XSF Discussion - 2019-02-20

https://dot.kde.org/2019/02/20/kde-migrating-matrix

> Other alternatives, such as Telegram, Slack and Discord, although feature-rich, are centralized and built around closed-source technologies and offer even less control than IRC. Pretending XMPP doesn't exist

btw, they used to have an XMPP server

My first or second account was on their server

They kind of ignored me when I tried to approach them, so this does not surprise me

Sad news in my opinion but it is up to them in the end

I personally don't care, KDE is not representative

What is a KDE?

I made this table with help from others here https://github.com/SeveFP/KDE_IM_Requirements/wiki/XMPP-Table and sent it to them and try to see if we could do something together but they didn't show any interest on it

when they announced they wanted to move to some solution

Let's try to focus on increasing our exposure, instead of discussing the exposure of Matrix. Let's learn from what they do right.

(That's actually a nice table, Seve!)

Guus: they have money

I mean, we could have polished UIs if more people were working full time on XMPP

Goffi: given that there are _several_ commercial providers of XMPP solutions: we have money too.

but I know only 2 clients which can do this: Conversations and Converse, and both are only done by one people as far as I know.

And there are ways to gain more money (we'll discuss this in upcoming board meetings too)

Try to think of solutions, not of problems.

you need to know the problem to find a solution

You identified two recurring issues: lack of money (which I doubt we have, to be honest), and a need to improve UI / UX in our clients.

I'll add a third one: we're not showing off enough of what we can do.

we have money in servers, not in clients.

"not having money" is not a problem. It's something that prevents us from finding some easy solutions.

"not having a good UI" is a problem.

let's worry about that.

I've tried reaching out to OpenSourceDesign, where I need to to follow ups

Ge0rG et al. are making progress on UX guidelines

Guus: it's link to being able to work full time. You need time to focus on polishing.

linked*

goffi that helps, but isn't going to fix UI/UX things. You need expertise, too.

where do we find these?

do we integrate our solutions at customers, where such expertise is available?

maybe we can ask for help there.

We have several pretty successful projects that are making money. Maybe we can start sharing experience on how to do that, so that more projects can reach that level.

We can think of plenty of things that can be effective. That's way more stimulating that just being jealous at Matrix for "having money".

I'm still convinced that the XMPP proposition is at least as good as Matrix's one. So with that in mind, we must be able to at least match what they're doing.

but we'll have to put in the effort.

I'm going to visit a commercial org tomorrow, that I've met at FOSDEM. They expressed interest in stimulating XMPP. They offered to host a sprint, and they like to see the UX of the IOS clients to be improved. We'll discuss that tomorrow.

let's find more of these people, and reach out!

Guus: who's jealous ? it's not about being jealous, it's about explaining why they have more polished UI and attraction.

Guus, oh is that the one that we discussed with?

let's do more outreach - write more blog post, show off stuff that we _can_ do moar, better, faster.

goffi: so let's fix that. But I'm convinced that there's a level of jealousy in the XMPP community. In some MUCs, "matrix" is mentioned more than the topic of the MUC, which I think is such a shame.

pep. the two guys who talked to you after your presentation, yes

There have been lot ot unfair attacks from Matrix community (and leaders) toward XMPP one, specially at the beginning (it's slightly better now), so there have been reactions to that. And it's important to know what going around in other protocols/software.

Could someone help me get job postings at https://opensourcedesign.net/ for clients that want/need help from designers?

I want help from them as a server operator, for now :x I need to post that someday

please, do!

I agree with goffi - money is the problem

a few full time devs and you get your UI

and we obviously don't have money, I don't understand why Guus said the opposite

zinid Money is not a problem, it offers a possible solution to a problem.

We do critically lack resources

At least in the public part of the community

so, let's find resources.

Guus, money is the obvious way to get human resources

Guus, other methods have failed

we tried a lot of them in the past 15 years

don't get me wrong: I'm not saying that we should not try to get more money. That's why I put this (sponsoring, financing) on the agenda of the board again. But I am sure that we can find more ways to attack the problem.

Guus, well that boils down to other people spending this money (time)

s/other //

pep. As XMPP is used pretty much everywhere, that must be possible to do.

All the sprints we do are not free, for example, time, transportation, accomodations, that's all on our own. Even if at the end I'm happy to do it, that comes with a cost

I think the sprints are great to have, and I ❤ the people spending time/resources to make them happen!

Do we have employers (or customers) that are somehow invested in XMPP, and have UI/UX employees that they can spare for a day per week?

commercial organizations do OSS sponsoring all the time. This could be one way of doing that.

