XSF Discussion - 2023-05-04

XEP-0482: Call Invites <3

Hello, almost Board time ✨

Yep!

Until then... Bored time...

:-)

📯🔔📢 Board is on time - right?

0. Welcome

Hi!

⚡

Who do we have?

Present!

MattJ, Arc?

Here o/

👋

HI!

I actually don't see Arc in here and he doesn't show as online to me, so let's continue.

yup

1. IETF MIMI WG / MLS / DMA

Anything to report at this front?

I can report on my recent activities.

Please do

I’ve had two conversations with the MIMI WG chairs.

(Everyone must have got my reminder for the meting yesterday) ✏

After the first conversation I talked with MattJ.

And MattJ has started an email thread with folks who are interested in working on an XMPP profile of MLS.

(emus: I just used my calendar :-) )

> (Everyone must have got my reminder for the meting yesterday) per e-mail? ↺

I’ve also reached out to the MLS WG chairs but I haven’t heard back yet.

Here is a short summary….

> ralphm: > 2023-05-04 07:04 (GMT+02:00) > (emus: I just used my calendar :-) ) (yes just to be sure after last meetings)

Arne: just to board

The MIMI folks are basically hoping that the threat of DMA fines forces the “gatekeepers” (big tech companies with chat systems) to get involved with the IETF effort.

But I am skeptical that will happen.

> stpeter: > 2023-05-04 07:05 (GMT+02:00) > I’ve also reached out to the MLS WG chairs but I haven’t heard back yet. > Here is a short summary…. on which regard exactly have you reached out?

Supposedly some of the gatekeepers are open to a standardized solution, not one-off bilateral agreements.

But gatekeeper representatives are not involved yet.

I think it hinges on several things: a) MIMI offering a viable approach, b) the EU Commission forcing their hand

And we don’t have anyone from the XMPP community who wants to dedicate lots of time to the MIMI effort. I was asked and I declined because I’m semi-retired, y’see. ;-)

And my personal feeling is that a) kinda requires XMPP

> stpeter: > 2023-05-04 07:07 (GMT+02:00) > And we don’t have anyone from the XMPP community who wants to dedicate lots of time to the MIMI effort. I was asked and I declined because I’m semi-retired, y’see. ;-) Are there further other active actions we should start over to encourgae that? What is preventing?

So IMHO MIMI is sort of dead in the water until (b) happens

Well, who wants to work on MIMI? It’s basically volunteering a whole bunch of your time to work on technical solutions that will save Google and Facebook millions/billions of dollars, but you’re going to do all that work for free!

And we’ve been down this road before!

emus: it requires someone who can navigate the IETF, the XMPP community and its standards process, as well as someone apt enough to work through the politics of this all. Most of the problems aren't really technical.

@ralph right, which is why they tried to recruit me. ;-)

Is it about paid time for people from our side?

As to MLS, that seems much more viable.

MLS is also kind of a requirement to get MIMI done.

I think it's a very time expensive thing. I see value there, but I see more value in other things most people are spending time on (sorry if I should be silent during meeting I will be after this if you say)

Eventually we might want to have MLS-based solutions to MUC encryption and so on, and we have people who are interested in working on that, either at the XSF or the IETF.

Because the DMA requires an e2e component

Do we still need to develop an official position from "XMPP / XSF side

It turns out that the MLS might not be the place to work on MLS profiles (like MLS over XMPP), so it might be fine to do this work at the XSF for now. This is why I have asked the MLS WG chairs for their insights.

singpolyma, no, board meetings are open to participation from "the floor" :)

s/the MLS/the MLS WG/

And as Ralph says, MLS over XMPP is a good first step toward DMA compliance.

> stpeter: > 2023-05-04 07:11 (GMT+02:00) > It turns out that the MLS might not be the place to work on MLS profiles (like MLS over XMPP), so it might be fine to do this work at the XSF for now. This is why I have asked the MLS WG chairs for their insights. But noone wants to spent time on it?

