XMPP Council - 2024-10-08


  1. daniel

    Hi.

  2. daniel

    It's time

  3. dan.caseley

    Howdy πŸ‘‹

  4. daniel

    1) Roll call

  5. moparisthebest

    Hello!

  6. daniel

    singpolyma, larma you around?

  7. larma

    πŸ‘‹

  8. singpolyma

    hello

  9. daniel

    No agenda. No editors notes. No new items for voting.

  10. daniel

    We do have pending votes on the new happy eyeballs proto xep

  11. daniel

    those expire today. at the current state it would be rejected simply because there aren’t enough +1

  12. daniel

    personally I didn’t get around reading it

  13. daniel

    so i'd probably give us another week?

  14. singpolyma

    IIRC I did +1

  15. daniel

    singpolyma, that correct. and I recorded that. we have 2 +1. we'd need at least 3

  16. moparisthebest

    I really dropped the ball on list feedback, but tldr I think that whole XEP should be "when you have a domain+port to resolve and connect to, refer to happy eyeballs RFC" the rest seems excessive and not useful?

  17. dan.caseley

    I'd rather extend than reject a XEP because we failed.

  18. daniel

    certainly.

  19. daniel

    the alternative would be that we get one other +1 right now :-)

  20. moparisthebest

    I'd rather extend also if we can, if I had to vote today it'd be -1 because I see no reason for it to have all the detail it does

  21. moparisthebest

    I might be missing something though

  22. daniel

    ok. I’m giving us another week.

  23. daniel

    maybe do send your feedback to the list moparisthebest

  24. dan.caseley

    There's some XMPP extension to the RFC, e.g. SRV weighting.

  25. moparisthebest

    I will send on list tonight for sure, just set an alarm 🀞

  26. daniel

    anyway. moving on for now

  27. daniel

    …) Date of Next

  28. daniel

    +1w wfm

  29. larma

    I'm +1 on happy eyeballs. I don't think it's strictly needed, but doesn't hurt either

  30. moparisthebest

    +1w wfm

  31. daniel

    larma, noted

  32. larma

    +1w wfm

  33. dan.caseley

    +1w wfm I'll likely miss +2w and +3w for family holiday.

  34. daniel

    …) AOB

  35. daniel

    i do actually have two of them myself

  36. moparisthebest

    You should reapply for council if you want

  37. daniel

    council election is less than a month away. if you consider applying for next year do it soon

  38. moparisthebest

    oops double (:

  39. daniel

    the other one I had: I was thinking about starting Last Calls on Message Displayed Sync and Message Reactions (the latter obviously only with approval from the authors)

  40. daniel

    that obviously not a vote but I was wondering what everyones feelings are on those

  41. moparisthebest

    Sure, a last call can't hurt

  42. dan.caseley

    Certainly on Message Retraction. A quick look at the history says it's been Experimental for a very long time.

  43. daniel

    ok.

  44. singpolyma

    I think daniel said reaction not retraction

  45. daniel

    no other AOB I guess?

  46. daniel

    …) Close

  47. dan.caseley

    > I think daniel said reaction not retraction right you are... Maybe we should do retraction first πŸ˜‚

  48. daniel

    thank you all. see you next week

  49. moparisthebest

    Thanks all!

  50. moparisthebest

    Why not message retraction also 😁

  51. dan.caseley

    That was my reaction!

  52. dan.caseley

    I'll get my coat...

  53. moparisthebest

    Nice

  54. singpolyma

    If we do LC on retraction I'll surely vote -1 since it's not widely implemented and IMO not even needed

  55. singpolyma

    Reaction xep I don't love but it's well implemented and it's fine

  56. dan.caseley

    Message Retraction has 20 implementations according to the website

  57. singpolyma

    ...how? Are there even 20 clients anyone uses?

  58. Zash

    clients, libraries, servers?

  59. daniel

    Note that we need fairly active authors to both request the LC and actively and timely incorporate feedback if there is any. That's why I picked MDS and reactions specifically because those happen to be council members

  60. larma

    Regarding reactions: IMO we don't have a lot of implementation experience of restriction and rejection of reactions.

  61. daniel

    I was meaning to ask nicoco about that. Because a bridge would probably be the likely (if not only) candidate to need that

  62. Kev

    FWIW, We're in the early stages of implementing Reactions at the moment.

  63. Kev

    > I was meaning to ask nicoco about that. Because a bridge would probably be the likely (if not only) candidate to need that I think you can legitimately need it outside bridges too. It's easy to imagine non-private MUC cases where having unrestricted reactions would not be good.

    🀯 1
  64. Kev

    Or enterprise situations where a restricted reaction-set is seen to be desirable.