> spare for a day per week? That's not how development works

you need some focus and dedication in order to produce something valuable

zinid: then find a collaboration that works better.

It's already better than nothing.

but in any case, 1 day per week is better than 0 days per week.

I don't buy that argument honestly, I hear it a lot

my point is: let's try to find ways to pull in resources.

And that's already how lots of the free software projects around here work

what you can say is 1 > 0, that's it

zinid please offer better solutions then!

I'm open to anything

as long as it's not "we don't have money so we're doomed"

I'm working part time (80%), that mean that I have one day a week I pay on my own to work on my project, and I'm overwhelmed and have no time for now to work on UI polishing. I don't say we are doomed, but I say it's not sustainable on the long run without paid full time dev.

pep. I'm interested in your reason for searching for UI/UX experts as a server operator. What exactly would you like to see improved there?

See my wonderful design skills: https://cluxia.eu/

I'd like to have a color theme / logo

I have a few ideas, but no clue how to put that together

that seems like a solvable problem 🙂

goffi, I'm guessing somebody else apart from you working on the project would also help

I understand is hard to find collaboration, I know

pep. I think I can set you up with someone that can improve that page

pep. could you send me a brief "this is what I would like to happen" summary?

Sure I'll try to write something down. Thanks

> The requested page "/2019/02/20/kde-migrating-matrix" could not be found. Looks like somebody pulled the plug

Guus: "we have money" is illusive, because we as the XSF have a complex process to access that money, and it is (rightfully) demanded that the money is spent in a fair and impartial way. We can't just throw 10k€ at improving Monal, even if we have consensus that we badly need a good iOS client, nor can we throw 10k€ at me implementing my favorite yaxim feature.

I can see how sponsoring (open-access) Sprints and Summits and conferences is a Good Thing.

But this only goes so far in bringing forward XMPP

Ge0rG all true. On top of the XSF getting more money, I'd also like to set up a way for individual projects to share knowledge about making money for themselves.

Some projects are more successful than others.

We for sure can help to promote crowdfunding campaigns though, if Monal has one, for example

We can learn from eachother there.

Guus: I know I'm sounding like a broken record, but this would be a great task for The Jabber Software Foundation™

Ge0rG make it happen, please.

Ge0rG, you mean sponsoring etc.? Or teaching people how to get money

Because I also want in on these tips

Guus: I learned the hard way that You Shall Not Create A Jabber Org.

Ge0rG doesn't stop you from preparing to get people together, start collecting / sharing information, right?

pep.: promoting sponsorship awareness as well as redistributing money

the name is just a name.

Ge0rG, yeah well that was discussed at the summit, and I think we can do something with what we started already

\o/ progress!

I'm patiently awaiting (I long) for the time I have _any_ time to put in it.. hopefully soon enough.

Guus: I agree about sharing knowledge, but people may fear to see councurent projects raising and taking there grants/customers.

their*

goffi we don't know before we try.

Guus: what pep. said. We all have things on our TODO lists. But they always get overridden with higher-priority non-XMPP things.

goffi I'm also more thinking of sharing _techniques_ rather than _customers_

Ironically, some of us might be able to invest more time if it were paid ;)

to give a very practical example

So does anyone have a cache of https://dot.kde.org/2019/02/20/kde-migrating-matrix - the page was removed :(

someone told me it'd be a good idea to have a 'professional services' section on the igniterealtime website, so that people that search for ways to spend money on Openfire, Smack, etc would find people willing to accept money.

we put up a very basic section: I've already been approached through that a couple of times in the last few weeks alone.

Guus: what do I need to do to get enlisted there?

...with my employer's hat on?

Ge0rG a credible link as a service provider of ignite-realtime related projects.

Guus: could the XSF help to put people willing to spend money for service/feature reach people offering them? Like a dedicate page on XMPP.org?

(who's your employer?)

Guus: https://rt-solutions.de/en/

goffi I don't know - we can consider it.

Ge0rG from a quick browse on the website, I'm not seeing obvious Openfire/Smack service providing offerings?

Guus: as a person, I'm heavily invested in Smack (and obviously experienced in XMPP as a whole), but there is no way to rent me out except via my employer.

Guus: the latter is expensive, and heavily targeted at German companies, but there is a market for that, and if you need some kind of page on our homepage about competent XMPP support (I'd prefer XMPP over specific Ignite projects if possible), I can probably make it happen.

