Reactions / References discussion on-list: generalize "this message belongs to past message X" into a small self-contained XEP, change LMC, Reactions, References etc.

+1 to that!

namespace-bump everything!

Link Mauve, To what?

dwd, +1ing Ge0rG’s AOB.

Maybe a bit early. ^^

Link Mauve: now it's too late to pull back

.

Finally managed an agenda.

No worries, we've got some more 20mins

You rang?

1) Roll Call

I'm semi here.

Link Mauve, ?

2) Agenda Bashing

Apologies for the very late one, but I think we just have two items of AOB to discuss.

Does anyone have anything else?

Yes, I’m here. o/

not from me

None from me either.

I need to catch up on a Lot of votes.

Well, most probably expired. Sorry I've been mostly absent recently, there's been a lot of Stuff.

3) Items for a vote

AFAICT, we have nothing.

4) Outstanding Votes

Kev, you have a couple expiring today; #797 and #796, plus two expiring next week.

FWIW, I'm pending on the last two proto XEPs, and unlikely to veto either, but haven't had a proper chance to look yet.

Everyone else seems to be up to date.

There was list discussion for both.

Indeed, I've not properly digested all that yet.

5) Next Meeting

Next week is fine for me (31st), but I'll be away sunning myself the week after.

I should actually be here next week.

+1W WFM

\o/

+1w wfm

6) AOB

We have two items, both fropm Ge0rG:

a) Road to Compliance Suites 2020 [Georg]

Ge0rG, Take it away.

(As in, start the discussion)

When 2019 was published, we had a heated discussion regarding how to do 2020. Should it still be a XEP? A different kind of XEP? Something completely different?

There were great ideas, but they'd require moderate to significant changes to our process and tooling.

(https://mail.jabber.org/pipermail/standards/2019-April/036074.html for context)

pep., Thanks.

Nobody has stepped up so far to make these changes happen, therefore I propose to either a) ask for a volunteer to do so, or b) ask Ge0rG to fork CS2019 into CS2020 and to go on with the old process

I suggest (b).

> Kev would like to schedule another discussion on this topic for the start of June, if only to determine it has failed.

Luckily, it's around the start of June now.

Heh.

I like the idea of living documents, but not with version numbers.

And living documents with year numbers is sufficiently close to what our current process & tooling gives.

I like many things, but the enemy of the published compliance suite is the dream of the perfect one.

Which is why I'd suggest to use August as a time window for somebody to step up and make that dream happen

And after that I'd be glad to fork 2019 for 2020, and add a "Future" section

I think the enemy of the published compliance suite is actually the way we currently publish compliance suites.

Kev: I hear you volunteer to create new process and tooling?

..until end of August.

Does it *need* new process and tooling?

Nobody ever does much in August, in the best years, though, and I have a feeling that once we hit September, and start the run-up to Council elections, etc, we'll never get anything out before 2020.

Hence, if Ge0rG is willing to crack on and do it, I'm firmly in favour of that.

I don’t have a strong opinion either way.

Kev: do we have process for "living documents"? Our process for RFCs sucks, and the XEP process is not well-designed for living documents

I don't want to manually keep track of changes in the History section.

Could we maybe schedule a higher bandwidth (short) discussion at some point using $UnpleasantVideoSolution with whoever feels strongly about this?

(soon)

And work out what the minimum needed to make the change is?

Kev: process-wise, we'd have to figure out a proposal and present it to Board.

I know I'm fairly outspoken on this, but I believe our current compliance suites, and especially the noise they generate every year trying to get them published is actually harmful.

I'm not sure that noise propagates outside of the core XMPP community.

Ge0rG: I don't think we need Board to sign off on a process unless it's overlapping with something we already have.

No, I think the noise it generates and the distraction for Council is the harmful bit.

If we want a formal process enshrined somewhere, yes it needs Board.

But I don't think that's a given.

Kev: besides of burning out CS maintainers, I haven't percepted that as a huge problem.

I think any new process for publishing a Thing is going to add a lot of overhead to whatever we do.

I don't object to scheduling a video conference for this, anyway.

Formally, we'll have to bring it up before Board.

Not if we just want a page on the website, potentially.

But I think doing this over not-text might be a better way to iterate.

Kev: we could have that today, with the tables extracted from latest-CS-XEP

It is an option.

OK. Is this going to be Council, or shall I open it up to other stakeholders?

I mean, the video call.

dwd: I suggest the minimum possible group to try to understand the issues.

e.g., Kev, Dave, Ge0rG would be fine.

