XSF Discussion - 2019-07-08


  1. rion

    according to "Council Minutes" emails it seems it's hardly possible xsf will accept any significnt changes to jingle s5b. So I think what if I propose another s5b based protocol under different namespace, will it have better chances to be accepted?

  2. jonas’

    unlikely

  3. jonas’

    that’d probably fall under duplication

  4. rion

    then probably I have to hurry up with sctp+dtls transport implementation while it's still "Deferred"

  5. jonas’

    rion, to be clear, I don’t like this situation at all myself

  6. jonas’

    I’m not sure what to do about it though

  7. jonas’

    The list is often rather silent on Jingle-related issues, presumably because of its complexity. So there are a few choices: (1) council trusts the submitters of patches and allows them through, assuming that those who patch jingle are those who are most familiar with it because they implement it; (2) we replace jingle by something simpler, causing yet another rift in the ecosystem; (3) we stall jingle until we find some trusted reviewers for the patches or a similar mechanism.

  8. jonas’

    currently we seem to be kind of passively ending up in situation (3)

  9. Daniel

    jonas’: I've observed the same pattern with any of the more complex / less implemented xeps

  10. Daniel

    More in-depth questions for example regarding pubsub seemingly hit the same wall of silance

  11. jonas’

    and I find pubsub not even *that* complex

  12. rion

    fippo had some review of my change. at least I trust him and I think others too :) According to him there is another way with different pros/cons and he implemented it in some client.

  13. jonas’

    rion, until last week, I didn’t even know who "fippo" is :)

  14. jonas’

    so it’s not as easy and clear cut sometimes

  15. Daniel

    We should probably block a lot of those until we have more (open source) experience with them. For example once the last call for message init it out I will suggest to hold it. I'm planning an implementation for later this year and based on the fact that my implementation will only be the second one that's a good reason to gain some more experience with it imho

  16. rion

    message init?

  17. jonas’

    jingle message initiation

  18. jonas’

    a non-iq flow to initiate jingle sessions

  19. Ge0rG

    IMHO, the XSF Council should be, in theory, knowledgeable in all aspects of XMPP. However, with over 400 XEPs, this is simply not feasible. So I think that (3) is the most viable way forward, and for Jingle we should maybe have three volunteers with sufficient experience to chime in

  20. rion

    ah. I'm going to implement it too but not sure how soon

  21. jonas’

    I wonder whether we should use Last Call to find folks who are knowledgeable in the spec and not advance it unless there are any

  22. jonas’

    but that of course does not help with the typical human churn in ecosystems

  23. Daniel

    jonas’: isn't that literally the point of last call

  24. jonas’

    Daniel, no, the last call is for feedback (in my understanding)

  25. jonas’

    not for writing down the names of the poor souls who *provided* feedback as victims for future review requests :)

  26. Daniel

    i mean it's almost funny. with simple xeps like http upload you got tons of bike shedding which prevents it from going through call because everyone seems to have an opinion. and the complex xeps 'nobody' understands just go through

  27. Ge0rG

    yeah, that's rather sad.

  28. jonas’

    that’s the specific definition of bike shedding, isn’t it?

  29. fippo

    the situation wasn't much different for RFC 6120...

  30. fippo

    message init has been implemented and from what i discussed with lance it was with mam enabled even

  31. fippo

    and the case that was raised on list (need to reply...) seemed unlikely unless you consider cases like hunt groups where multiple people use the same account

  32. Ge0rG

    fippo: I'm very much interested in the interaction of message init (actually all non-directly-IM use cases of messages) with MAM and Carbons

  33. Ge0rG

    especially under the new Carbon rules which are apparently widely ignored.

  34. fippo

    ge0rg: its a bigger topic, yeah. it was originally designed to interact well with both ("did i have a call from that person" - mam; "i took the call on the other device" - carbons)

  35. ralphm

    At VEON we used dedicated messages for the latter. We called them CDRs, although I'm not sure if that was the right call (haha).