Ge0rG the way that we've set up that page at IgniteRealtime, we expect that visitors of the 'professional support' page to be searching for very specific solutions / expertise. So, if you'd like to be listed there, I suggest you have a landing page at your employers site, that lists the (Smack-based) services that you can provide.

I shared the IgniteRealtime "professional support" thingy as an example of how other OSS projects could also try to attract more money.

I'm not saying that this is the only solution. goffi already suggested having something that's more XMPP-generic - which seems to suit you better.

So, although I'd certainly consider adding your company to the Ignite listing, I'm also/primarily using this as an example of things that other projects could also do.

and no, that won't fix all of our problems overnight

but by sharing ideas like this, we _will_ improve things.

goffi From a quick browse, I'm not finding a way for me to hire someone on my salut-a-toi problems (if I have one), at https://salut-a-toi.org/

Maybe considering adding that to the website?

Maybe consider adding that to the website?

Guus: yes, I'm currently redoing the website so I can add this. But SàT itself was not ready so far to be used in production, it should be alright starting with the incoming release.

We've made a couple of strategic mistakes, starting by not doing the famous "release early, realease often"

goffi I have no experience with it, so I can't judge. One thing that I can say is that I am sometimes amazed at how people use pre-release, not-ready stuff in production anyway 🙂

Wait, isn't that what production means? :)

pep. you also deploy on Fridays, don't you?

I deploy on weekends, even worse

yes, that what one of my friend well aware of business stuff told me: you should not wait to have something bullet proof, nobody does.

goffi sounds like you have a friend that could offer ideas to more of the XMPP community!

> zinid please offer better solutions then! > I'm open to anything > as long as it's not "we don't have money so we're doomed" Guus: I have no solutions except money, and even that is not easy to collect and redistribute

maybe you should invite him to share ideas

zinid I just shared an idea on how to make money. We can use more ideas like that.

Guus: he already wrote interesting reads on his blog: https://blog.addictedtointer.net/2017/09/26/goodbye-cozy/

I'll share another one. I've heard of several people in companies that would like to 'sponsor' an open source project that they're using, but setting up sponsoring through their corporate procedures is difficult (not standardized). What is, however, easy for them to do, is to buy a 'support contract'. So, OSS projects could start offering 'support contracts' with the idea that this really is just sponsoring.

I'm currently not doing that, so I can't tell if it's effective - but again, it's somethign that you might consider for your project.

As a third, and final, idea / food for thought: I've learned about idea 1 and 2 when I met up with people at either FOSDEM or (what then was not yet called) sprints. I highly recommend getting together like that!

Guus: those things should be documented in the xdf maybe?

https://xmpp-developers.foundation

Maybe. Go for it. Not sure if these are thoughts to be stamped as 'approved by the XDF'.

Ge0rG it's discourse link is broken: https://discourse.xmpp-developers.foundation/

Somebody needs to open an issue on https://github.com/XMPP-Developers-Foundation/website I suppose

I am already ready to oppose the use of discourse, fwiw

We already have a venue for that, which is jdev@

https://github.com/XMPP-Developers-Foundation/website/issues/1

Which I was planning to discuss with daniel

is jdev@ publically logged?

I mean the ML

Ah.

Maybe it's time to revive it instead of pretending it's not there

off to lunch with me.

There were six mails this year.

Lunch is an awesome idea. BBL

Or.. if you really want to kill that list, it's not impossible to have bulletin boards on XMPP, and goffi has a barebone implementation of one in SàT. Porting discourse to XMPP would be another idea.

pep.: Flow mentioned something like that for Ignite Realtime earlier

Using smack

Like which part?

XMPP Discourse integration

I see

By that I meant not just authentication right

I also meant using XMPP as the storage backend (pubsub)

I think Flows Idea was to post replies to discourse, using XMPP

Well if you have it using pubsub, yes that'd be possible

I think it already offers an API that you can hook an XMPP client library in

Unsure, talk to Flow. 😉

The thing is that I'd just want discourse to be a dumb frontend

Then we could use xmpp as the transport and matrix as the backend! 😜

Unsure if that is achievable

It is if you write it in Lua

:D

but that's a point which is annoying me a bit, we have tools not perfect but usable and which can be easily improved, and community want to use non XMPP software instead. How can we promote XMPP if the community itself doesn't use it?

(lunch time)

XMPP based forums is something I dream with since long time ago actually, I will ping you goffi I think, I'm really curious

we do use XMPP, don’t we?

jonas’, how is jabbercat doing?

oh, 2 months without a single commit

It's dead! Don't use it it's unmaintained!!1!1

I won't, okay

Is there anybody in the XMPP community who does *not* have their very own client? That's so nice about XMPP: You can choose betwenn many, many clients, all bad in their own way. Fixing issues would be boring, so why not write a new one?

