XSF Discussion - 2018-05-14

marmistrz, well, the XSF can’t and won’t give you legal advice on that matter

we will work on a guideline of things you should definitely look into, and a few non-legal opinions on things

but that’s in no way legal advice

Ge0rG, any news on your schedule w.r.t. GDPR meeting?

(tomorrow, 12:30 CEST)

Can't do tomorrow 12:30 (lunch break)

Ge0rG, winfried, pep., what would work for you instead?

any, not later on friday as usual

I've got some time after the 25th.

Sorry, not funny.

Ge0rG: any way we can keep you involved while not attending the meetings, or is your schedule too packed for that too?

(BTW, I do like the 25th pun)

winfried: my problem is that I'm fully booked on a project that just began today, so I need to be available on short notice. I can't schedule anything for the next days, except Friday and *maybe* Thursday.

jonasw pep. Ge0rG I propose we will meet tomorrow without Ge0rG and friday in full? (both 12:30 CEST)

It might work for me at 11:30 or 13:30 tomorrow, but I can't say for sure.

both 11:30 and 13:00 tomorrow wfm

But please don't nail me on either.

Just saying that it won't work at 1230 for sure.

Ge0rG: I understand, won't nail you for it

winfried, pep., I’m fully available except Thursday and late Friday

Except maybe with NIN.

so 1130 CEST would workforme tomorrow

or 1330 CEST, I think I’d prefer the latter slightly

Or is it NIИ?

Ge0rG: NIИ, that is a youth memory!

jonasw: would friday 1330 CEST work for you or is that to late?

winfried, that would work

jonasw: oops, meant to ask 12:30 on friday , but 13:30 would work for me too ;-)

winfried, 1330 works better than 1230 on friday, I think :)

pep.: would tomorrow at 13:30 CEST AND friday at 13:30 fit you? Ge0rG may then attend tomorrow and friday for sure (correct Ge0rG ?)

winfried: besides of the music, I also loved the Quake1 ammunition...

winfried: I can't promise it 100% yet, sorry.

Ge0rG: sorry, I am from the wulfenstein era ;-)

winfried: "Mein Leben!"

my new portable digital audio player unfortunately cannot run doom anymore

not enough colors (1 bit color depth)

https://op-co.de/vr3/files/doom.gif - I had doom running on my first Linux PDA.

that’s more than one bit ;-)

Last modified 2004-01-31 18:48. Wow, that's less than I thought it'd be.

(granted, people *did* run doom on black/white LCD TI Voyage 200 calculators by tricking the display into displaying 2 bit grayscale or something)

https://news.ycombinator.com/item?id=17063761

edhelas: there we go again...

ZZzzz...

> That said, I think that main problem with Jabber is lack of a single entity pushing it. So it'll naturally lose fight against corporation with billion-dollar marketing budgets. With Matrix it might be better.

:D

let's built a decentralized network, but handled by a single entity :-°

Wait the performance of the matrix server is so bad that had to disable presence?

Where can I find old versions of XEP-0045 (which define the GC1 behavior)

Is this for real 😂

maybe that's just me but I think they have a point, Arathorn is very active both on social media, and getting publicity (French government case), they operate like a corporation but Matrix is still federated (anyone can run the instance etc.)

catching a bit conversation on standard@ after a while, for the record I'm strongly willing/needing a deletion mechanism for file sharing (I'm not using HTTP upload, but would be nice to make it generic enough so it can be used by HTTP upload). I was planing to use ad-hoc commands for that, and propose a XEP after testing on the field.

What came first, Jabber Inc. or the JSF?

https://xmpp.org/extensions/attic/ only goes back to 1.16 :(

daniel you don't have performance problems anymore if you remove the heavy features

I mean honestly I want to like matrix. And I actually hope they will succeed. They do some (non technical) things a lot better than the jabber community or the xsf

But from all I hear their stack is just polished shit just as xmpp is

Marketing?

It'd be nice if they had learned from XMPP, but from what I can see, they're following the exact same path.

Yeah Marketing fundraising to employ full time people. Their way of extending the standard that you basically have to submit a PR for the spec, the implementation in the server and the client at the same time

I'm sure Jabber was just as cool back when there was one client, one server, transports were all the rage and Jabber Inc. wasn't bought by Cisco yet.

