XSF Discussion - 2020-12-04

November was when the announcement was sent out: https://web.archive.org/web/20021230183123/http://www1.ietf.org/mail-archive/ietf-announce/Current/msg21379.html

Damn, you've nerd sniped me. It was before the drat anyways; here's a reference from Jan 2002: https://mail.jabber.org/pipermail/jdev/2002-January/069346.html

Sounds like it might have been an IETF term and "Jabber" was just one candidate protocol to be the IETFs "XMPP"

Oh no, definitely named within the Jabber community sometime before Jan 2002: https://web.archive.org/web/20020305041511/https://perl.jabber.org/logs/conference.jabber.org/foundation/2002-01-09.html

IETF had the IMPP and CMP(?) things

So that February draft is the one mentioned in that jded post as for archival purposes

> [13:32] <stpeter> ick, i hate XMPP :P

:o

Well, don't we all?

Hahaha. There's a discussion about what to do about inactive projects in listings in that chat log.

Well, we should come to a conclusion on that topic in another decade or two.

> [13:17] <Dave> Maybe we should have a "dead projects" section?

(Not me, in case anyone's confused)

I'm confused

There can be only one Dave!

Zash, Sadly not true.

https://sodocumentation.net/xmpp/topic/3036/xmpp-addresses-aka--jids--jabber-identifiers- nice documentation

I didn't found any PHP PRECIS libraries unfortunately

oh cool, they used my package as an example and I think used my graphs with some modifications, that makes me feel good

oh I see, this is taken from the old SO Documentation site before that was shut down, so definitely my graphs. Cool!

SO?

stackoverflow

So I guess that "they" didn't so much "use my graphs" as "I wrote that" :)

heh

hmm, CC BY-SA violation?

It's got a link to contributors at the bottom (and we're both on there). Took me a while to find.

Happy Friday everyone

ÄNTLIGEN

0. Welcome

Hi! Who do we have, and what's on your mind?

Present

What's this? Board meeting? On a Friday?

Zash: new round, new changes

Seve, dwd, MattJ

Hello

Hiya.

Dwd I think that is your turn to do the minutes this week

It seems we're at:

1. Minute taker

:-D

OK.

2. Appointment of Officers

I sent out an e-mail to both Alex and Peter.

Peter responded with "Yes, of course!", so I motion we re-appoint him has Treasurer for the 2020/2021 term.

+1

ralphm, received no mail. can you send it again?

+1

But I will volunteer again

+1

+1

found in SPAM 🙄

:-(

Motion carries. Thanks Peter!

I motion we appoint Alex as Secretary for the 2020/2021 term.

+1

+1

+1

arc?

+1

Motion carries. Thanks Alex!

dwd: please make that full names in the minutes

Yup.

3. Voting Process

This item came up last week. Go!

It appears as though the bylaws description of how we vote doesn't entirely match the memberbot implementation. Which is correct, and what do we do about it?

For Board/Council or Members?

(FWIW from the peanut gallery, I feel quite strongly that the two options are doing what the Bylaws say, or updating the Bylaws, I don't think continuing to not follow them should be an option we consider)

Possibly both, but certainly Board/Council was mentioned.

Kev: definitely, with a strong lean towards doing what the Bylaws say

From the peanut gallery: I say the opposite: if it ain't broke don't fix it. Just ignore the bylaws and do what we're doing today unless an actual change is needed.

SamWhited, that's not how things work usually.

¯\_(ツ)_/¯ who cares? I'm not trying to be silly, I really don't see the point of messing with it unless there's some actual reason to change things.

I am not interested in legal challenges, TBH

We have enough undocumented lore in the protocols.

Why would there be a legal challenge?

Anyways, just my opinion; don't make more work for yourselves.

Right, "fixing" the bylaws is easy enough.

I don't know, but this is not a book club

Ralph: You mean 'not interested in' as in "don't care that there's the posibility of" or "don't wish to see"?

Kev: the latter obviously

Ralph: For the sake of clarity :)

Having the bylaws match reality is a good thing, regardless of which part(s) need to be changed for them to match.

My point is that we are still a legal entity with rules and just ignoring them is not an option

SamWhited, So the problem is that people noted that the Council, for example, wasn't voted in in the way that the Bylaws say.

Ok, so let's discuss what we have and what we think it should be.

I think the sticking point is that currently we might see a Board or Council made up of individuals who aren't eligible to be on them, based on the number of votes they received.

SamWhited, So someone who stood for, but didn't get into, the Council could complain. Nobody has, but still.

Every member of Board and Council must have received a +1 from the majority of (voting) members.

So the bylaws appear to state that we select the people with the most votes, as long as they have more than 50% of the vote.

Whereas the memberbot only allows people to vote for 5 people.

In principle, the bylaws allow a member to vote for all the candidates or none of them.

I'm not sure the bylaws suggest how voting has to happen, do they?

Merely how B/C must be selected based on the votes cast.

They *imply* how it has to happen, for it to not then break in the face of many candidates.