ME!

but I have my own server. Does that count?

Real server devs have their own client too

And I have only a bot.

only one client? you need one on each platform

Also, I'm wrapping every web-based client that I can find as a plugin... 😛

MattJ: what's yours? :D

clix, for the command-line

Also scansion?

scansion counts too, I guess :)

But isn't for users

...as opposed to clix, right?

Absolutely :)

Uhm. Yeah.

zinid, currently the improvements are more happening in aioxmpp ;)

looking forward to implement MIX there and then in JC

jonas’, you need to work full time on JC, JSF (created by Ge0rG) will pay you

zinid: and you will pay the JSF?

Ge0rG, no, I'm affiliated, I cannot pay, I can only receive

zinid, sounds like a plan

btw, the current state of matrix.org patreon is ~$3800 per month

I wonder if the xsf can collect similar money

jonas' why doesn't JC work on JC? :)

> I wonder if the xsf can collect similar money we easily can

Alex: I don't find the xsf at patreon 😁

I see only sponsoring page where the amount starts at $1000

Seve: https://www.libervia.org/forums/list it's really basic but working. The hierarchy of forums is stored in a pubsub node linked to the language (so you can have different topics in different languages), and then it's XEP-0277 blog nodes with open access. Experimentation is concluant, so that one of the protoXEP I need to write and propose.

Ge0rG, looks like the blog post was edited: https://dot.kde.org/2019/02/20/kde-adding-matrix-its-im-framework

They say matrix provides "presence" in that article, I wonder what they mean by that

Is it just the status? online/unavailable/offline/whatever

Ge0rG, can you confirm real quick that you’re happy with this version of the '156 diff? https://github.com/xsf/xeps/commit/160145b9152fbae62620637a70830e91b597987e

jonas’: 👍

non-emoji please

jonas’: yes

thx

jonas’: the English is a bit bumpy, but I'm OK with the content

feel free to fix the former

Or... I'll just go back to work

or that

Ge0rG, also to answer your question, http://logs.jabber.org/jdev@conference.jabber.org.html, but that seems to have stopped in 2015, not sure why

I wonder what this was, http://logs.jabber.org/devcon@conference.jabber.org.html

hmm, and https seems borked: "The certificate is only valid for the following names: logs.xmpp.org, muc.xmpp.org, www.xmpp.org, xmpp.org"

ohhh, logs.xmpp.org ~

less links though :(

What's techreview@muc.xmpp.org?

No description, no name

hmm, and topic says "visit xmpp:xsf@muc.xmpp.org?join instead"

https://jitsi.org/news/telephony-support-on-meet-jit-si/ anybody tried their gateway support? Also is it possible to self-host that dial-in gateway or is it all closed?

pep. although I'm not sure, I think they're using a third-party dial-in gateway

I'm guessing that they're pulling in the media through their SIP stack (but that's a total guess)

What's the state of each of the XMPP IOS clients?

quite bad

what's lacking in clients, specifically

ChatSecure is abandoned and Monal has strange distribution policy and full of bugs

_apart from money_ 😉

ChatSecure had a release last December?

that's what I'm told 🙂

Guus: maybe, but that release didn't fix some show stopper bugs. And general development seems to be down to yaxim speed

What are the show stopper bugs?

is someone in contact with their developers?

(are they here?)

zinid: what's with the Monal distribution? The author fixed GDPR some months ago

there was something about OMEMO the other day

Ge0rG, he invented yet another problem recently

Ge0rG, https://monal.im/blog/omemo-and-french-laws/

Guus: https://github.com/ChatSecure/ChatSecure-iOS/issues/906

Guus: I've chatted with Chris some time ago, but he looks too busy to push forward

zinid: that's a problem invented by the French government, enforced by Apple...

Ge0rG, maybe

Ge0rG that's just one bug. That surely isn't stopping ChatSecure from being the Conversations equivalent on IOS?

And regarding bugs, yeah, Monal is very fresh. Don't use it for MUCs yet.

Guus: it's generally crappy, has a horrible backend store performance problem, doesn't quite get stream management right and is more abandoned than not. That bug is just my canary for a simple issue that's very annoying for users and easy to fix for a developer.

