XSF Discussion - 2017-03-08

https://bloggeek.me/webrtc-cpaas-report-update/ -- "Vidyo.io makes use of XMPP for its signaling". I knew jonathan lennox is smart (-:

fippo: I can't find the quote in that site

it's all fake news

ahh, it's in https://bloggeek.me/download/11252/

any urn:http:upload:0 implementations out there? The prosody module seems to be stuck at the non :0 version

Flow: unfortunately no

hmm, guess i've to introduce a compat layer in smack then

Or patch a several

*server

hey, look! https://wiki.xmpp.org/web/Meetups ;-)

There's an upload:0 now?

nyco: the SCAM has started?

ralphm: Ping; can you make an editors board in the XSF team trello and possibly add me to the XSF team so the boards stop showing up in my personal boards?

is it board meeting time once again? :D

-certinfo conference.prosody.im

Zash: conference.prosody.im has a valid certificate with a 4096-bit RSA key and a RSA-SHA256 signature that expired 6 hours and 57 minutes ago issued by Let's Encrypt Authority X3

onoez

#shitfuck

Language?

right

I should probably explain that’s a quote and deliberately used as exaggeration, because it came from a student who came into a room we were learning in, glanced at us, and basically shouted that and left.

no idea what was going on there

(and probably inappropriate in contexts where that story isn’t known)

Ge0rG: i think you enjoy the board meetings too much

arc: sorry, I'll try to be more negative next time :D

I may not make the meeting today.

aww... :((

im happy to chair if you need, assuming we can get quorum

Tx

hey Alex

Martin: MattJ: nyco: you here?

Here

sir, yes sir

gavel?

Here

ralphm: are you here and able to chair?

Ok roll call, we're all here but ralph.

dwd are you able to take notes today?

he's not present. can anyone else take notes today? we'll try to keep this brief in ralph's absence

ok throwing the notepad at random people, jonasw can you take notes today?

can do

thanks

give me a sec

anything special I should know?

otherwise I’m ready

nope, just minutes. there's plenty of examples on the mailing list

minutes, or hours, or seconds...

Topics for Decisions - is there anything to be done fro Google Summer of Code today?

/ms stops here

first, please, who’s present?

all except Ralph

cause I still don’t have a reliable mapping nick -> realname yet

jonasw: myself, martin, mattj, nyco

noted

Martin Hewitt, Matthew Wild, Nicolas Verité, Arc Riley

done, thanks

ok, Kev: is there anything of note with GSoC? are the students discussing things with projects? Mentors signing up?

I looked into making editor tasks as Dave or somebody suggested last week

I couldn't come up with any that I actually thought would fit GSoC or keep a student busy for the summer

arc: I've not asked mentors to sign up yet, I need to clarify with the Google folks if that's really necessary. I typically ask them to sign up only once there are suitable project ideas coming in.

(e.g. I wouldn't bother making Swift devs sign up if there are no Swift projects).

We've had a couple of students start approaching us, yeah. Although if Board can think of ways of widening the net here, that'd be grand.

there's no harm tho either. if they volunteer, eg they're talking to students who are interested, they'll still get a tshirt etc

Fair.

yah, I got a code-in tshirt from arc for showing up and saying hi in the room a handful of times :)

in the past if I wanted to widen the net I'd toss some cash at google advertising or put up a youtube video

but it sounds like the program is on-track for this stage

is that all for gsoc this week? no decisions necessary?

well, make some noise?

As long as Board are still happy to delegate to me, I'm still happy to do stuff.

on our blogs, social network accounts, etc.

I am happy

nyco: would you like to take that on?

there's certainly some college students out there who'd love the cash to work on xmpp related work.

