XSF Discussion - 2018-02-06

some people at this university want to establish some team chats, with the following requirements: - instant messaging - availability/presence status - video chat - desktop and mobile end devices they already have XMPP in their mind. I’m afraid of the "video chats" requirement. Any suggestions what I tell them at the 15th?

I doubt there will be any funding to be made available, given that Matrix et al. could possibly be adopted without funding.

video chat on all devices, I suppose?

I guess.

Jitsi?

sorry, jitsi is completely horrible as day-to-day IM client

(at least last time I checked, which was around 1.5 years ago)

edhelas, does movim already do voice/video via webRTC?

Openfire comes with a plugin that installs the Jitsi Meet suite locally

Guus, is there a client which can make use of that?

Jitsi Meet is a web client that does webrtc

oh, that’s a web thing

okay, that’s interesting

what’s the deployment experience with that? does it work well?

I wanted to set it up once, but it conflicted with my prosody.

Jitsi meet itself is pretty stable ( https://meet.jit.si ) - The Openfire installation of it has some quircks, but it's managable (notably, you _need_ a proper TLS certificate, and I'm lagging behind, version wise, a couple of months)

who’s developing that?

(jitsi-meet that is)

And the documentation wasn't telling me how it's supposed to work, only what commands I should run to "make it work"

The jitsi team, now acquired by Atlassian.

With Openfire, you can set it up in two minutes (literally).

that looks seriously neat

I'd actually love to get your feedback on the Openfire integration. It's primarily one guy and me that's building it, and still has rough edges.

I haven’t looked at openfire, sorry, only at the meet.jit.si thing

but that looks very polished

being acquired by atlassian, I’d be worried they might drop XMPP?

the only downside of jitsi meet is that while it uses a MUC, the chat is quite limited through the web interface

(imo)

Well, I've had a discussion with them about that. I believe that the are underutilizing XMPP, and could have a better integration, from a XMPP point of view. Their project lead pointed out that he was aware, but that their goal is not to build a strictly XMPP-optimized solution. Instead, they want to build a highly performant web-based video confernence solution.

XMPP is just a tool for them, not an end goal

the chat features demand of a web-video-conf is rather low

Guus, so they’re going to ditch XMPP as soon as it is more convenient.

the only way to make that less convenient is by having great interoperating clients.

jonasw, no - their setup is heavily integrated with XMPP

We really need modern and usable mobile and desktop clients

A dozen of half-clients don't count.

Ge0rG, pay me

but I'm not sure if they'd manage interop with other clients than their own, for the audio/video bit (most of it is Jingle based though, so it might actually work - and they're very happy to work with the community and add features/changes, as long as it's not conflicting with their interests of course.

Guus, okay

I heard rumors of a Jitsi-Meet-Integration with Spark, is that a thing yet?

jonasw: I offered you a job already, just not XMPP related ;) Maybe it does pay well enough to work half-time and to have the rest of the time for XMPP :P

but yeah, get yourself a host, disable the firewall for starters, install openfire, install a proper certificate, install the ofmeet and offocus plugins, done.

Ge0rG, I’m already just going to work 75% in my job; right now I’ve got masterthesis and other things on my plate though.

jonasw: almost, but it's not what you expect. We simply open the web view in a browser in Spark :)

Guus, that’s already a good start, if it provides a good flow from "we’re in a MUC, but we need a video/voicechat to discuss things now"

jonasw: personally, I can't afford to pay somebody to do the work I'd gladly do myself if I had time.

Ge0rG, heh

Ge0rG, point is, I think most a are painfully aware of the issues.

jonasw: it was even discussed at the summit, but I didn't get the conclusion, if there was one.

I don’t recall either

jonasw: maybe we should reduce fragmentation and focus our efforts. Stop writing JabberCat and help out making Gajim Great Again.

stop writing yaxim and implement video chat in conversations.

jonasw: yaxim came first :P

jonasw: I certainly would love your feedback on the Openfire / jitsi meet integration bit

give it a try if you've got like half an hour to play.

Guus, I’m still looking for an openfire-docker-thing

jonasw, surevine might have one, but there's no official one from us

we do have .deb, .rpm and .exe installers - or you can build from source.