But yeah, it's the Conversations equivalent on iOS! 🤣🤣

What flavors do we have on IOS? ChatSecure, Monal, Beagle, ... ?

depends on flavors you prefer 🙂

I'm primarily trying to see if there's one that's most likely to be pushed over the finish line

Monal, seems like

but I'm not sure you can communicate with the author

there is also Xabber, which looks like a prototype mockup

Xabber doesn't even support MUC

Guus: currently, Monal has the highest trajectory. However, the developer is on a multi-week trip right now

What are the current issues with Monal?

MUC support, apparently?

Holger, did you try Monal recently?

Guus: MUC support is very rudimentary, OMEMO is apparently very freshly in.

I can't evaluate the latter.

zinid: Yes, kinda better than ChatSecure, not good of course. The author is more active right now.

Holger: please elaborate

Guus: Better than ChatSecure because (a) I didn't see it loose messages and (b) it has that VoIP flag that allows it to use silent notifications (unlike ChatSecure which produces "New message!" notifications). Still not good because bugs (duplicates and whatnot), UI is meh, and MUC still rudimentary as Ge0rG said.

Tx

Does anyone other than Guus and Ralph have expenses that need to be reimbursed from FOSDEM? I plan to go to the bank on Friday to send reimbursements.

(nothing pops to mind - I think I reimbursed everyone that spent money that wasn't sponsoring - but I might be forgetting something)

OK, thanks.

intosi - didn't you buy stuff for the stand, perhaps?

> Xabber doesn't even support MUC Their dev told me they invented something better than muc and mix

Yeah. Sure.

if it's been implemented once it's already better than MIX ¯\_(ツ)_/¯

🙂

> Their dev told me they invented something better than muc and mix the whole point of xmpp is standards, I don't see any points in inventing proprietary extensions, eventually you'll end up in supporting both: you proprietary wheel and the standard one

rion, they re-invented group chat 1.0

Isn't part of the point of XMPP the freedom to invent your own extensions?

Zash: in theory yes

in practice I already said how this will end up

and I know that because we at processone have such problems now

note that was against that shit from the beginning 😁

*that I

we have currently proprietary SM, proprietary push, proprietary MUC. some customers want that or this, interaction between the extensions is a nightmare

ouch

Interactions between standardized extensions can be complicated as well:)

Zash: true, and having proprietary don't help at all, quite contrary

Just because you have the freedom to do something doesn't mean it's a good idea. ;-)

> they re-invented group chat 1.0 I didn't see the protocol. But he promissed to add support to every piece of xmpp software.

lol

> if it's been implemented once it's already better than MIX ¯\_(ツ)_/¯ It's already implemented, working on server and one client, covered with tests, has nicknames, badges, history search, messages retraction, flexible system of permissions and restrictions, can work pretty well with legacy clients, can work if participants server does not support it, can work with multiple devices (seamlessly even if one client does support it and another does not)

Everyone interested may come and see. Like, right now.

rion, since you are Russian I can give you access to Google doc

> What are the current issues with Monal? My biggest issue with it was that it just doesn't work.

It works much better than Xabber ios... 🤷‍♂️

"implemented on 1 client and server that work together" makes it better than MIX regardless of anything else

you can go into endless discussions about whether MIX is superior protocol-wise, but none of that really matters when it seems like no one will ever implement both ends

Ge0rG, > It works much better than Xabber ios... 🤷‍♂️ It is a false statement.

Andrew Nenakhov: less false than: > My biggest issue with it was that it just doesn't work.

Xabber for iOS is in pretty reliable state, and my Monal has never-ever received a push notification.

I've tested both, latest Testflight

Can record you a video like right now of monal and Xabber running concurrently

Xabber didn't even connect to my account, had to register one on the Xabber server

And even then nothing worked, most functions were just mockup

Maybe some non standard configuration, whatever

Ge0rG, liar )

Yeah. Right.

Andrew Nenakhov [21:38]: > Ge0rG, liar ) zinid, is it you?

Andrew Nenakhov: I take it as a hint that you don't want me to retest and open issues on the tracker? You are probably already fed up with my issues on Xabber android anyway... 😁

No, I just want to to stop lying

Andrew Nenakhov, maybe don’t assume the worst right away

Just a sec, uploading a video taken right now

instead, consider: - there may be a situation with some devices which break things in your software - there might be a language barrier of some type - any other technical weirdness which might influence how the app is (not) working

https://photos.app.goo.gl/Z4kh4CF7nb8mvRUX7

jonas’, indeed it is possible that he has some weird setup for connecting. Besides ours, we test against jabber.ru, xmpp.jp and jabber.at servers.

hm I thought that would be a video of a group chat.