I am only the master of the places I control, but still I can ask people to do, even provide some content as example/templates

my "audience" is not students though

fantastic. remember any non-mentor can participate in GSoC, so long as they're enrolled for at least one college credit for the current or next season, undergrad or grad. we can also reach out to our own community

ok, that seems wrapped. moving on

IEEE

I met with William for about 5 hours last week. He's an XMPP evangelist who's been working to get XMPP standardized for IoT for awhile now, but only orbiting the XSF

cool

Eg he's met PSA, Peter Waher, etc

Alex has him setup with a wiki account and he's applying for membership, and said he's going to pull some of the 100+ members of his IEEE IoT WG into the XSF as well, but it raised an issue that we may need a greater presence within the IEEE

SamWhited: you commented on the card, share to speak to this as part of the meeting?

I don't really have anything other than what's on the card; Dave suggested we appoint a council liaison, but there weren't many people at the council meeting today so I'm going to send out an email to council@ and see if we get any volunteers.

for context, what card?

on Trello, https://trello.com/b/Dn6IQOu0/board-meetings

Presumably we'd work with the IoT SIG to come up with a strategy for improving the current XEPs, or creating new ones.

I will advise the board when/if this happens, but I'm not confident that anyone will volunteer (they didn't last time this was mentioned when the IoT SIG was forming)

its possible that some of the new members we can draw from the IEEE side will volunteer

I deferred a lot of his queries about the XEPs to the IoT WG during our meeting, but one thing he reiterated was there's an existing IEEE IoT XML format, which is why XMPP is getting so much traction.

ok it feels that we're dying off, AOB for today?

yes

meetups sync

https://trello.com/c/PyQpztdh/262-meetups-sync

copy-paste:

* Starting point? https://wiki.xmpp.org/web/Meetups * Collect slides and content * Share and cross-reference

agree? disagree? comments? suggestions?

what is "AOB" for a control code?

Any Other Business

ah!

I'm fine with it. I'll note tho that at least one of those listed is a dead meetup

yeah, lots of jargon, sorry

oh

which one?

the SF meetup

oh

what is meetup sync about?

nothing wrong with listing it.

I talked to Justin, he liked that here

maybe we can resurrect it?

Leave it, or leave it with a note on the page

yep, good idea

Better to have it linked to than not visible

done

well there's a bay area realtime meetup that was intended to have some xmpp content

https://www.meetup.com/Silicon-Valley-Realtime/

it more or less took the place the xmpp meetup used to have, but with wider scope

so, call to all: the slides or any other content, let's share that! increase our visibility, let people talk about it

That should include any videos/slides from FOSDEM

i think its a great idea, have a centralized place for people to drop links to their slides and to the ever-shifting local meetings

ah good point... we generalise it to ALL events?

yea why not. does meetup still have their API?

or are you editing by hand

dunno

a calendar of XMPP events worldwide could have some limited utility

goooood!

still a reference, a list, generate SEO

or by having a list of slides, it would double as a list of xmpp _speakers_ which could serve conference organizers trying to reach out

yes

are you volunteering nyco?

why not?

jonasw, I count on your notes! ;-)

alright, AOB?

nyco: you live dangerous :-)

we're at time. +1W?

+1

+1

Works for me

alright, thanks everyone. Next week, same place, same time.

thx all!

nyco: quick question: regarding the GSoC-social-media-visibility point, am I correct to say: "Nicolas is happy to make some posts and spread the word, providing examples and templates."

?

oh... huh... yeah... :'(

just making sure for the minutes

i can also help, im just not volunteering for much given that we're in the middle of a 5000km+ move across the US

and I'm going to be doing a lot to organize PyCon activities once there

ralph and dave suggested we get US-side beanbags/etc for US conferences, and step up our game on this side of the pond

+1

I absolutely volunteer to "store" one of the beanbags and ship them wherever they need to go…

I am central, which means that makes sense somehow, I'm sure.

minutes are sent

we'll have to work out those details soonish. I'll be in Portland so i can gather materials before and after PyCon

jonasw, thx !

our booth last year was pathetic

need to ru, bye all!

pycon is on the wrong side of the ocean for me, sorry

SamWhited: are you coming to pycon this year?

arc: No, probably not

good notes jonasw

SamWhited: just to note, we have one free pass still available

the booth comes with 2, im using one.

thanks arc

wow, ok so there's some incompatabilities between exi implementations. which isnt suprising

it looks like most of these are simply incomplete

which isn't necessarily bad, so long as they're not used for server implementations.

wow, ok.