So to follow the bylaws, for Council we should've had the possibility of casting 7 votes explicitly

IIRC the board was dedicing on how many seats there are for board and council. A long time ago it was 3, and then got increased to 5

Kev, I think the most obvious reading of the bylaws would lead to being able to vote more than 5 times.

but not sure where this is covered in the bylaws exactly

Alex, IIRC, it is the members who set that limit not the Board.

The Bylaws state that the membership decides the number of seats. I think we did that, long time ago.

dwd, possible as well, we would need to research the old member mails

Section 3.13 Voting Procedure for Election of Board and Council. Election of individuals to serve on the Board of Directors and on the XMPP Council shall proceed as follows. First, the number of individuals to serve on each body shall be limited beforehand by the Members as specified in Section 4.4 and Section 8.1 of these Bylaws for the Board and Council, respectively. Second, the Members shall vote on the candidates standing for election in accordance with Section 3.9 of these Bylaws. Third, the individuals elected shall be those receiving the highest percentage of votes cast, up to the limit set by the Members and with the proviso that no individual receiving less than a majority of votes cast shall be elected. Fourth, in the case of a tie for the final remaining position, the final individual shall be chosen in accordance with the procedures defined in “RFC 3797: Publicly Verifiable Nominations Committee (NomCom) Random Selection” published by the Internet Engineering Task Force.

the 50% rule definetly can be an issue when we have a huge numer of applicants, then in theory none of them could get 50% of the votes

Alex, Yes, but only if you restrict the number of votes. Otherwise the risk is more ties, but we have a solution for that.

Bylaws made for a larger organization than what we are now?

If it's approval voting which this seems to suggest you could get less than 50% no matter how many members there are

no matter how many positions, that is.

Zash: I don't think the XSF was much bigger when founded

SamWhited, That is true.

And we have saved basically the same size sense

Since

If we'd done the Council voting as we do for (re-)electing members, it would have been according to 3.13, I think.

As a concrete proposal, then: I motion we change the memberbot to handle Board/Council votes as a simple y/n for each candidate.

dwd: that's equivalent to what I wrote, right? In that case +1

ralphm, I think it is equivalent, yes.

+1

+1

+1 from me as well.

Motion carries.

FWIW, I cannot remember who raised this originally - I'll have to go look in the chatroom logs for this room.

Now the remaining question is, do we need to re-do or re-evaluate the way these elections were done?

<hat:member>Anyone considered a nominating committee?</hat:member>

(And from the peanut gallery, seems sensible to just change memberbot)

ralphm, Board is unaffected. Council might have been affected.

For board no one received under 50%, for Council only I did (and I do not contest the results, because that would be stupid).

SamWhited, hehe

Neither Council nor Board is affected.

Years ago Memberbot required exactly five votes for both board and Council. I'm the one that pointed out that error, and we continued to run the ongoing vote and just corrected it for the next year

The only people with less than 50% vote were not in the top 5.

Kev, Yes, but the ordering might have altered, I suppose. That would affect my seat most likely.

ahhh, that's fair.

So I would propose that we just say that we will apply this going forward

dwd: I don't believe the limit on the number of votes per person is actually a conflict with the Bylaws, it just leads to a problem with 50% votes.

I happy to keep the results as-is. If the membership disagrees, they have the ability to appeal.

Note that memberbot did allow for having less than 5 yes votes

(I'm wrong, 3.9 says that what memberbot does is wrong, but I still agree with just changing memberbot and moving on)

(Or at least arguably says that)

I am forced to argue that we should actually re-do the Council vote, despite it really not being in my self-interest to do so.

ralphm, Want to propose a motion one way or the other?

Dwd I would argue that that would break precident

I already made a motion

arc, Oh, so you did.

arc: indeed

Let's vote on arc's motion

At this point, I'm inclined to agree with Sam's "Don't make work" :)

+1

+1

arc, dwd?

(yes, you have to vote on your own motion)

+1

Ugh. Redoing the Council election would be a lot of work, but I feel it is the correct thing to do, so -1.

> I happy to keep the results as-is. If the membership disagrees, they have the ability to appeal. It would be fair, dwd, but I also agree with ralphm, membership can appeal.

dwd: can't you abstain due to conflict of interest? ;)

Ge0rG, If I were going to vote in favour, I would abstain for that reason.

Noted. Let the record clearly show that Board recognises that the Board and Council election hasn't been done in accordance with the XSF Bylaws (for multiple years) and that Memberbot shall be adjusted to allow everyone to cast a yes/no vote for each candidate, instead of a maximum of 5 yes votes. If the membership disagrees they may appeal Board's decision to keep the results nonetheless, noting that Dave voted against.

SamWhited, you have the right to speak up during Board meetings.

SamWhited, Oh, no - I *want* to do both.

Yes, but one should probably avoid unnecessary jokes that don't contribute to the meeting :)

With that:

4. AOB

We're over time and I have fries waiting for me.

Cold fries are dead fries, let's wrap

None here

5. Date of Next

+1W

6. Close

Thanks all!