@emus we do have people who want to work on MLS over XMPP

My feeling about MLS on MUC should not be hard. I.e. not as hard as it is to get Matrix going (hence the dMLS stuff). But unfortunately I haven't spent enough time on it to make a proper determination.

Ok, I understood the opposite

But we don’t have people who want to define a new S2S protocol that Google and Facebook can use.

stpeter: sure

Right, because they can just use XMPP s2s as is

I mean, ideal this new s2s protocol would be "XMPP" but I guess they want a big chunk of it written down in a certain way

My understanding was that we had Tim and Marvin interested in working on MLS. Correct, MattJ?

Yes

One piece of tech-ish feedback from the MIMI chairs is that the gatekeepers think XMPP s2s is a non-starter because it’s too hard to scale. They don’t know how to manage long-lived sessions across multiple servers etc.

moparisthebest, singpolyma: agreed, though there are some people (involved in MIMI) who are sceptical about that

Well, my only concerns are that this very important topic is not actively adress to our best capabilities ✏

Because we don’t have direct connections with the gatekeeper people, it’s hard to have a real conversation even if we wanted to.

The gatekeepers opinions don't matter, they aren't even involved and likely won't use any of this anyway

@moparisthebest I tend to agree.

moparisthebest: except that it's being done "for them"

stpeter: I am surprised that this is still the argument. Is that really a problem? And if so, we could redefine s2s streams to not require permanent connections.

I don't think it even requires a protocol change.

I think you can already close s2s stream when you like and reopen as needed it's just less efficient. Something something quic

@ralphm Right, but someone would need to define that option in XMPP - basically negotiate it upfront or switch to non-persistent version.

s2s streams don't require permanent connections already? you can time out whenever you want. with direct tls and bidi it's super fast too

Plus "won't scale" etc is just provably false ie https://www.process-one.net/blog/ejabberd-nintendo-switch-npns/

Google Talk proved it scales just fine

singpolyma: sure, defining XMPP on top QUIC is something we should do anyway. The problem is probably in proper library support for the relatively new IETF QUIC.

@Zash I believe the concern is keeping a user session alive across the s2s restart etc.

Anyway, this is getting into a technical discussion.

Getting a real contact at any of the gatekeepers would be a huge win independent of MIMI and DMA stuff

And the basic problem here is people / political.

Zash: Google literally used "XMPP is not ready for the cloud" as an argument for shutting down federation and eventually Talk.

ralphm: we already have XMPP on quic defined and implemented and deployed...

So we need someone to ~bribe~ lunch with the right people in the right lobbies then. Who likes to eat? :)

@singpolyma We had great contacts originally with the Google Talk team but those atrophied and their commitment to XMPP wavered once a new product manager took over. Etc. I don’t see a lot of upside here.

stpeter: (as you are writing miniutes, can you liate the technical implentations mentioned)

“We’re trillion dollar companies, please do all this work for free to save us from EU fines” is simply not a great offer.

stpeter: right. That's one issue with all political work. New people always come and go and it needs constant renewal. It's not a project more like a job unfortunately

In any case, we're more or less where we started: nobody that has the cycles and/or desire to do this work.

stpeter, ... "or we'll do it an even worse way!"

If money is the problem, however, we can apply for funding, and likely get that sorted.

EU funding?

Indirectly

So don't know the problem? or is it money?

These companies are likely to create bilateral agreements or one-off APIs in order to comply. They will not be pretty, but they will help the gatekeepers avoid fines.

Unless XSF wants to become the kind of org that has staff I think funding probably won't help this particular kind of issue. Unless it's just about flights or whatever

there is also the german souvernaity fund ✏

what about the sovereing tech fund?

Arne: yes, mentioned that

May I ask, how far is the progress of the application? ✏

> Arne: > 2023-05-04 07:21 (GMT+02:00) > May I ask, how the progress of the application? no application

Well, there’s also a question of who the hell wants to do this kind of political / standards work for literally years on end.