  36. ralphm

    https://ralphm.net/publications/xmpp_chat_voip/#/6/4

  37. ralphm

    (example below this slide)

  38. ralphm

    We didn't get to submit a proper XEP on it, but I probably should.

  39. ralphm

    (Ge0rG, fippo)

  40. Ge0rG

    > I probably should This happens everywhere in the XSF, doesn't it? :)

  41. ralphm

    Ge0rG: a company shutting down your project and development team tends to change its one's priorities.

  42. ralphm

    Ge0rG: a company shutting down your project and development team tends to change one's priorities.

  43. Ge0rG

    ralphm: not trying to criticize you, it's just a general sentiment

  44. fippo

    ralphm: i never dared to call anything CDR inside a telco :-)

  45. dwd

    fippo, I called things CDRs, but only because they were.

  46. dwd

    Ge0rG, It's nice when the Council has members with direct knowledge and experience of a subject area. Where we don't, I hope people at least know someone they trust and can ask for a review. That's certainly how I work - I don't claim any serious understanding of VOIP, for example, but I'm pretty confident that fippo knows a thing or two.

  47. ralphm

    fippo: haha

  48. ralphm

    I read most messages by fippo and it made sense, but I didn't respond. Should I do more "I agree" messages?

  49. Ge0rG

    dwd: that's a great way to work indeed. But maybe we can formalize that out just a little bit, with "trusted people" being loosely defined for certain topics like Jingle, and "ask for a review" happening on-list?

  50. dwd

    Ge0rG, I think if we wanted to formalize it, I'd say we should form a SIG and/or Work Team for it and formalize it that way.

  51. dwd

    ralphm, Yes. Speaking personally, a whole bunch of "I agree", and "Me too", is very useful for me to gauge consensus on things.

  52. Zash

    Silence is agreement!

  53. dwd

    Zash, It's not though. It's just silence. And if we choose to interpret silence in a particular way, it becomes quite complex to judge almost anything.

  54. dwd

    Zash, IoW, All that is required for evil to triumph is for good men to do nothing.

  55. jonas’

    especially on topics I’m not familiar with, I don’t know if silence is due to other people also not being familiar, or whether there’s truly a horde of folks who know things and simply agree in silence

  56. Zash

    s/for .*/negative parlamentarism/ ? :)

  57. ralphm

    Ok, I'll speak up more.

  58. jonas’

    we need simple mailing list reactions :)

  59. Zash

    `Content-Type: human/reaction; charset=utf-8`

  60. Ge0rG

    ralphm: that'd be great

  61. fippo

    a jingle directorate? but remember, its already hard work to get people to review stuff

  62. dwd

    jonas’, I sometimes wonder if we shouldn't switch to thumbs up/thumbs down/comments on PRs as the feedback of record.

  63. jonas’

    but that would be fully moving to GitHub

  64. jonas’

    but only for PRs

  65. dwd

    jonas’, Move off mailing lists totally, and maybe see about integrating chatroom archive fragments into PRs.

  66. jonas’

    which is kind of meh, because there’s then no single source of truth for technical discussion

  67. dwd

    jonas’, Well, PRs and issues.

  68. jonas’

    I see

  69. jonas’

    I’m not fond of that

  70. Zash

    I do not approve

  71. dwd

    jonas’, I'm not wild about the idea either. But I wonder whether mailing lists are really the most appropriate thing these days.

  72. jonas’

    haven’t seen anything else with that fine-grained multi-device synchronising read marker support.

  73. Ge0rG

    dwd: essentially anyone can +1 or -1 on github, and sometimes PRs go viral and get thousands of useless votes

  74. dwd

    Ge0rG, For sure.

  75. dwd

    Ge0rG, Any alternative I could suggest to mailing lists would get the same push-back, of course.

  76. Daniel

    maybe github allows us to limit that to members

  77. dwd

    Ge0rG, My question is not, therefore, "Is X a perfect solution?" but "Is X better for us than mailing lists?"