Thank you all! Have a good weekend :)

And the minutes should be in your inboxes.

regarding that codepoint/byte discussion

obviously bytes would be retarded solution for python

but i recognize that the solution needs to fit all languages

all talk only about the one direction, finding a reference with start and end point

I'd just go with Unicode characters. Should be good enough.

but how woudl the other way around work in python

i have a string i know i want to quote 5 - 20

now i encode that to utf8

how do i find out start and end in the utf8 bytes

Loop? The answer to all computer questions is loop.

i could encode 0-4 and then get the length off that

start := str[0:5].encode("utf8").len end := str[0:20].encode("utf8").len ???

yeah

Counting UTF-8 sequences is easy if you have bytes too, using the magical construct called "loop".

Was there prior art from elsewhere in the beginning of this thread?

(or, the previous thread rather)

If you don't care about efficiency you can encode it then decode it. I assume it has a unicode package or something where you can loop and do it quicker as Zash said too. Been a long time since I've done Python, but something like this if you had a range of 0-6 bytes: "ⅧⅧⅧ".encode()[0:6].decode()

i dont see any point to make this as efficient as possible

its not like i have to resolve 10.000 references per second

of course i would want to operate on strings with python

If you select some text in your GUI framework, what data do you get?

because its natural and easy and needs no calculations at all

strings Zash

but this is a matter of the bindings

Some people work in different environments than you do.

In GTK you can generally operate on bytes or on some string like concept that they have that I don't know how it works internally that lets you access "characters". If you do operations on that they're all cached to make them fast enough to be okay on the UI thread.

it totally depends on the bindings you use

Sure, bindings might do extra work to adapt that to whatever language you're using. But in GTK itself it works as I described.

*a*̈

feels like a discussion with no correct solution

some will have more work with one solution, others less

and efficiency is no argument at all for me

Unicode code point array looks like the middle point to me.

we decode and encode so much to put it on the stream or read it or display it in the GUI

here and there a reference is a drop in the bucket

Another possible way of thinking about it (although I'm not yet sure if it matters or is helpful or not) is that a range in bytes corresponds to the XML where you have UTF-8. A range in Code points effectively corresponds to UTF-32, which is not what exists in the XML. I'd like the range to make sense in an XML document representing a session, not just after we've done some work at a higher level.

Something just gave me a headache. Good night.

Maybe that doesn't really matter, but semantically it makes more sense to me.

i have to say i dont care also as long as its doable

and works

There's one perfect solution, it's called "Bring back XHTML-IM"

Yah, that would be really nice<script>haxor.zash(123)</script>

<font background=white color=yellow>if only we had XHTML-IM</font>

:)

<span id="ref1">SamWhited</span>: Everyone else is using HTML (not XHTML) in JSON and get away with it. Surely we can too.

And surely in our XML environment we can get away with XHTML.

Are they getting away with it though, or is the web a mess and every random product is generally trivially easy to break?

Styles can be filtered out.

But jokes aside, I don't even want to use this for styling but generally feel we do need a way to select ranges of text

<span>

solved

I kind of hate the idea that you'd have to duplicate the body if you wanted fallback support, then you run into all the issues with clients putting different things in each body, but maybe you're right.

There'll be problems regardless of method.

Indeed

People showing 393 into HTML-passtrough Markdown libraries tells me that entire approach didn't solve the "XHTML-IM is insecure" problem.

I don't think that's solvable, that's the web being .. the web.

Has anyone done that? I mean, yah, that would be bad

Yes.

Someone did that.

But I really didn't think people would do that because it obviously wouldn't be 393. You'd just end up with markdown and like 2 of the styles would be the same and the rest would be different

Link?

"This looks like Markdown. I'll use this markdown library. Done! Ship it!"

Converse.js plugin IIRC

cool, I'll go look. I've seen their implementation and it wasn't using a library, but maybe there was an older one

Thanks

Probably mentioned in the logs here or council@

I don't think I've ever been able to find anything in chat logs ever, but I'll look :) THe earliest commit I see linked on the repo was a custom implementation anyways

Aha: https://github.com/igniterealtime/Pade/issues/79

That specifically says it's not XEP-0393, doesn't look like a mistaken implementation, looks like they just wanted markdown?

Note the edit

oh cool, I don't think I realized you could see history. "As much of XEP-0393 as possible" still makes me wonder, but yah, fair enough

Not sure if I've seen other cases of 393 vs markdown confusion, but I do remember that this one

"My use case will be markdown code pasted from GitHub or elsewhere" still sounds like they deliberately wanted markdown

The confusion was pointed out and a bunch of the comments have since been edited

Still. There was confusion at first.

Aaaaaaand then: https://github.com/markedjs/marked#usage

yah, you're right, found an edited comment where he specifically said that it was markdown but would *also* give XEP-0393 compatibility, looks like the author just messed with the history

Maybe 393 should have a red blinking comic sans warning :)

THIS IS NOT MARKDOWN!!!!!1eleven

Still, they noticed and it didn't get shipped, so 🤷

How many others might have made that mistake without me or someone being around to panic about it?

I dunno, I just assume they'd at least try styling once and realize the styles didn't actually work, but maybe that's optimistic for "the web"