Yes, they aren't going to federate with the open XMPP network no matter how much we butcher XMPP s2s to please them, so why expend any effort?

stpeter: there was discussion on this at the DMA meeting in Brussels, and all parties there, except Meta of course, made it clear that they didn't think that bilateral APIs were going to be an acceptable approach. The summary by the EUC representative made me think that the C would agree.

Remember their goal is user lock in

I't rather start with a funding and work towards this if that is what we all believe is the right way

Even if we think there is a high likelihood of success from the XMPP perspective, we’re talking about 3+ years of unpleasant political / standards work. I’m pretty sure everyone here would rather spend their time writing code and making their software better for real people.

If we put in a good proposal NLNet will likely provide funding. But as I said, and Peter is as well, somebody has to want to do this. And also have the qualifications I mentioned above.

> Well, there’s also a question of who the hell wants to do this kind of political / standards work for literally years on end. Yes, it's a kind of work that's extra hard to do long term with volunteers ↺

ralphm, stpeter: but please lets define what we believe are required and meaningful questions and how to adress. then we can call and see how we could reach this in a second step

If I hadn’t been working for JINC and then Cisco, I don’t think I would have done all that IETF work for the last 20 years.

I don't have anything further to say and I don't think we have any action items here (other than generally keeping on top of MLS)

@emus Sure, we can do that.

So I'd focus on the people part first, then arrange the funding.

But in any case I think a good first step, no matter what, is to jumpstart the MLS over XMPP work.

Action, people, funding, rather? ✏

That’s a concrete, useful effort, and we have people who want to work on it.

Should we make it an official team? I mean its really important right?

I think it makes sense to take a wait-and-see attitude w.r.t. MIMI.

emus: so far the team is called Board, and I don't think that creating a dedicated work team will change the outcome at this point.

ok, but you mentined others too right?

@emus It’s a XEP. Potential authors can submit a proposal anytime. Not sure we need a team for it.

sorry, matt mentined

ok

emus: I mentioned people who have previously indicated a desire to work on MLS. I don't know what the current status is, but maybe MattJ knows

ok

And I am no longer convinced that the MLS WG is the right place to do this work, until we have a conversation with the MLS WG chairs.

For MLS we may form some kind of official team, but there is no pressing need

@ralphm we have an email thread going

stpeter: open for all? :-)

(Which I need to reply in given my recent discoveries.)

Various people in the MLS WG have indicated they'd help out if needed. Including Raphael Robert.

(whos that?)

@emus Just an organizational thing among potential spec authors. Once they get organized, I’m sure a XEP will be forthcoming.

@ralphm Yes, Richard Barnes offered to share insights, too.

Anyway, I think we have enough to go on for the moment, let’s see how the next few weeks shake out and we can discussion again in the next Board meeting.

(Anything I can do even though I lack may required competences?)

stpeter: where is the thread?

😄

(I may have missed it completely)

Agreed on moving on.

2. Infra

I understand there have been issues and work was done. What's the latest?

fine, but lets write good mnutes here and cycle them

@ralphm It’s not on any list yet.

stpeter: ok, then it makes sense I haven't seen it :-D

> stpeter: > 2023-05-04 07:31 (GMT+02:00) > @ralphm It’s not on any list yet. lets move it there please

@emus Yes, I will write up some minutes based on the discussion here, likely tomorrow.

sure

MattJ: infra?

(typing :) )

There are no major infrastructure emergencies right now, but some ongoing issues of various severities. Top ones that come to mind are the lack of web-accessble archives on the mailing lists since restoration, and a lack of IPv6. The first one is within our control, the second seems to be some issue at USSHC and so far Jerry has not responded to a couple of pings about the issue (but I didn't press too hard because he was helping us with more urgent issues). I'll try and get on top of both issues as soon as I can.