I’m asking for docker specifically because I can do that in travis-ci without hassle

(and having it would make it easier to play around locally indeed)

I don't have one, sorry :)

regarding testing jitsi meet+openfire, if it requires a working TLS cert, I’m not sure how that’ll work in half an hour ;-)

letsencrypt/

?

but that requires me to do that on an internet-reachable server

(also, the cert requirement will be dropped again after we are able to fix an internal issue)

Guus, I do want to make existing desktop clients interop with Jitsi Meet.

For now just in one to one, so not with Jitsi Meet, but MUC video is part of a longer term plan.

jonasw: well, you can do a self-signed one, but then your browser will need to recognize it

does it need to do that beforehands, or can I simply accept it when it asks for it?

Link Mauve: in which client do you want to make that happen?

jonasw, also, if you disable all encryption in Openfire, it'll work - not something you want to do outside of a local setup, obviously.

ah okay

Ge0rG, probably Gajim first, since it already has good Gstreamer integration.

I’ll just do that I guess :>

Also to get up to speed with Gstreamer personally.

Link Mauve: I was frightened for a moment you could say "poezio"

:D

Link Mauve: that'd be great. As I said, we'll integrate the Meet web client in Spark, which is a start. Proper integration would be better!

Link Mauve, come over and do it in Jabbercat. We’ve got Qt and the chromium engine. WebRTC for free!

Ge0rG, libaa!

jonasw, the idea is exactly to add interoperability between desktop and web, not just web. :p

Link Mauve, that’s still desktop?

jonasw, as in, using jitsi-videobridge to make Gstreamer clients talk with WebRTC clients seamlessly.

can’t you talk webrtc from any client?

jonasw: libcaca please.

Ge0rG, does libcaca have a sweet demo like bb?

jonasw: dunno, but you can mplayer the matrix trailer with it

and there is an awesome visualizer called "cacaball". I'm not kidding. ✏

Ge0rG: what's wrong with mobile clients?

I agree desktop clients suck

zinid: Conversations is okay, minus voip/video. The others all suck.

zinid: especially on iOS

Ge0rG: ah, iOS, right

zinid: the situation on other platforms is even worse, but fortunately nobody expects that any more.

Ge0rG: you can always suppress your ego and join conversations team 😁

zinid: I'd use all my time to keep up with the changes in the code base and couldn't contribute anyway.

I really don't understand the need in yet another toy project, it was clear for me 15 years ago when I joined ejabberd project and it's even more clear now

zinid: I'm not sure what you are talking about.

Ge0rG: I want to say that it's better to join existing project

“10:12:28 jonasw> can’t you talk webrtc from any client?”, Gstreamer doesn’t support everything yet AFAIK, although I’m not fully sure what’s still missing.

Especially if it's superior

But jitsi-videobridge integration will be needed anyway for MUC video, so why not start with that.

> Ge0rG: I want to say that it's better to join existing project tell that to daniel :P

Ge0rG: so you're just fulfilling your ego and now cry why we don't have mobile clients?

zinid: I have contributed code to conversations, I have tested and reported a number of bugs on the iOS clients. I really don't know what you want from me now.

Ge0rG: VoIP stuff 😀

zinid: not gonna happen

Ge0rG: why? Ego again?

jonasw, Ge0rG, nowadays we have 256 colours available (some terminals even support 24-bit, but ncurses doesn’t), we don’t need terrible things like libcaca anymore. :p

have you guys tried the IOS client from Tigase? They showed it to me at FOSDEM briefly. Looked nice enough

See poezio’s gorgeous avatar display. ^^

Link Mauve: which doesn't work

(I don't have an iPhone, so I've not really played with it though)

Works for me.

Guus: IIRC I enrolled for the beta test a year ago, but then was disappointed that there is no bug tracker to report into

> I really don't understand the need in yet another toy project, it was clear for me 15 years ago when I joined ejabberd project and it's even more clear now This 'client a is not written in my favorite programming language / ui toolkit thus I really need to start my own from scratch' is one of the reasons we don't have a lot of good clients.

zinid: as opposed to you, I'm not driven by my ego :P I just know the insanity that awaits me, should I start writing VoIP/video code

daniel: but it's the reasons we have *a lot* of clients!

Ge0rG: ah, so you're just a pussy 😁

daniel, the reason for that though is that we only have free-time developers.

daniel: and out of all people, *you* shouldn't complain about client diversity :P

zinid: and you are a dick. So what's the problem, let's do it quick!

lol

Rammstein?

