XSF Discussion - 2017-03-14

efrit: a Slashdot "debate" is taking place around this, and the failure of XMPP

Slashdot still exists?

And it's full of dumb trolls. Nothing changed in the last 15 years

In Soviet Russia, trolls dumb you. ✏

Ge0rG: or paid trolls? I will not be surprised if it's some company's PR marketing

Fake news regarding our demise?

Aand someone linked to this https://m.imgur.com/nfivrKQ?r

as I said, someone needs to create a client which comes bundled with an XMPP server and hand-picked and automatically set up transports.

Do it, then talk about it.

Zash: it’s on my todo actually. at least IRC, because I have usecases for that myself. Prosody would be the bundled server ;-)

I believe libtransport is designed almost for this.

is that spectrum?

yep

uc: that was the original purpose of jabber, back in the day, to unite all different chat systems using bridges

my experience with spectrum is unfortunately rather bad

uc: I really can't imagine a company being dumb enough to pay for /. propaganda.

uc: OTOH, marketing is usually led by a huge amount of heroine, so who knows.

doesn't spectrum also bundle libpurple?

I still have nightmares about exposing an internet service that's linked against libpurple.

it depends on it, yes.

Ge0rG: but are trying reinvent the wheel lol

I think the multiline makes your client QUIT :with a nice message issue was also due to libpurple. Or maybe that was libcommuni and libpurple just had the /quit on the first line to quit your client with a nice message

issue.

Does it still depend on gtk & co?

but no point in bashing projects which at least try to do something better. especially if I haven’t taken the time to look into that multi-line thing

So, psa is stepping down. Now what?

Board choose a new ED. That's about it :)

what's the job of the ED?

(what he asked)

"Get stuff done". More or less.

Bylaws, section 6.6 has the details, but the ED's job is to ensure that the Board's resolutions are carried out, and signs any contracts.

so, what's the consensus of yesterday's "how much do we trust a remote MUC" debate?

At least I hope so. The alternative would be moxie haunting me.

According to some rapid grepping; Peter is an author of 59% of all XEPs. Perhapse we should consider minting the position fo "Executive Director Emeritus"

+1

yeah, I also feel that PSA might deserve some permanent honouary mention somewhere

Generally, things like that have happened when someone leaves the XSF, I believe.

In that case, let's make sure we never mint the position of "Executive Director Emeritus" :)

We do have Emeritus members.

Those should perhaps get some more exposure?

Kev, Yes, but they're not "normal" members - you can't be both.

dwd: See what I said earlier about 'when someone leaves the XSF' :)

https://xmpp.org/about/xmpp-standards-foundation.html does not list them, nor does https://xmpp.org/community/membership.html

Kev, Ah, yes. Missed that, sorry.

Who are our Emeritus members?

They should be listed on the website.

So I wonder if this was something lost in the Great Website Improvement.

it is why I asked - couldn't quickly identify them on the website

There is this awesome announcement for summit 19 though: https://xmpp.org/community/events.html

https://github.com/xsf/xmpp.org/issues/275

https://github.com/xsf/xmpp.org/issues/276

emeritus members are listed at the end here: https://xmpp.org/about/xsf/members

in the past they were proposed on the memberlist and we voted on them.

Ah, thanks Alex

should we, perhaps, dedicate a bit more of virtual space to them? Define why they're considered emeritus members, for instance?

Probably, yes.

moparisthebest: Possible future update for XEP-0368; I was adding a section to register the SRV services and realized that there's no actual example of what a full direct-TLS SRV record should look like. That would probably make things a lot easier on people creating one.

yea it could, I didn't know how much of the SRV spec I should put in there

I don't mind more examples or anything though

I don't think you need any of the SRV spec, just an example of what the thing you're describing actually looks like so I can copy/paste it into bind9 or wherever and call it a day and not read the spec :)

I started to add an SRV registration, but I was looking at the wrong thing (protocol registry, not service registry) and I don't think there's such a thing as registering a service, so maybe there's nothing to do here. I can't find anything that creates such a registry anyways.