The lack of IPv6 affects all our USSHC servers, so for us that's primarily our mail server, DNS server and everything we host at xmpp.net (beyond the XSF, jabber.org is also affected). The website and XMPP service are not affected.

Is lack of IPv6 a problem?

(or a desire)

Yes, it means the service can't be accessed by people with IPv6-only internet connections, for example

Right. I have no idea how large that group is

and we can't deliver (or receive mail from) IPv6-only mail servers

It's not the largest group on the internet, but it's certainly a group

One small part of holding us back in the Legacy IP dark ages!

Yep. Ok. Thanks for the update.

3. AOB

Any OB?

Thanks!

Well, another issue is that we still haven’t fully configured mailman - we need to install/configure nginx and re-enable list archives and perhaps a few other tasks. I’ve been traveling so I have been remiss in working on this.

(At least I think those tasks still need to be completed, but I might be out of the loop.)

I did mention that above, I'm planning to look into it

stpeter: Maybe we can ask singpolyma for help if thats required

stpeter: I think that was in MattJ's message

OK, sorry, my attention wandered.

No worries

In finance news, I just reimbursed Ralph for FOSDEM/Summit expenses.

Any other business?

YES! And got it, too

That was somewhat expensive. :-)

Well yeah, it is has been a while since we needed to pay for a venue

wait I may have one extra announcement

So instead of $16k+ in the bank we now have $13k-ish in the bank.

But we can talk next time about sponsors again and I can provide a more complete finance report.

About finances, I'm thinking about contacting a potential investor. He has some small political contacts and he came to me once after I won a price for my project. I could try to introduce him to xmpp. Maybe anyone here could help then?

Na not yet

?

Arne: sounds intriguing.

(nothing to announce yet from my side)

Arne: nice!

@emus OK, but the suspense is killing me!

So yes, we can help

No more AOB here.

emus: guess the next meeting then.

4. Date of Next

ralphm: later hopefully :-0

> So yes, we can help Thanks ↺

June 1 or 8?

June 1 works for me

7th in my cakender

but should be 8 righy?

AFAIK we haven't picked a date in June yet

I thought we did. ok wait

id prefer 8th

WFM

Next meeting is on June 8, 17:00 UTC.

5. Close

Thanks all!

Thanks!

but actually, didt we agree on first Wed? thats how it is in my calendar

Thanks and bye!

I thought we agreed on Wednesday?

I mean Thursday.

Anyway, we can work it out on the Board list.

I need to run, cya later!

fun :-) I got blamed sending a reminder :-0

kk bye

emus: We only recognise 64v3vs15qlalgqv0j7r99ikm1c@group.calendar.google.com

ralphm: well didnt worked for arc I think

arc always has calendar issues. One of the constants in the universe

🙈 people don't believe we work in tech

Good night

Tech workers have the most tech problems, probably.

Relevant https://www.moparisthebest.com/images/enthusiasts-vs-engineers.png

A printer? Someone likes to live dangerously

Printers often make problems

^^

ralphm, stpeter: But we used to say that we want to have a strategy before be start asking for a funding in any direction. now I understood it was mention to "simply" look for people and funding. Im a bit confused... Btw, I want to remind that nicola, a lawyer and politician from Italy offered support in eg. legal questions if that is required.

emus: I strongly believe that we should first find a person that would want to lead this effort, and has time to do this. This could mean that instead of doing other work, this person will dedicate his time for this *if paid for it*. But we'd first have to find a person that is _willing_ and _capable_. Then if funding is the "only" concern, we can try and find funding for that.

But I'd rather say lets set the goals or what ever we want to work towards and set the scene. why only one person. maybe the could be done by people part time. or people act on certain parts, not all of it. I mean we are looking for one champion, but maybe thats too high of an expectation?

Are we still talking about doing political/ietf work? Because that's at least focus of one person to realistically happen in full

Rag tag group can do a lot but some stuff needs continuity

sure, not ignoring that