Just wait until there are more than one of independently developed servers and clients, it'll be exactly the same as XMPP.

Rumor has that they are actively discouraging new server implementations

Zash: that's one of the main issue I have with matrix (with their initial aggressivity with XMPP), it's all controlled by one company and a few people.

Wasn't the excuse for the reference server being slow "but we're rewriting it in rust!!!"

Go!

Samesame

Yeah their server story looks like a desaster.

But otoh, Jabber was started as an open source project, not as/by a company with a marketing budget.

> List of acquisitions by Cisco Systems (redirect from Jabber, Inc.) Heh

2019, Cisco is buying Matrix.inc

What about going through the IETF and having a (Software|Standards) Foundation?

I mean in a way it's interesting to see how their approach pans out in the next 5-10 years

If it's still a somewhat open standard in 10 years that you could federate with their approach might have been proven more effective

I don't think there is anything inherently bad with a single company providing the reference implementation. As long as you always have the *option* of forking while still federation

*federating

I think the real problem is the difference between an open chat application like Matrix and an ecosystem like XMPP. Getting started is *much* harder in the ecosystem, but in the long run you want an ecosystem not just one implementation.

winfried: Doesn't that depend somewhat on who "you" are, and what their goals are?

Matrix uniformity comes from it's relative obscurity. Once it will have multiple implementations, it'll be absolutely same as XMPP.

except without the specs/documentation ?

moparisthebest: you mean that https://matrix.org/docs/spec/ ?

And of course they are less obsessed with this crypto bullshit

Are they?

didn't they develop OLM themselves?

Zash: don't think so, when maintaining a system for a longer time then 5-7 years, you want an ecosystem, not just one application. Up to that point funding, acquiring technology/knowledge etc is easier when you use a monolithic code base. Once you are up to your first big overhaul, you get your regrets. No matter who you are and what your goal is....

What about some projects applying for MOSS funding

Xmpp projects

winfried: I mean like, a company wanting ultimate control, do they want an ecosystem? Or like, Signal.

Zash: If I want ultimate control, I can still use the XMPP ecosystem as codebase. In matter of fact, I expect most XMPP deployments are exactly like that!

.

Is it me or is the xmpp.org website slow?

I think it might be you

1.78s

loads fine here

MattJ: You've just gotten spoiled by how blazing fast the new prosody.im site is ;)

About 1.8s here for prosody.im’s home page. :p

Hmm, the link to compliance suites on https://xmpp.org/extensions/ is outdated

Links to XEP-0375 instead of XEP-0387

MattJ, PRs welcome!

I'm looking for the page

MattJ, https://github.com/xsf/xmpp.org/blob/master/content/pages/extensions.md

PR in

So, I'm still looking for JEP-0045, version 0.4

archive.org?

And no, the git log doesn't go back *this* far.

Zash: do you know the URL?

duno

explore

I'm not sure why but "jabber groupchat history" leads to completely unrelated results.

https://web.archive.org/web/20021004130423/http://www.jabber.org:80/jeps/jep-0045.html

That seems to be as far back as it goes

Zash: you are awesome, thanks!

I'm sure I've seen the "Groupchat 1.0" protocol somewhere too, but where?

Seems /attic didn't exist back then

Ge0rG, have you seen https://mail.jabber.org/pipermail/standards/2002-September/001485.html ?

(not entirely sure what your current mission is, but seems like it's related...)

MattJ: There are tickets here: https://github.com/xsf/xmpp.org/issues

Neustradamus: Too slow: https://github.com/xsf/xmpp.org/pull/436

-> for https://github.com/xsf/xmpp.org/issues/398

SamWhited: ^^

Any news for the XEP diff tool? https://github.com/xsf/xmpp.org/issues/412

Ge0rG: Hmm https://web.archive.org/web/20000617033408/http://docs.jabber.org:80/

MattJ: thanks, that's another useful pointer

In particular this message from that thread: https://mail.jabber.org/pipermail/standards/2002-September/001499.html

which I'm sure you'll print, stick to the wall and throw darts at

https://web.archive.org/web/20000823073740/http://docs.jabber.org:80/jpg/chgc.html Well what have we here

ohmygod mcversion="2" ✏

!praise Zash

I see that the redirection of docs.jabber.org is missing

https://web.archive.org/web/20000919185743/http://docs.jabber.org:80/jpg/x273.html examples!