Kev, Does that mean I don't have to come? *hopeful face emoji*

Not if you don't care about this.

Or don't care to care until later.

Kev, I'll come along and play chair. You don't think it's worth having any other client or server devs involved?

"Involved" makes it sound like excluding them from having input, which isn't my intention.

it's about process, not about what XEPs to put into CS

But I think a very small number of people understanding the problem would be sane. Two, Three, Four people. That sort of number.

OK. I'll arrange something.

👍

Ta.

b) "This message belongs to past message X" [Georg]

Ge0rG, Yours again.

https://mail.jabber.org/pipermail/standards/2019-July/036285.html section 1.

general agreement from my side

I think we should have a single consistent and archive-understandable way of doing this.

Ge0rG, big +1 from me too.

and it’s not References.

I think thats good design as well.

My preference would be to do it with 372, even if that means stripping 372 back or something.

references may be able to re-use that, but References is more specific than that

I'm currently very short on time, so I can't step up to write that proto-XEP. But it would essentially just provide a way to DAG messages to some root, and allow servers to emit the full DAG on request of any of the messages from it

jonas’: I think the reverse, no? References is more general than that.

Kev, depends

Kev: stripping 0372 to "this message belongs to that message" is actually an option I've thought about

DAG? So you can reference a reference?

dwd: you can LMC an LMC as well ;)

Kev, references wants to select a range of the body, which I consider more specific, but then again, that could be seen as more generic ;)

it’s not useful for non-body stuff in any case

dwd, in general yes, you may want to react to someone attaching a message to another for instance.

jonas’: No, it allows selecting part of a body, it doesn't require that (or the concept doesn't).

the only thing that References doesn't define is references to other messages.

Kev, if we can re-do References addressing the feedback I gave last year or so on-list, I’m game

Or, well, it says "If you have a URI, you put it here", but we don't have IDs for messages that are actually usable.

and we don't have a message ID URI scheme.

That is the easier problem to solve, I suspect.

I'd rather rename 0372 to something else ("Message Content Annotations"?), and make a new "Message References" XEP that's sole purpose is to reference messages.

Indeed - take <stanza-id/>, and do `xmpp:${from}?msg=${id}`

Ge0rG, we’ll need such a way to reference a message, whether it is encoded as a URI or not shouldn’t affect it in any way.

Link Mauve: Doesn't affect *this*, but I think there are advantages to being able to URI-encode. ✏

Same.

Using URIs will become a major distraction because then developers end up wondering how to accept those URIs from the outside.

And then security issues and maybe other problems.

I'd like to avoid opening up that can of worms.

which is why I think `<reference id="X" />` with something like https://xmpp.org/extensions/inbox/reactions.html#business-id to define what X is supposed to reference.

Ge0rG, the outside may be a future XEP which links messages, or something like that.

It may also be quite useful to have “the outside” for clients which want to scroll to a specific message.

inside XMPP, we can always clearly define semantics, we don’t have to wrap an ID into URIs to give an otherwise unstructured string semantics ✏

Link Mauve: this is another can of worms.

Guys, we're running out of time - might I suggest continuing this discussion on list or in another chatroom?

Using Discord a fair amount, which does this, I can say that I find it very useful (not very often, but not infrequently) to be able to drop a URI into a chat message as a URI.

Kev, Slack also.

But anyway.

I use Slack less these days, I find, but sounds plausible :)

Kev, but that doesn’t mean that "Slim References" necessarily has to use URIs internally

Unless anyone's any other AOB?

I'm not saying it's bad to have message URIs, I'm saying they are the wrong mechanism for the problem at hand

no other AOB from me

I don't have other AOB.

OK - go continue this on a list and encourage others to join in.

7) Ite, Meeting Est.

thanks dwd

Thanks. :)

tnx dwd

"dwd> I'll be away sunning myself the week after.", you haven't had enough sun already these past few weeks? :P

Thanks all.

pep., Luckily, I live in Wales, so it's been raining the entire time.

lucky you

It's been 30ish here.

rain?

what’s that again?

My trips to London, though, have been hothothot.

Kev, Dropped down today, for me. Rain last night, cloudy much of today. Sunny now, though.

Rain last night for me, was fine this morning, been getting slowly warmer through the day, but nothing unpleasant.

Yesterday was unpleasant, though, and tomorrow it's meant to go back to that.

I suppose the unexpectedly early end of meeting means that somebody needs to write up things and go to standards@?

s/early/on time/ but I accept, with some shame, the "unexpectedly".

dwd: not an attempt to blame anyone, or you, for it. It's ending at that time every week.