  78. jonas’

    one thing which mailing lists give me which nothing else has been able to so far is what I mentioned above: I can read parts of a thread, mark them as read when I not only read/looked at it, but also understood it, and have that state synchronised across my clients.

  79. dwd

    Ge0rG, And, of course, discussing such things on a mailing list are a recipe for survivorship bias of sorts.

  80. jonas’

    which is vital for how I work

  81. jonas’

    no web-based system has had that yet.

  82. dwd

    jonas’, Yes, I can understand that.

  83. Ge0rG

    dwd: right. I tend to side with jonas’ here though. While ugly, I think that establishing a sort of +1 one-liner comment on-list from competent participants would be good

  84. jonas’

    it’d break my ability to participate.

  85. dwd

    Ge0rG, By the way, if we had an XSF PR that garnered thousands of votes, it'd be a problem I'd love to have.

  86. Ge0rG

    it appears to me that the Matrix has a very active and vocal marketing department, leading to a percetion of it being the only federated chat system out there.

  87. dwd

    Yes, indeed.

  88. Daniel

    they also had the genius and very unique idea to create bridges to other IM networks

  89. jonas’

    very genious, very unique

  90. Ge0rG

    I'm sure we all know what we have to do. And we lack the time to.

  91. jonas’

    Ge0rG, more marketing?

  92. Ge0rG

    > matrix-appservice-purple@conf.xmpp.matrix.org/Ge0rG: cancel: Server-to-server connection failed: Connecting failed: closed

  93. Daniel

    fwiw we are going to be at froscon in august

  94. Daniel

    with both a booth and a talk

  95. jonas’

    Ge0rG, you’re not helping the lack of time by being cryptic :)

  96. Ge0rG

    Daniel: we = the XSF?

  97. Daniel

    yes

  98. Ge0rG

    Thanks for reminding me. I'll try to make an appearance as well. Are volunteers still needed?

  99. Ge0rG

    Maybe somebody should issue a tweet? A mention in the newsletter?

  100. Daniel

    probably. the booth certainly could need 'stuff'. we might have enough 'staff'. but last year or when ever that was we were sitting at what was basically an empty table

  101. Ge0rG

    don't we have SCAM for that?

  102. Daniel

    well scam is just people

  103. dwd

    Yes, but if you need Stuff, they're the people to ask.

  104. Daniel

    scam could certainly pay for a poster or get it printed. but someone still has to design it

  105. dwd

    I think Ralph and/or Guus have Stuff, for example.

  106. dwd

    Daniel, Again, yes, but if nobody knows that's what you need, it definitely won't happen.

  107. Daniel

    i mean i have a flag. i will certainly get more stickers printed

  108. dwd

    Daniel, I think Guus has a load of stickers he can send you.

  109. Daniel

    but also like a general 'this is xmpp' poster would be nice. something like that

  110. Daniel

    but yes. we will try to be more organized this time.

  111. Ge0rG

    A nice pictogram of the federated nature of XMPP?

  112. Daniel

    i think Ge0rG was offering his help. and this might be something he could help us with I guess

  113. dwd

    Pictures of doctors, soldiers, and so on. :-)

  114. Ge0rG

    Bonus points if the background is black with fading green hieroglyphs...

  115. Ge0rG

    dwd: soldiers are not PC any more

  116. Guus

    Daniel: let's try to create reusable designs for this, so that we can use them for other conferences too. Stuff like posters etc are very nice.

  117. dwd

    Ge0rG, Yes and no. People are usually impressed with the fact it's XMPP used for MEDEVAC in battlefields. Something that WHatsapp or Matrix can't claim.

  118. Daniel

    terrorists and nato

  119. dwd

    Guus, Those pop-up stands are quite cheap these days, I think.

  120. Ge0rG

    dwd: Matrix can claim the French government.

  121. Daniel

    and matrix only has the french goverment

  122. Ge0rG

    also they have 100bps and we have only 75bps.

  123. dwd

    Ge0rG, Yet XMPP can claim the French military, as well as every other NATO one.