“14:54:34 daniel> the reason i'm asking is because ejabberd now supports merging and parting multi sessions nicks. and Conversations behaves pretty badly when you actually try it”, IIRC when I tried with poezio and Gajim with Prosody’s trunk code months ago, these two clients were behaving correctly.

It’s also the way biboumi does IRC-side nick changes.

I’d guess this is an Ejabberd issue more than a client issue, like it wouldn’t put the correct status code or something.

I was referring to PSA's resignation

Link Mauve: yeah maybe

I didn't actually boil down the exact problem

Have you tested Prosody trunk yet?

Or biboumi?

From what I've seen discussed, it looked like an ejabberd issue to me as well. Not creating a join presence for the other client when splitting, or some such

"""Otherwise, determine the set of characters for each immediate pattern facet of the target datatype definition according to section E Deriving Set of Characters from XML Schema Regular Expressions. Then, compute the restricted set of characters for the string value as the union of all the sets of characters computed in the previous step. If the resulting set of characters contains less than 256 characters and contains only BMP characters, the string value has a restricted character set and each character is represented using an n-bit Unsigned Integer (see 7.1.9 n-bit Unsigned Integer), where n is ⌈ log2(N + 1) ⌉ and N is the number of characters in the restricted character set."""

"""The characters in the restricted character set are sorted by Unicode [UNICODE] code point and represented by integer values in the range (0 ... N−1) according to their ordinal position in the set. Characters that are not in this set are represented by the n-bit Unsigned Integer N followed by the Unicode code point of the character represented as an Unsigned Integer."""

ok so it appears to operate as follows: you can construct an optimized character set of 1-254 characters. no more than 8 bits per character, which simplifies implementation considerably since you need a simple uint[256] for a map.

there is a reserved top value, if that value is passed, then the full unsigned int of the unicode codepoint follows.

"""as the union of all the sets of characters computed in the previous step"""

so while the XML regex format allows for multi-character sequences, this does not. all the valid characters from the regex are combined into a set, then sorted by unicode codepoint.

if fewer than 8 bits is needed, then that many will be used. but without multi-character sequences theres no "compression" available at this step beyond the obvious one of using a charmap

but it does allow some cool effects; Base64 might be encoded using 63 (not 64) values, accepting that "/" and the terminating "=" will use 5 bytes, so while more complicated than it really should need to be it would only add 6.4% to the size vs 20% as is used with utf-8 representations of Base64..

or if 3.6% difference doesnt matter, 7 bits could be used, still reducing the 20% overhead in half.

nyco: someone in the gsoc channel is interested in mix for mongoose xmpp:gsoc@muc.xmpp.org?join

Thx!

yup that looks like the end result; EXI strings can be either all UTF-32, or you can specify an optimized character set of up to 254 glyphs formed out of unicode glyphs from the BMP plane (0x0000 to 0x9999). implementation-wise, its a simple ushort[n] table to map because theres a 1-to-1 value for mapping the stream value to a unicode character. any character not in that mapping can be expressed by giving the highest value +1, followed by a UTF32 character.

the XML Schema containing XML Regex can and should still be supplied for typed values like Base64, SVG paths, etc. and the character set will be derived from all the characters possible. provided all the characters in that regex are in the BMP plane, and there's fewer than 255 of them, it'll use one byte per value for most values

for 126 or fewer, less than 8 bits per character will be used for bitpacked streams

but it doesn't validate, and it doesn't constrain what the stream can contain - the encoder will just plop down a max+1 value followed by UTF-32 for any character not in the mapping.