The real spec!

so simple

Ge0rG: Looks like the origin of the direct invite namespace https://web.archive.org/web/20020214222508/http://docs.jabber.org:80/draft-proto/html/conferencing.html

<message type="groupchat" from="act3@gc.denmark"> <body>Hamlet has arrived.</body> </message>

interesting

a precedent for messages from the bare MUC JID :)

jonasw: yeah.

<iq type='get' to='roomname@server'> <enter xmlns='jabber:iq:conference'/> </iq>

jonasw that would be nice indeed, introducing each participant by all his familly name, relations and positions

> get > enter

jonasw, that's why most clients still support those messages (and it's handy)

"still"

Heh

for new clients it isn’t that obvious ;-)

*hint hint*

M-Link certainly sends messages from the room's bare JID.

Messages from a MUC bare JID are often used for CAPTCHAs and other entrance impediments.

So-called "System messages". I hadn't realised those were folklore at this point.

Ge0rG: https://web.archive.org/web/20020207105531/http://www.pipetree.com:80/jabber/gc.html

How many groupchat protocols are there even?

3 or 4 by 2002?

I think it was three wasn't it? gc, iq:gc and MUC.

But maybe I missed one.

"conferencing"?

Or is that what became MUC?

Zash, MattJ: thanks very much for your archeological support. Its result should hit standards@ any minute now.

oha

this will help skip the time waiting for the raspberry pi to update

> 33.7 kB/s 9min 42s

I can't see it yet.

me neither

but looking forward to it

May 14 16:21:24 bender postfix/smtp[9754]: EC5C814540E1: to=<standards@xmpp.org>, relay=atlas.jabber.org[208.68.163.215]:25, delay=3.1, delays=0.28/0.04/2.3/0.43, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as C40468FC)

,oO( Atlas Shrugged )

https://web.archive.org/web/20020602221836/http://docs.jabber.org:80/draft-proto/html/sxpm.html Hah, neat

Ge0rG, maybe it’s running an apt update on a raspberry pi and waits for the result before forwarding to introduce some random mixing delay

jonasw: that would be some interesting interdependency.

“16:09:10 jonasw> a precedent for messages from the bare MUC JID :)”, mu-conference used to do that if some entity joined with a GC1.0 join.

Ge0rG, ooh, the line above the last participant was that null participant, not a ncurses bug!

Thanks for making me notice that.

Too many bugs..

99 bugs in the code, take one down, patch it around, 103 bugs in the code

Zash: how is potato farming going

Link Mauve: haven't you been at the meeting where we discussed doing the avater thing?

Congress?

At*

Yes

pep.:

Zash:

daniel:

daniel, I wasn’t expecting that to break.

I was wrong.

Nasal Demons!

memberbot is online for our Q2-2018 voting

So, I want to use XMPP for my IoT. Who is the right person to give implementation advice?

depends on what IoT means

I’m doing IoT-ish things with XMPP

Means whatever you want it to mean

I want to control multiple LED lighting controllers, get them automatically discovered by OpenHab, and just have to give them names.

Ge0rG, the LED lighting controllers... are based on what?

Mine hang off an Arduino, and you won't be XMPP'ing from those (unless you use one of the fancier models with Linux on)

MattJ: esp8266

No sensible XMPPing on that either, although it runs Lua and I considered some very hacked port of a minimal subset of Prosody...

Once you factor into account that JIDs can be multiple KBs, there's no chance of a compliant implementation

Prosody already doesn't normally work with that.

True :)

Depending on who has the 3KB JID

Ge0rG, in my case most things happen over MQTT, and I bridge to that from XMPP/whatever needs to. It's the simplest option

MattJ: I'm not going to use the Lua runtime, and I don't need 3KB JIDs on my LAN

Then you can probably manage some form of XMPP, if you're willing to cut corners

https://xmpp.org/extensions/xep-0060.html#example-216 shouldn’t this be in a pubsub#owner rather than in a pubsub namespace?

I mean, jonasw should be able to give you some tips for XMPP in constrained environments :)