Andrew Nenakhov: my statement was as true as your statement about Monal. I've tested both, Monal worked, Xabber didn't. So who's lying?

You are, of course.

my guess would be neither, isn't it perfectly possible both of you are telling the truth?

Send me a video of monal receiving push notification.

surely you aren't saying Xabber works 100% of the time everywhere and has 0 bugs

Like after 10 minutes of being in background.

moparisthebest, of course xabber has lots of bugs

But calling it a mock-up opposing to monal, it's an insult

Andrew Nenakhov: Monal is using VoIP priority push. Works reliably for me even when I kill it

My managers suffered for years with monal

Andrew Nenakhov: my iPhone isn't charged and hidden somewhere in a bag

Andrew Nenakhov: Monal significantly improved in the last two or three months

Andrew Nenakhov: but calling me a liar won't make me create a demo video just for you, sorry.

Whatever

hug it out :D

I spent a few euros, but not worth the trouble. See it as sponsorship.

Andrew Nenakhov, means nothing coming from me of course, but I also tend to like running code first, specs later, just hoping you release specs soon and are open for feedback etc :D

^ Guus

> hm I thought that would be a video of a group chat. Left my laptop in the office,so not now. However, group chats can be experienced live on https://web.xabber.com/develop/ right now

moparisthebest, I'm very open for feedback

Ge0rG, well, I've installed monal just now, set up my account, waited till roster loads, pressed home button.

10 seconds later I send a message... And 2 minutes later it's still not delivered.

Andrew Nenakhov, sorry if I missed it but are the specs someplace public?

I'm not surprised. I never saw it work even remotely reliable. Not that pushes not at all work, sometimes am hour later it explodes and loads something. But ... Very unpredictable

Andrew Nenakhov: nobody will implement non-standard extension, just accept it

Andrew Nenakhov: Testflight or store? Do you have the push XEP active on your server.

moparisthebest, you see, it's in Google doc, can share it but it's all in Russian

Andrew Nenakhov: I told you that already btw 😁

Ge0rG, store, 0357 supported

zinid, that's no problem if he plans to standardize it, you should submit it as a XEP as soon as you can Andrew Nenakhov

also, MIX will be released with ejabberd 19.02 next week, so "nobody implemented it" won't hold anymore

zinid, non standard extensions can very well become one.of course xsf can still pursue that dream of mix or whatever.

Andrew Nenakhov: too late, mix will become a thing

Anyway our group chats has pretty decent compatibility with any legacy client

ejabberd and conversations will implement it and it is documented within the xsf

zinid, which clients work with ejabberd 19.02 MIX ?

moparisthebest: conversations branch

zinid, but nobody can use it because you don't know if everyone else's server supports it

(or client)

interesting, so then maybe it's no longer "no one implements it" that would be a step forward

surprising for a 4+ year old XEP

in a good way

Andrew Nenakhov: I suggested you to release a xep back then, you did not

Our group chat can be installed on any serer )

Problem is almost never the XEP itself in my expierience, its just that devs have no time for experiments

still can release a XEP, who knows, maybe it's far superior protocol-wise to MIX

zinid, it wasn't ready. Also my xep guy is fucking seriously ill for the last couple of months

I wonder why I cannot cite MattJ, seems like he is not joined 😁

Blame MUC :)

zinid: welcome to GC1.0

MattJ, "nobody can use it" appears exaggerated…

flow, hmm?

MattJ, your response to zinids annoucement that ejabberd will support MIX

I think it's fine (even good) for closed environments, custom clients, etc.

... i think he meant why you think that

I hope that I, using ejabberd, will be able to join a MIX at some other ejabberd once it is released and i've updated my ejabberd

:D

flow, you will, of course

But you'll be alone :)

Then I hope that I am not nobody :)

As soon as you want to invite someone from another server, either their server or their client (or one of their clients) won't support MIX

Right, but I am willing to continue with that experiment

does the other server need support?

honest question i didnt read all the XEPs yet

It sure would be nice if a MUC successor wouldn't require home-server support

I'm so glad I demanded full muc compatibility back then.

flow, please don't think that I'm not happy that people are experimenting, or implementing MIX

And maybe someone will come up with a solution that does not and that solution experiences high traction in means of implementations

flow, xabber gruop chat does not require home server support

I'm glad that something is finally coming of all the work that has been put into it

But it doesn't change this simple fact that it is badly designed if we ever want to see it widely adopted on the network

(within the next 5-10 years)

MattJ [22:15]: > But it doesn't change this simple fact that it is badly designed. There, I fixed it.