TIL the ejabberd-MUC has an avatar

zinid: the fact that you recognized that just added a sympathy point to your score.

damnit, zinid stole my sympathy point!

daniel: but in Android you have the same language and toolkit lol 😁

jonasw: we only have free-time CADT developers... https://www.jwz.org/doc/cadt.html

Ge0rG: what about funding? How much can we collect per month?

zinid: who is "we" and what do we want to collect for?

Ge0rG: we - the xmpp community

Collect for the salary

zinid: the XSF is collecting some amount of money. Last time I checked, there was strict opposition against collecting money to fund development of a single client.

zinid: it is easy to f*ck this up: assumed we have funding for one or two developers, how are we going to distribute that?

Ge0rG: why? We can leave donors to control the development depending on their money contribution, matrix does this

should we give the money to daniel because he's writing the best Android client, or maybe to me because it's good to have alternatives? To JC for improving the web client? To lovetox for Gajim? To somebody else?

zinid: the XSF is "vendor neutral", so it can't really solve this problem

Ge0rG: well, we should collect applications then donors will decide

We can create JSF as you suggested

zinid: but then again the JSF will be vendor-neutral as well and have the exact same problem!

Ge0rG: what problem?

zinid: with Matrix, there was THE client, and then they switched from THE matrix client to riot.im, by whatever mechanism they used to decide that riot is better now

zinid: the XSF is lacking volunteers even for the regular tasks we have. There is nobody left to run another Org

Ge0rG: actually they have the neat patreon policy: you can control the development if you contribute some decent amount

zinid: that's great. So now they are owned by this one big crypto money scam org?

Ge0rG: then we can just open patreon project and will control the development via bugtracker

So no org is required

depends if you want accountability or not

zinid: yes. Please create a patreon project, invite developers of different XMPP clients and let them define tasks and prices, then go out and collect funds

Ge0rG: what's wrong if the project will be aquired?

patreon might not be the right thing for this

but the general concept probably applies

Ge0rG: I'm actually asking here on purpose, if there are people wanting to contribute

zinid: to contribute money? I'm not sure you'll find anyone in here.

I’m not sure what tax implications it’d have. I’m also pretty sure that I’d have to run that by my employer, which feels to be too much hassle.

Ge0rG: here - in the XSF

jonasw: giving or taking?

Ge0rG: nobody, really?

I actually think that it's not a bad idea if the XSF helps individual people to sign up for Patreon or similar things

Guus: is it something the developers can't do on their own?

What, even 10$/month is a terrible burden for an XSF member?

Ge0rG, taking

giving is trivial

actually, https://converse.js lists both Patreon and LiberaPay on its website. Everyone can start with contributing to that.

zinid, if we shuffle funds around inside the XSF, I’m not sure that’ll help in any way

Ge0rG: devs _can_ do that on theirselves, but everything is hard if you don't know how to do it. Simple guides can go a long way.

jonasw: regarding taking: it's a secondary source of income that you need to income-tax. You need to make a tax declaration at the end of the year and pay something between 0 and 40% of it back

okay, I’ll have to make tax declarations starting next year anyways.

Guus: I've never interacted with patreon. Is there some kind of tag system where we can group all xmpp developers?

so that is not overhead which counts

but employer things would be annoying.

(not that annoying probably)

jonasw: IIUC you only need to tell your employer, not ask them.

yeah, probably

damn, I didn’t consider I’ll have to file taxes now anyways. I might as well give it a shot the.n

I would be curious which features users even want.

boom, I've just pledged to converse.js. Now you guys.

Is it "just work better" or "vidcalling" or "moar crypto"?

Guus: there are two problems with existing project: they cannot attract donors and the quality is mostly shit, sadly

Ge0rG: I don't know.

SaltyBones, all of the three

with "moar crypto" breaking the other two

zinid, you were telling people to pay developers to allow them to improve quality. Converse.JS lets you do that.

Ge0rG, we could make a page on xmpp.org where ways to support XMPP clients are listed

jonasw, https://www.youtube.com/watch?v=3_Af1qeouMc

;)

jonasw: I'd love to have that, but I suppose we need some open and egalitarian process to enroll developers, so that we don't violate our "vendor neutrality"

Ge0rG, add a field to the client list, boom done.

jonasw: I'd like to have another field in the client list: a short description

what other xmpp projects have Patreon-like pages?

and a logo

Ge0rG, do that

logo is a good idea actually

I'd be happy to spend a few bucks/month on several

we can base64 that and have the python stuff put it in an image directory automatically

Guus: still, the first problem remains. That's why I asked if XSF members can contribute, that will be a great start

Guus, conversations maybe?