  65. Zash

    I'm sorry Kev, that reaction is above your pay grade

  66. larma

    Yes, MUCs are also explicitly mentioned as a usecase for restrictions. Also potentially might be useful to completely ban reactions in a room

  67. Kev

    Now that is an interesting approach I hadn't thought of. Tying particular reactions to clearances :)

  68. Kev

    "You're cleared for COSMIC TOP SECRET, therefore you can use the πŸ’© reaction"

  69. Zash

    /PRIVATE//NOBUS/πŸ‘οΈ

  70. Guus

    moparisthebest: I'm somewhat surprised that your concerns regarding Happy Eyeballs may be reason to prevent the text to be accepted as an Experimental XEP. If my reading of the process is right, a -1 vote at this stage would prevent the text even from being published / made available for public discussion through XSF's channels. Does that properly reflect your concerns with that text? (I am under the impression that to qualify for acceptance, the bar basically isn't much higher than "the text is formatted correctly").

  71. Guus

    (sorry for butting in, I had this typed and was waiting for the other discussion to die down - but I need to leave to pick up kids from practice)

  72. singpolyma

    Guus: it is already published in the inbox for discussion

  73. Kev

    While that's true, my understanding was that the XSF was always wary of discussion (and especially further development) of specs pre-IP assignment.

  74. singpolyma

    That seems backwards. Surely you'd want a XEP to be properly developed before accepting it

  75. Kev

    That would certainly be a new, interesting direction :)

  76. Kev

    The usual barrier is just "doesn't unnecessarily duplicate existing XEPs" and "isn't harmfully specified" isn't it?

  77. Kev

    We haven't even always required a XEP to be fully-specced enough to be implementable, if we wanted to start discussion on something.

  78. daniel

    This council has indeed put a fairly high bar on accepting a proto xep

  79. daniel

    There has always been room for interpretation and this council has had a rather strict one

  80. Kev

    Well, I suppose it was inevitable. First Draft became the new Final, now Experimental became the new Final.

  81. daniel

    Apparently

  82. Zash

    Wasn't Experimental originally meant to be like the IETF draft stage?

  83. Kev

    In all seriousness, though, why? I know orgs do tend to get more conservative and bureaucratic as they get older, but it sounds like in this case it was a deliberate decision to make the change?

  84. daniel

    For a while we did fairly poorly on advancing things to draft / stable. Thus people started to implement experimental xep. Instead of fixing that and do better on advancing things to draft we raised the bar for experimental

  85. daniel

    Personally I disagree with that and would much rather prefer that we do better on LC+advancing to stable. And I hope that my actions and my voting history reflect that

  86. singpolyma

    > Well, I suppose it was inevitable. First Draft became the new Final, now Experimental became the new Final. I hope not, since final means we can't change it and we definitely need to change lots of the experimental ones :)

  87. Zash

    Final doesn't mean we can't change it, just that it needs council approval to change, no?

  88. singpolyma

    Hmm, no as far as I have heard?

  89. singpolyma

    even Stable we mostly can't change. has to be backwards-compatible

  90. singpolyma

    final is like editorial changes only

  91. Zash

    I find it sensible for the process to make it progressively harder to break backwards compatibility, which also happens when there are more implementations.

  92. Zash

    > Note: Once an XMPP Extension Protocol has been advanced to a status of Final, every effort shall be made to limit the scope of modifications; in particular, backwards-incompatible changes shall not be made. However, limited modifications may be made as long as they are optional, backwards-compatible extensions rather than modifications to the core protocol itself. Therefore, a Final protocol is safe for deployment in mission-critical applications.

  93. Zash

    And if breaking changes are really needed at that point, make a new Experimental XEP that superseeds it

  94. singpolyma

    Yes, I agree at least that with a process like ours that's the right choice

  95. singpolyma

    I was just expressing that "experimental is the new final" is definitely not my PoV

  96. singpolyma

    But I don't think experimental is a scratch pad either

  97. singpolyma

    We're publishing these as XEPs with a number for people to implement

  98. Guus

    > The granting of Experimental status must not be construed as indicating any level of approval by the XSF, the XMPP Council, or the XMPP developer community. Implementation of Experimental XEPs is encouraged in an exploratory fashion (e.g., in a proof of concept) in order to gain experience with and iteratively improve the protocol defined therein, but such implementations might not be appropriate for deployment in production systems.

  99. Guus

    I'm not saying it is a scratch pad, but we're also not actively promoting experimental XEPs as being production-ready.

  100. Zash

    Don't they put IETF drafts into production too?

  101. Zash

    Yes, they!