  124. Guus

    dwd: payment won't be the problem, design will be.

  125. Ge0rG

    dwd: we can also claim Nintendo, EA and some other evildoers.

  126. Ge0rG

    Did I mention Google?

  127. dwd

    Ge0rG, Oh, yes, indeed.

  128. Guus

    I'm taking my kids to the swimming pool now, I'm getting anxious looks

  129. Guus

    But yes, let's do create more eye candy

  130. dwd

    Guus, Go drown them.

  131. Ge0rG

    I have this from CLT 2016... https://op-co.de/tmp/CLT-2016-preview.jpg

  132. Ge0rG

    froscon is probably a slightly different audience, should be something with the spirit of FOSS

  133. Daniel

    clt and froscon are pretty similiar

  134. fippo

    EA uses xmpp? Epic does for sure

  135. jonas’

    froscon always sounds like frogs to me

  136. Ge0rG

    fippo: https://xmpp.org/uses/gaming ;)

  137. jonas’

    maybe we should remove EVE there for a while, people are complaining that EVE wants to make the in-game chat slower ;)

  138. Ge0rG

    jonas’: they want _what_?

  139. jonas’

    https://www.derstandard.at/story/2000106030972/eve-online-etliche-spieler-drohen-mit-kuendigung-ihres-abos-wegen I only have that, which is both cookiewalled and in german

  140. Ge0rG

    Daniel: I haven't been to froscon, but CLT is very user oriented, not so much targeting nerds

  141. Daniel

    froscon is user oriented~ish as well

  142. Daniel

    froscon is not fosdem

  143. Ge0rG

    alright

  144. Daniel

    maybe a little bit closer to fosdem (or ccc events) that clt

  145. Daniel

    but certainly not fosdem or congress

  146. Ge0rG

    Daniel: do you have any other marketing material that we can make use of?

  147. Daniel

    i have a flag. lol

  148. Ge0rG

    Daniel: what about a description of the booth, and what we are allowed to put there?

  149. Daniel

    it's just one table. I'm not aware of any limitations regarding on what we are allowed to put there. you can’t advertise commercial services. but that's the only limitation i'm aware of

  150. Daniel

    also technically i'm not the organizer of the booth. one of the dino developers is

  151. Ge0rG

    Daniel: is there a MUC, or some other collaboration tool?

  152. Daniel

    Ge0rG: not yet

  153. rion

    > [15:40] <Ge0rG> fippo: https://xmpp.org/uses/gaming  I saw this xmpp for Fortnite. Unfortunately I can't unsee it.

  154. Ge0rG

    Bad people use good technology for bad things...

  155. waqas

    Ha, that's a neat page. I only knew about a few of them.

  156. Ge0rG

    There are also significant numbers in IoT, like Google Cloud Printing. Somebody just needs to collect evidence

  157. Zash

    Wasn't 0 AD using XMPP too?

  158. Ge0rG

    Zash: was that released yet?

  159. Zash

    Website says alpha, but gameplay seems fairly solid.

  160. ralphm

    We were next to them at FOSDEM, right?

  161. Zash

    Yes

  162. Ge0rG

    So were they using xmpp?

  163. ralphm

    Yes

  164. Zash

    https://trac.wildfiregames.com/browser/ps/trunk/source/lobby

  165. edhelas

    actually most of the project at FOSDEM are using XMPP, but we don't know yet

  166. edhelas

    except the Matrix guys

  167. Ge0rG

    Do we want them added to the list? Number of users: ~10

  168. moparisthebest

    *technically* matrix's xmpp gateway uses xmpp right?

  169. moparisthebest

    So matrix should go on that list of companies using XMPP

  170. Ge0rG

    moparisthebest: the gateway is down for months now

  171. moparisthebest

    The software still exists though?

  172. waqas

    That's a bit of a stretch

  173. Ge0rG

    waqas: you are outdated. It's a bit of a buster now.

  174. neshtaxmpp

    hi