Guus, you’re tempting me to set up liberapay just now ;)

jonasw: I buy conversations, the client.

ah

good :)

jonasw, seriously, go for it.

(I don’t, because I’m not sure shuffling around money between XMPP developers is going to be a good thing) ✏

hm...can you gift something to people on the app store?

SaltyBones: no :(

There are still a few people whom I would like to be able to reach with conversations...

bummer

i'd try to come up with actual business models instead of relying on donations

jonasw: the idea is that non-members also make use of it. :)

daniel: don't you have one already?

daniel, is there a way to have vouchers or something?

daniel, obviously. But every bit helps, right?

Ge0rG, yes. that's why i'm recommending that approach over donations

SaltyBones, sadly, the only way may be to wait for the next offer. I was lucky that C was free around the holidays.

Guus, Gajim has three options here: https://gajim.org/dev.php

jonasw, what I'm saying is, I bought it and I would buy it for other people to support daniel and get them to use it...

SaltyBones, I see

(my issue was that family doesn’t have any payment info associated with their google account)

Guus, sure. that wasn't a comment on you pledging towards converse.js. that was just a general statement on the 'how can we fund xmpp clients' discussion

What's wrong with donations? Even if we fail to init a project we can always give the money to some existing project depending on elections

when you buy Conversations on Google Play, Google will keep 30%, and Merkel will keep some ~30% as well. It's better to meet Daniel and spend him a beer ;)

except beer ;-)

a beverage.

Maybe I'll try getting a play store gift card and see if that works...

Ge0rG, i love beverages. but beverages don't pay for rent

at least i never tried that…

if anyone is interested, maybe create a wiki page that links to all donation pages of XMPP projects?

Since it's supposed to have a lot of donors, the elections supposed to be fair enough

daniel: depends on your landlord. If you can make them drunk every time they want your money... :P

Daniel: depends on the land...whathesaid

Guus: that's a sensible approach as well. Maybe we should ask the Board whether adding links to donation pages would be appropriate for the client/server software lists?

Ge0rG: wouldn't be the worst of ideas

Guus: the point is to find a single important task and donate, that is iOS or desktop client, and if we just list all the projects this will not accumulate a lot of money within a particular project

zinid: so who is going to choose which client gets featured, then?

zinid, i feel like at least some of the client developers wouldn't even want to work full time on their projects

zinid, you're right - but it does help, and could be a first step that we could do _today_. We can always do different things too.

Ge0rG: polling

Ge0rG, could you create a PR with logos?

if I’m gonna work full-time on a client, it needs to be a safe employment; I couldn’t quit my job only to find that two months later, the client doesn’t pay enough anymore.

daniel: I think there will be some who will 😁

Ge0rG, SaltyBones, if you do a PR with the logos as base64 inside the json or as files in an appropriate structure below content and refereneces to the file names in the JSON, I’ll take care of the rest of the templating/python magic

jonasw: yeah, that's the problem

jonasw: SVG or PNG? What resolution(s)?

zinid: never trust a poll you didn't manipulate yourself!

creating a business is always risky. i think i spend like at least two years working full time on Conversations before i made any significant amount of money from it

daniel: if there is not secret, how do you make most of your money ? With support ?

daniel, convince some of your clients that they really want vidcalling ;)

Ge0rG: wut?

or maybe jonasw tell your university to hire daniel to add it! ;)

zinid: see, there is money involved. People will try to cheat your poll so they end up receiving the money.

Ge0rG: the elections will be forged? 😀

Ge0rG, whatever you find. SVG is fine, PNG would be safer though.

zinid: you know how it works from your state elections :P

Ge0rG: how is that when only donors can poll?

SaltyBones, I doubt they’ll want to pay a lot for that communication thing. otherwise I’ll be recommending daniel as consultant for sure.

jonasw, try to convince them!

zinid: So what you propose is that instead of just funding what they want, people are supposed to make a collective, put together significant money and then have a majority decide what that money will be spent on?

SaltyBones, hah...

It's usually not their money..

jonasw: you just need to find some source of money and redirect accordingly.

Ge0rG: majority? Everyone can poll

zinid: you are contradicting yourself.