Andrew Nenakhov, that is great, but I was told that the spec is only in russian? That maybe hurts adoption a bit outside the native speakers

Ge0rG, I don't think it's all that bad, but it made different design choices to MUC

MattJ, and by badly designed you mean the design as whole or the fact that it requires home server support?

Ge0rG, if you're developing certain kinds of closed group chat, I think it's pretty neat to have... compared to needing to implement all the problems MUC has for those environments

MattJ: I think the proxy JID "feature" is adding huge complexity for a very small user group.

flow, mainly just the latter

Then that isn't bad design in my eyes. It is one of two possibly design choices with that regard, each with their pros and cons

flow, we just don't have time to translate it. And we planned to start translating once it's really final

I personally would make the same decission. But I would also encourage other approaches. Ultimately it is not the XEP, not the council but the ecosystem which determines the survivor

xmpp => just amazing

flow, please re-read what I wrote

flow, I didn't say it was "bad design." I said it was "badly designed if we ever want to see it widely adopted on the network"

I said in a later message that I think it is great for some use-cases

so we will have muc and mix and the xabber-stuff-which-is-the-best and some other upcoming alternatives, because mix is badly designed for wide adoption

MattJ, ah yes, ok, sorry, but I am also not sure about that. The implementation effort for MIX-PAM is manageable

at least that is my impression

flow, it's not just about implementation, it's about deployment

Ok, but you have to deploy anything that is new, how is MIX different?

Because many of the problems with MUC can be solved in backwards-compatible ways

which allows for progressively improving the experience, as users upgrade their clients, etc.

As it is right now, MUC will still be a fact of life for everyone until we reach some high percentage of MIX support in clients/servers in years to come

That is probably true, and I fully support the efforts put into MUC.

Yep, fully aggree (minus the timeframe)

and if everyone is using MUC daily, because MIX is unusable for general-purpose chats, there isn't much to motivate developers (assume client developers) to maintain both MUC and MIX in their clients

It could be 5+ years, could also be 2-3 years

Let me know how you plan to upgrade all those Pidgin users who still don't have MAM :)

which is a significant portion of the XMPP userbase, like it or not

apt remove pidgin

It doesn't work like that, but we all wish it did

I spoke to multiple people at FOSDEM who directly went from Pidgin+XMPP -> Matrix

because Pidgin was their experience of XMPP, and it was rubbish

No yes, of course, but I am not arguing against MUC and MUC+, but for being adventurous and progressive regarding a new persistent groupchat protocol

Pidgin was an extremely popular client, most people who used XMPP "back in the day" used it, and don't keep up with modern client development like everyone in this room does

which may or may not be MIX

This room is a bubble in the XMPP sphere, trust me, out there it's Pidgin all the way

(ok, maybe I exaggerate a little here :) )

No I trust you, I was also a pidgin user when I was in the computer room at my university

But you can't underestimate the deployed userbase. Except for group chats to be used only by XSF members, you're basically always going to have at least one occupant who isn't able to join a MIX room

for quite some time to come

Obviously the sooner we get the ball rolling with MIX, the sooner we can switch to it

That is why I hope for multi protcol rooms

Which is why I'm glad it now has implementations

Yes, multi-protocol rooms will help somewhat

But if you can access every MIX room via MUC, and MUC is already implemented and compatible with every server... why maintain two group chat implementations in every client?

It's a hard sell if you ask me

Well I am speculating that MIX(like) protocols provide better posibilities than MUC when it comes to hibernated mobile devices

Or for huge participant rooms where you often do not want pushes of every message, and only read the backlog when you open the window of the channel/room

For sure

But if over the next year we solve both those things for MUC...

That is the point, I am not sure if they are solvable, or solvable in I way I would like

There where a lot of improvements to MUC in the past years, but I could imagine that at some point there is just no more one could fix without either sacrificing backwards compatibility *or* really dirty hacks

you only need one android, one ios and one cross platform desktop client

MUC is based on flawed ideas and must be destroyed.

I could be wrong, that is why I say "let's go down both holes and see where we come out"

Xmpp does not have 2-5 years for this. Our ship is burning.