Or rather, I think it's just part of the port number registry, but since we don't actually have port numbers for this it probably doesn't matter.

hm, it may matter

we should probably get a port for this then

or raise the issue with the people responsible for SRV, because having fixed ports allocated does not make sense in this context

I don't think there's an issue to raise; we don't have a port, so nothing to do here.

what I mean is: there are several use-cases for SRV records. you don’t want conflicts in the Service field. but you also don’t want to have a port allocation for each of these use cases, because it doesn’t make sense with SRV-native applications (as the port is in the record itself). I don’t think that is covered by the SRV-RFC (by looking at it quickly)

it’s not like we *have* to do anything here, but I wonder what the SRV peoples stance on this is

We have the protocol registered, so there won't be conflicts in the service field because we'd have to write a new XEP to create a duplicate service under the xmpp-client or xmpp-server protocol.

Glancing through RFC 2782, it seems to suggest that if there *is* a name in Assigned Numbers, that's the name to use, but otherwise there's nothing. So maybe nothing is needed.

Although https://tools.ietf.org/html/draft-gudmundsson-dns-srv-iana-registry-04 - but that seems to have stalled. Seems to be an individual effort - we should probably ask our friendly Area Director what to do.

we need a dedicated registrar who knows this stuff; I can barely keep just the editor things in my head

SamWhited, Agreed. The "XMPP Registrar" is meant to do this, and is currently sort-of the Editor - finding a body to do this would seem sensible.

Sam's suffering a little bit from us deciding Editor was too big a job for one person and needed a team...and then Sam ended up being (more or less) the only active person on that team.

Kev, Indeed.

I think it's mostly a matter of tooling and process; the jobs small enough for one person (or could be done by a number of people pretty easily without them getting in eachothers way), there's just a lot of shit to do that could be automated (which I will eventually get fed up enough to do)

SamWhited: Hysterical Raisins, partly.

In the olden days, stuff got pushed to master as a WIP.

(I have never heard that term; this is great, Kev ++)

SamWhited: which parts could use automation?

SamWhited: I got it from a colleague of mine, although I'm happy to take credit.

.oO(<http://catb.org/jargon/html/H/hysterical-reasons.html>)

jonasw: All of it. Any time I have to touch a script (sourcing a python virtualenv, generating the XEP, sending an email, sending an IPR thing, etc.) is a place where I could screw up, and none of it requires human intervention.

SamWhited: I am not familiar with the editor process, but I’m familiar with automating stuff

Oh reeaallly… be careful what you say out loud :)

*shrug*

I know why I said that :)

SamWhited, I can take on XMPP Registrar if nobody else will, I think. Although I'd appreciate borrowing jonasw's automation skillz if he's offering freely.

although I’m not sure what you’d need a python virtualenv for in the editing process.

jonasw: it uses some texml library that can only be found on an ancient sourceforge repo and doesn't exist in any packaged form that I can find

But the build process isn't something I want to touch personally; just automating it.

holy the heck wat.

http://getfo.org/texml/ is how we generate PDFs

yeah, figured from grep + wikipedia

jonasw: But yah, if you're actually interested in helping automate some stuff, a good one that I think could be done first and easily since it's sort of separate from the rest of the process is signing IPR releases. Just setting up a CLA bot or something so that the first time people make a PR on GitHub it makes them sign the release would be super helpful.

I forget to send it all the time, and I'm sure that's a bad thing for some reason.

what is CLA?

Contributor License Agreement

SamWhited: fun though, I think you owe me an IPR release then.

Yah, we don't have one of those, but the bots could be used for making people sign our intellectual property release just the same way

jonasw: Do I? Probably…

are you specifically talking about that thing here? https://clabot.github.io/

not specifically; there are a ton of them. No idea which ones are good or nto.

*not

SamWhited, so re https://github.com/xsf/xeps/pull/447 my only nit is it doesn't say it needs registered in the ALPN section anymore, there are a lot of TLS extensions on that page

moparisthebest: Oops; yah, not sure how to do that and make it an entity which can be linked elsewhere. Maybe it doesn't matter; the main thing I was trying to do is just not make the link the word "here" (these are printed too)

yea I don't like linked here's either

maybe just mention these ALPN Protocol IDs need registered then?

makes sense, I'll try to improve the wording. Or we can just ditch this PR, I only made it because I thought I needed to add an SRV section, but then I didn't so I just removed it and submitted the tweaks I'd done for readability

I like the rest of the tweaks