XSF Discussion - 2021-01-29

Hello, I just read https://tigase.net/tigase-im-mix. Until now I didn't know that a public server was already offering MIX although XEP-0369 is still flagged as experimental. Will this year see the broad rollout of MIX supporting servers and clients or is that expected to be to early?

Maybe, I guess we'll see.

Zash, what's the trend on developers side in regards to the experimental status of the XEP? Is that hindering them or is the XEP so close to final that they can develop full steam ahead anyway?

Not hindered at all. Lots of Experimental XEPs deployed to production. You just have the risk of the XEP changing and everyone ending up on different versions, creating a fragmented mess.

Experimental mostly means that the author is pretty free to do whatever they want.

Zash, Ok, thanks for that explanation.

In the end, features users want tend to get implemented and deployed, regardless of the state of XEPs

Zash, thanks for explaining what I meant to lovetox :)

your words were better than mine

hi all, I have a quick question about MAM, xep doesn't seems to says the Fin element should have a queryid attribute, still I see that ejabberd set one. Anyone knows if it's ok to have one? (maybe I missed something in the xep)

paul, that's probably a leftover from when <fin> was carried in a message

aah ok make sense, (prosody is setting it to)

thanks

Yeah, an implementation that supports/-ed multiple versions of MAM returned the same <fin> payload in a message or iq depending on version.

Does anything depend on it?

correct me if I'm wrong, but currently, sending a fin in an iq with a queryid is against the xep? Or should the xep says it's allowed?

I think it's technically against the spec. But it's also easy to ignore

Yeah

I'm using the xmpp-rs lib, and parsing is very strict :(

Maybe it should be added to the spec 🙂

paul, a library should not reject unexpected attributes anyway

Really?

You should generally ignore what you don't understand.

Within know namespaces?

Holger, yeah, learnt that againts an ejabberd instance ;)

But we probably had this discussion 1.000 times.

it still sent the @code on <error/>, which isn’t in the recent RFCs anymore

but also cases like these with <fin/>… the choice is to only be compatible with 100% bugfree software, or ignore unknown stuff :)

You might consciously choose to cope with bugs, either with arbitrary bugs or (IMO better) just with known bugs.

right

But that's different from saying "I can send what I want because the spec tells ehe recipient to ignore crap".

jonas’, technically it's not an unknown atttribute but an entire elment

Daniel, huh?

ah sorry. never mind

I hadn't even heard of this issue until today, in all those years. Because everyone just ignores all crap :-)

We sorta added back error/@code to Prosody, the new error model library is shared with the HTTP stack. And MUC uses numeric codes.

Link Mauve: Absolutely.

Holger, that’s the very reason for xmpp-parsers to be that strict, to find and eventually fix issues in other implementations. :) ✏

There's a dead link to <http://www.xmpp.org/schemas/archive-management.xsd>, I guess that's just a placeholder for when the XEP is no longer Experimental?

> Zash: /me deletes it

thanks, that was quick \o/

And then I got complaints immediately from the CI 🙁

xD

Holger: https://github.com/xsf/xmpp.org/issues/413

Isn't that yet another story?

Board time

No way

Where did a week go?

Right?

paul: Oh I just committed (but not yet pushed) an ejabberd fix. Hadn't seen your PR. Yours breaks the mam:0 case where the queryid attribute must be included.

I'm glad to hear you're having enough fun to forget the time 😀

paul: Would you like to fix that, or would you rather avoid the CLA and/or the work anyway? :-) (I don't care at all.)

paul, Holger: Either of you got stats on mam:0 usage?

(Sorry this should be PMs I guess.)

Zash: Not me, sorry.

Or non-latest usage.

ejabberd supports :tmp because 'some customer' ;-)

There he is!

0. Welcome

Hi!

o/

Who do we have and do you bring any agenda items?

FOSDEM

dwd, MattJ ?

Here

Seve?

I second FOSDEM/SCAM

3 is a narrow quotum, but let's go

1. Minute taker

I forget who's up

dwd: ?

I don't think I've done it for a while, so I volunteer

yay

2. FOSDEM / SCAM

arc?

Actually last week we didn't do this one because we needed you

The only thing I can cover is Outreachy

Outreachy and GSoC should be on the agenda too

ok

Main question I have is: did someone apply for a FOSDEM "stand"?

So, on (virtual) FOSDEM, I didn't put in much effort. I guess primarily because I see most value in face-to-face events.

So I didn't apply for a "Stand"

Ok, that's a shame

There _is_ a RTC devroom

Yes, and mine is the only XMPP talk in it afaik

Right

(only XMPP mention in all of FOSDEM afaics)

I don't think we can just pause all engagement with the community until face-to-face is possible again

I do agree, I've just been busy with other stuff and wasn't sure if the FOSDEM stand thing would bring anything virtually

I would have contributed content, and I'm sure other projects would have also

But anyway, what's done is done, at least we know

Yup, dropped the ball.

Sorry

I think that means there is nothing else XSF-related for FOSDEM this year

SCAM can work on the summit (or not), I think other community members are pressing forward on that

Which is good to see

yeah, I saw that. Yay

The summit feels more important anyway

I agree there is not a pressing need to locate an online summit temporally with FOSDEM

But something needs to happen

What are you thinking of? A communications push?

Some kind of online replacement for the summit. Otherwise we'll basically have a year off making progress on various protocol matters.

I think the idea of focus groups would be good to organize

Maybe just as we do IRL, asynchronously gather topics on the wiki

and we assess interest, and schedule some times for each topic

I am kind of relying on the summit to gather sprint topics for March

But we could always build the agenda purely on the wiki and do a breakout over 6 weeks on topics

That makes sense to me. I think it would be good to coordinate with theTedd on his tales/talks effort

(as he's scheduled it for next week)

or indeed pick a later time

Sorry, I did not have Internet connection

Sure, we can work it out

I'm not aware of discussions in SCAM (Guus, nyco, Daniel), but let this be our notice of intent. Doesn't mean they need to do it.

