XSF Discussion - 2014-01-27

I don't think we captured start times on http://wiki.xmpp.org/web/Summit_15#Agenda

ISTR discussing it in here (I vaguely remember talk of 9:30 for a 10am start Thursday, and 9:30ish Friday. Does anyone want to take ownership of that?

I said 10

As an executive decision, or suggestion?

If the former, fancy updating the wiki? :)

If the latter, let's pick 0930 for 1000 as a decision and goto 10.

Kev: suggestion. Feel free to set the time in the wiki, Mr. Chairman.

BASIC. Unsupported command. Try logo.

I have to admit I've largely forgotten Logo.

LEFT 180

FD 30, RT90

I remember what one did with it.

I just forget the syntax.

I had a German-language module without a manual.

http://www.calormen.com/jslogo/

Had to dive into hexdumps to discover command names

Maybe I wasn't the target audience.

So both LEFT and LT turn out to be legal. But Simon's comma is in error - no statement seperators.

So, we have a general decision to start at 10 both days? (Aiming for folks to get there for 9:30 to settle down?)

REPEAT 36 [ FD 10 LT 10 ]

Sounds comfortable. 5 of us flying from Munich will be there at about 9am or so.

Do we know that this is OK with Cisco for us turning up?

What did folks arrange there?

we'll camp at the airport and have breakfast or sumfing

I meant that a 9:30 start, rather than you particularly.

BTW, we've a new top dinner sponsor.

What nice folks.

And another sponsor confirmed! (Thanks ChatID!)

the MUC runners?

nice - thanks guys.

Simon, actually ChatID is in the process of switching over to MUC internally :)

Isn't that like IRC?

Except with presence, strong identity, extensibility and... standards? :)

10/10! (Just checking we have our marketing lines all practiced ready for FOSDEM.)

*giggle*

can i rant about nickchanges in MUC?

Go on

fippo: Was it going to be about 2 presences vs 1? :)

fippo: please tell us a storey about Psyc.

shouldn't MUC be multicast?

haha @simon. ok, MUC sucks completly :-)

simon: MUC is like IRC but lacks multicast!

waqas: it breaks jingle sessions for example.

fippo: Ha, yes

So a clients maintaining a fixed unique id for occupants works around this?

I'm more curious about your name on http://about.psyc.eu/XMPP :)

Oh, and +1 to Psyc stories

simon: see document history ;-)

Thursday evening. Drinks at XYZ bar and fippo explaining what the following means

Philipp Hancke (as author of 99% of psyceds net/jabber module): note that the following text is heavily biased, may contain inaccuracies and generally tries to persuade you that [[PSYC]] is so much better than XMPP.

fippo also popping up in random Austrian huts: https://plus.google.com/photos/+SimonTennant/albums/5973593611259914241/5973593611994742770?pid=5973593611994742770&oid=101190913792731232738

fippo: I could see something /like/ user nicknames in MUC. In fact, I think it would not be odd to see if we can improve upon MUC in several areas while maintaining backwards compatibility.

Means your server ends up keeping lots of directed presence markers around which are bogus.

I like Swift's approach to nickname changing.

If you can't make up your mind, leave the MUC and rejoin it with the right nick.

I haven't seen Swift's approach yet

waqas: I just described it. No nick changing :)

ralphm: s/nickname/display name/

Ha

kev: i like it too. maybe you can explain emcho how cool it is :-)

"swift's approach" is what Google Talk expected, and caused presence fails

In any case, I think that room nicks / display names should be orthogonal to presence

ralphm, I agree, but that'd be a departure from the existing spec.

MUC2!

ralphm, FWIW, if we did split nicks and presence, then multiple occupancy becomes trivial.

and we can add multicast! :-p

Uh-oh

mattj: we just have to agree on a definition of what that means before ;-)

MUC based on xep-33?

MUC based on pubsub, after we've made it turing complete. That way MUC doesn't need to be a separate piece of code, just something which runs on top of pubsub.

waqas, "after we've made it turing complete" - just ask ralphm about node as code :)

Eeuuuh?

We haven't standardized a language for node as code yet. You can't make a portable pubsub implementation without that.

XSLT with some storage feature built in? ^^

waqas: I'm not particularly inclined to redo on top of PubSub

dwd: I think you can do a new join protocol, triggering an alternate mode. You can also do display name changes based on IQs, and block nick changes for legacy clients, like section 7.6

ralphm: You aren't seeing the full picture. Once it's turning complete, we can write an equivalent of emscripten for it. Just think: TCP over pubsub, Linux over pubsub. Portable across all conforming XMPP servers. The microblogging folks can finally implement the whole thing by running statusnet over pubsub.