And is going to be overtaken by matrix or something else if nothing changes

Andrew Nenakhov: relax, xmpp is in zombie mode now just like IRC 😁

zombies cannot die

yeah also the more people adopt it the more clients are out there, if they get another server impl except the reference

they will suffer the same problems as xmpp

its easy to make something work when everyone uses the same server software

we know this since whatsapp

and the same version of the same client btw

zinid, to be fair there are more than one client

"MattJ> I spoke to multiple people at FOSDEM who directly went from Pidgin+XMPP -> Matrix" I wonder how the Matrix support for libpurple is

but yes probably most use the same

lovetox, they'll suffer, no doubt, but as a protocol that has already won, not as competotor

Or did they change clients entirely

In which case they could have just changed XMPP clients..

why Andrew Nenakhov ?

Guus, here is an actionable item for you, put money on pidgin XMPP support :P

> lovetox, they'll suffer, no doubt, but as a protocol that has already won, not as competotor even so, I don't see how this can bother us, xmpp has always been a marginal protocol, what has been changed?

pep.: I was thinking the same

yeah xmpp is already in a very small niche

If you manage to revive libpurple's XMPP support, then finally we'll all stop ranting about it, and people might be a bit happier

Because they have an image of a shiny new protocol that currently has no interoperability problems and can be seen as a replacement for stale old not hipster friendly xmpp

Andrew Nenakhov: so good luck to them

yeah its called hype

Niche status won't change if xmpp does not roll out something that appeases to a wider audience

you can also learn a new programming language every year

because they invent something new that beats all the other

Andrew Nenakhov: but it will not, I already ranted about this yesterday in another conference

but truth is the people that write C since 20 years will probably never out of a job

xmpp is for nerds and their relatives, we just need to accept this and go this line

haha i like it zinid :D

I don't see the problem here 😁

you cannot fight WhatsApp or Google, that's not our league

yeah me too, there are billions that use whatsapp

zinid, I plan to ruin your nerdy lair.

Andrew Nenakhov: let me know when you succeed

:) ok

lovetox, analogy between computer languages and social networks is not valid.

Andrew Nenakhov: xabber group chat does not require home server support, but uses some group server? is the group server (component) open source?

It will be open source, yes.

If anyone's interested we can ship test version tomorrow

It works with a central group chat server. We have ideas to make it distributed in the future, nothing seems to prevent that architecturally, but not all at once

distributed means matrix style?

I don't know matrix style exactly. )

stupid full replica copying

Thing is, our group chat has a fixed jid. If that server dies, .. it's unreachable via that jid

In theory we can make failover jids so clients can msg any of those but it's clumsy.

zinid, how do they address group chats in matrix? I had an impression they have centralized index (or indexes) to discover channels

Andrew Nenakhov: I am not sure, but looks like they just replicate room history, just like rsync

Replicate between what? Servers? Every server in federation has own chat history?

yes

something like that if I understand correctly

https://www.phoronix.com/scan.php?page=news_item&px=KDE-Matrix-Support

edhelas: so slow

When we were discussing architecture of our Groupchats, thought experiment ran into too many potential problems, especially when retracting messages, and syncing changes and accounting for possibility of s2s failure

:( sorry

Andrew Nenakhov: distributed replication is hard

especially when you try to invent it from scratch 😁

Yes. And not really necessary.

Andrew Nenakhov: modern jabber implements distributed replication exactly (using xmpp as a transport), so it's kinda necessary

Twitter survived 7% total downtime in 2009, any decent server admin can achieve just 5% downtime

not everyone realizes this

> Twitter survived 7% total downtime in 2009, any decent server admin can achieve just 5% downtime yeah, tell that to chatme.im 😁

is it alive yet btw? 🤔

Never heard of chatme.im

Andrew Nenakhov: it's from operators@ mail list 😁

Andrew Nenakhov: I use it as an example of xmpp fragility

zinid you have the link ?

found it

that's also why I'm moving jappix.com and fr.movim.eu :D they were hosted on the same server as chatme.im

the admin just turned of the server without any notices, so I just asked him to turn it on again for a few days to let me migrate things before it's definitely stopped

yeah, my point is that from common user™ perspective Twitter or WhatsApp is more stable

than any xmpp server

Current Twitter is extremely stable.

Failwhale years were loooong ago

Andrew Nenakhov: true, and it's possible to build extremely stable xmpp service, just it requires a lot of work

no free-time admins can do that

@xabber.org had two major downtimes in 1.5 years, both caused by Linode

Of course it is maintained by paid admin.

Andrew Nenakhov: sure, because you're hosted in the same place, no diversity

We have a backup xmpp server on digitalocean but so far sole DB server, can't make it failover because of reasonz

I understand those reasons 😁

I mean no blame, it's hard

)) there is one more silly reason you don't know but I'm to ashamed to tell about it in public

okay 👍