Who'd like to take this on?

Agreed

I already have my hands full

Ok, not seeing a response, but we can do this async.

3. GSoC

I think we need to decide to do or don't do this, and we also require an admin for it.

Outreachy community sign ups are going on right now. I would be happy to be the admin for GSoC.

If no one else is interested

(No-one asked, but I don't have cycles to admin this year)

(But am always around to help if I can)

Outreachy differs in that it's not targeting college students, but women of all age. And in the United States, racial minorities traditionally underrepresented in the field

And we need to pay for the outreachy students

If I remember correctly, in 2020, we had flow and marvin as admin and co-admin.

Let's keep those topics separate

Ok

Last time there was a concern around the lowered rewards

We can apply starting today, until 19. Let's check if, besides arc and Kev, anyone wants to help out.

And we then discuss next week what we want to do in terms of payments.

Sounds good

I think we also need to punt on Outreachy in that regard, so let's have that as a second item next week.

We actually need to get the application for outreach today or tomorrow

4. AOB

Oh

Well, while I'd like to support any such effort, if we haven't even reached a decision on payments w.r.t. GSoC, I don't think it is realistic to do so today or tomorrow for Outreachy. Opinions?

I would suggest we start the process now, and discuss it further next week. Signing up is a multi-week process.

The final list of participating communities is published March 1st

After we are signed up as a community, we have to find funding. And that we can discuss next week. Signing up now does not actual participation

I'm ok with that

Because of the funding step it is not as simple as GSoC

Right

Seve, MattJ?

And having participated in gsoc in the past does make the process more straightforward. We don't have to be vexed as a foss organization.

Cool.

then, if there's no other business, I'd move to

Also just to throw this in, we don't necessarily have to fund every student we want. I can find other funding sources.

5. Date of Next

+1W

(arc: right)

6. Close

Thanks all!

Thanks ralphm and mattj

Thank you guys, and apologies from my side

fwiw, I'd be available to co-admin again.

Sorry, had to dash at :30, will write up and send minutes later

arc: Is there a TL;DR for Outreachy, that someone who knows GSoC could read?

https://www.outreachy.org/

It is basically GSoC without a single large corporate sponsor funding it, and focused on increasing diversity instead of solely focusing on college students

For most of the world, it is focused on women

Ta. But the basic premises are the same - OSS, definite deliverable rather than just 'internship' type stuff?

Yep

And prior participation in GSoC disqualifies candidates

Oh, looks like quite a number of the OSS XMPP projects wouldn't be eligible anyway, because you can't work on projects that have commercial versions as well.

(paraphrasing)

I guess Prosody doesn't have any features that are only available to people who pay, but Tigase, Ejabberd, Mongoose at least (Openfire and Spark in the past, but I don't think that's true these days) wouldn't be eligible. Swift we don't have an enterprise version yet, but it's possible we will, so probably shouldn't participate.

I guess the situation with Openfire/Spark is even a little bit more complicated, as there is no company behind it

e.g. what if someones writes an Openfire extension that is only available for money?

Extensions sure complicate it.

(The exact term is "There should be no difference in functionality between the free and paid versions of the open source project.")

basically that is also true for commercial FOSS projects where the company behind it makes everything free to use, but a third party creates extensions for money

So Conversations is eligible? Or not because of the push server?

Conversations is also interesting because of Quicksy, which has paid-only features, I think? (You can't use the directory unless you pay?) It does all sound a bit complicated compared to GSoC :)

Something as a Service

The directory isn't Conversations though. The software can be used with an alternative directory server completely unpaid

also the push server is open source, no? Don't see any issue for Conversations.

I think this is mostly about the spirit of the project. And Conversations surely is in the spirit of free software even if some hosted services require payments

Quicksy users don't pay anyway, only those who *don't* use Quicksy do :-)

> We do allow communities that provide paid hosting services.

arc, ralphm: Thanks arc for suggestion to sign up at least. I dont know how much I missed of previous discussions. However, I would like to recommend to call for any interested developers and projects in GSoC and at the same time review what the XSF can offer to make it happen. Then the discussion of whether it suits GSoC can start. XSF/XMPP people have mostly worked entirely via remote, we shouldnt let this pass away because of Corona if we actually have interested people. (But I agree that the situation is at least different due to mental loads/issues for many. I hope you got the general point.)

emus: GSoC have always been a remote program

Yes, thats what I meant

No actually I meant that counts for XMPP as well. We should make us of that and also that this is not the first time of GSoC, apart from the changes. And I also go with arc not to postpone things any further. If I can do something let me know.

so is GSoC still a thing or it's been replaced by this outreachy thing?

They're independent things.

GSoC's still happening (a little differently this year, but still happening).

ah thanks

The blog post transaltions are online: https://xmpp.org/2021/01/messagerie-instantanee-il-ne-sagit-pas-de-lapplication/

https://xmpp.org/2021/01/mesagerie-instantanee-nu-este-vorba-despre-aplicatie/

https://xmpp.org/2021/01/mensajeria-instantanea-no-se-trata-de-la-aplicacion/

https://xmpp.org/2021/01/instant-messaging-es-geht-nicht-um-die-app/

And here if one wants to share: https://fosstodon.org/web/accounts/180044

Are there any volunteers to post on Twitter?

emus, thanks for publishing all those translations!

could you merge my PR for article metadata by the way?

👍️ sure

I would, but I cannot technically review this, I need help from someone how can confirm this