All XMPP problems will be solved forever

waqas: what's your favorite food?

Depends. Ice cream is generally good all the time though.

haha

bear, You'll also note that ChatID's amazingly great amount is one less than &Yet's, at least in binary...

dwd: I was asking MattJ if someone paying 110+101+100+11+10+1 would get 6 mentions

It is quite tempting to fill those, instead of actually looking for the amount we *need*...

that said, donating more than once is possible

yes, I did chuckle at the 111 amount

I think I will donate at the 101 amount

Yeah, it's tempting to fill that end in that way.

Anyone in for a Euro?

dwd: here!

unless you turn down requests from persons

Individuals have sponsored before iirc

Put me down for €10 if you like :)

Yes, I've been pushing back individuals, though.

Peoples' OCD might kick in to fill the gaps in the series

bear, You serious about the 101?

yes

I could use a €100, really.

101 or 100 - whichever is needed

But the new sponsors are listed...

Isode : €1001 &Yet : €1000 ChatID : €111 Dave 'dwd' Cridland : €110 Mike 'Bear' Taylor : €101 Matthew Wild : €10 Philipp 'Fippo' Hancke : €1

11 is still open

Oh, good point.

I can do the 11...

waqas, You're not even coming.

I can also do the 1 spot :)

I'll join over hangout/webrtc room if we have one :)

For the dinner?

waqas: i plan to bring equipment and setup something on the jitsi bridge with proper xmpp :-)

Nice :)

fippo - could you also do a demo of the Psyc XMPP bridge?

simon: the one i'm using right now?

Simon, do you have any idea what you're saying? :)

MattJ: I've read the wiki history.

and I've stuck in the snow in the mountains and need amusement.

simon: of course I can re-add the smart presence thing and show you how xmpp could have been alot more efficient a decade ago

well, almost :-p

fippo, PEP.

Thanks, winfried!

that makes my OCD so happy

one spot left *hint* *hint*

One?

what about 1010

1011

1100

???

Question is whether I could allow a non-binary one...

or other numbers…

Well, we can allow conversion to decimal...

10000 ... ;-)

10000 = 16 in base 10...

I should stress the amounts *are* decimal...

Or else we get no dinner.

oh, I think we all will put on the appropriate "filter" if anyone wants to donate any amount

i.e. making it a table with the appropriate base? ;-)

sure, "any amount" is acceptable

emcho, so...to FOSDEM will Jitsi be able to call into jitsi video bridge and call to webrtc clients?

tobias: not that I know of :-/

:/

what's missing, it's standard Jingle, no?

no... media level interop weirdness between chrome vp8 and jitsi vp8.

Tobias: we have a prototype working but I am not sure if we'll be demoing that on FOSDEM

ahh...ok

fippo, interop? is there more than one vp8 implementation?

Tobias: not really ... and it's not really only a matter of protocol

Tobias: not really ... and it's not only a matter of protocol

tobias: chrome does some... interesting things at rtp level

emcho, was that using Last Message Correction?

nah ... some other bug

fippo, ahh..ok. RTP level strangeness

emcho, ahh..ok

the biggest hurdle is VP8

(which also comes with RED and FEC

that's what we are currently implementing

we already had VP8 but it wasn't complete and didn't work with chrome

we do have a prototype now ... but not sure if it makes sense demoing

not that stable

ah..ok

i find it very interesting....slowly approaching a skype replacement that has more than feature equivalence

Tobias, We're getting very close to parity from Skype's perspective - one-on-one and small conf-calls - but we're also able to cover other ground too.

Tobias, I'm personally hoping we can tackle the GotoMeeting/Webex/etc kind of capability, too, within the year.

dwd, yup...especially with conferences in rooms with light control...kind of a pumped up skype conference similar to mumble but with better usability

dwd: we've had ~13 persons using th bridge last friday

fippo, Yeah, I think the bridge is a big part of this.

and the experience was very unlike https://www.youtube.com/watch?v=DYu_bGbZiiQ

Conf call IRL thing?

yeah.

fippo, :)

Lance, I like your push stuff.

thanks

Lance, And not entirely because you've managed to make "XEP-0198 Zombie" a term of art.

Not entirely. But almost.