Ge0rG: amongst donors

Ge0rG: you just don't get it

zinid: indeed, I'm not getting it.

SaltyBones, i think at this point any money is better invested in a decent desktop client

Ge0rG: every donor can poll on the project selection, nobody else

because i'm actually doing ok and Conversations is good enough for now

daniel: still iOS left

And probably even more important imho

Daniel, kudos for that. Not everyone is gracious enough to say "i've got enough, give it to others".

daniel: you could start Conversations/Web now. Or Conversations/W32!

zinid, probably. in any case *not Conversations*. is my point

daniel: and not ejabberd, so what? 😀

Guus: I have also arrived at a point where it's not reasonable for me to accept funding, unless it's either at my employer's business consulting rate ($$$) or if it's on the order sufficient to buy a yacht.

Ge0rG: that be really stupid. If anything I'd contribute to dino or converse or what ever

daniel: but the network effect!!1

daniel, I contribute to these two, with time though, not money.

If only we could pool time as easily as we do money.

I contribute too! By yelling at the developers in their ticket systems!

Link Mauve: I suspect that many projects prefer time over money. I'd love to have devs for Spark, for instance - that badly needs a lot of work.

Guus: the challenge is to find somebody who can actually reduce your workload, not increase it.

Ge0rG: indeed. Which is why I like GSoC, but that's a future investment (and somewhat risky, if you ignore the 'its-also-fun-to-do' bit)

my third-party perspective on GSoC is that it rarely turns out really useful for the involved projects; is that false?

You know what - I think I'm going to blog a call-for-help for Spark devs. Most of our users are corporate. There's bound to be a few that have Devs

jonasw: I have one year experience. Two of the two students that I was working with last year were at the summit and fosdem.

but was the code useful?

yes, but that's not the point.

but that was my question :)

(in response to what Ge0rG said about "not increasing your workload")

The code that a student produces during GSoC can be created by a mentor in the same time it takes to mentor the student.

the point is that you attract long term contributors through GSoC

jonasw: GSoC is certainly a risky investment, in that respect: you have to first invest a lot of your time, in order to perhaps get a good deal in the long term.

Guus: we have never head a long-term contributors, they just disappear after the merge

I'm very skeptical for GSoC

zinid, that's exactly why I mentioned that both students are still very active with us. One of them is even in this MUC, both are still submitting PRs to our projects.

Guus: lucky you 😁

but yes, it is somewhat of a gamble.

(but also great fun anyways)

But keeping people engaged is also something that we need to learn, as the XSF, or as individual projects.

you can't simply expect people to stick around.

It's why I like that the XSF invited these students to come to the summit and fosdem - I think that that's a good way to further bond, in a sense

simply showing appreciation (not just to students, but to everyone in the community), goes a long way towards keeping people engaged, active, and happy.

(and it need not cost anything)

*zZZzzZZZzzzZ*

Guus, I fully agree with that. :)

Guus, agreed (with the engegaedness thing, not the sleep thing)

SaltyBones, https://www.ietf.org/mailman/listinfo/mls

zinid: > we have never head a long-term contributors, they just disappear after the merge Didn't Badlop get involved via GSoC, originally?

I think so

Holger, he was involved, but not via GSoC

he did GSoC later, being full-time contributor

so I'm not sure this is counted as "involved via GSoC"

zinid: Ah, I see.

Tobias, yep, thx, saw it!

is that list sufficient to stay in the loop for that WG?

probably

who knows if they are building more 'escape velocity' somewhere ;)

escape velocity?

jonasw, https://www.ietf.org/mail-archive/web/cfrg/current/msg09467.html :)

I see

hah, funny typo in the abstract of the draft (<https://tools.ietf.org/html/draft-omara-mls-architecture-00>): > This document describes the architecture and requirements for the Messaging Layer Security (MLS) protocol. […] It is eavesdropping, tampering, and message forgery.

OR WAS IT A TYPO

Good morning. You have been talkative here.

hi Zash

which one is the active one, the omara one, or https://tools.ietf.org/html/draft-barnes-mls-protocol-00 ? or are they even different?

moparisthebest, it appears one tries to draw a high level architecture picture (omara), while the other specifies the protocol in detail

is that the normal process for RFCs ?

moparisthebest, don't believe that there is any restriction for authors to split or not split their documents

moparisthebest: Not uncommon