XMPP Council - 2016-10-12


  1. psa

    I'm here. :-)

  2. Lance

    it is time

  3. Lance

    0) Roll call

  4. MattJ

    Here

  5. Lance

    Tobias Dave Cridland

  6. Dave Cridland

    Hello

  7. Lance

    I'm not aware of new items for voting

  8. Lance

    There is still https://github.com/xsf/xeps/pull/219 needing votes from Dave and Matt

  9. MattJ

    Oops, thanks

  10. Lance

    and from psa

  11. Lance

    1) Date of next

  12. Lance

    sbtsbc?

  13. MattJ

    wfm

  14. Dave Cridland

    #219 is fine by me. I've spent years arguing for it, after all.

  15. Lance

    ta

  16. Dave Cridland

    Erm, AOB.

  17. Lance

    2) AOB?

  18. Dave Cridland

    There's an IoT SIG proposal; whose call is that do we think?

  19. MattJ

    Ours?

  20. psa

    The Board approves new SIGs, I think.

  21. psa

    At least, I specified that the Board is the approving body this time (we haven't created a SIG in quite awhile).

  22. MattJ

    Tobias, nice try ;)

  23. Dave Cridland

    psa, I think it can be either, but I noticed you sent it to the Board.

  24. Dave Cridland

    MattJ, I was impressed by the effort.

  25. psa

    In the distant past, SIGs were all approved by the Board IIRC.

  26. Tobias

    MattJ, I sent it 5 min ago. Just had no connection

  27. Dave Cridland

    Who do we want to approve it?

  28. Dave Cridland

    I'll be asking the same question in Board later, I just wondered what Council folks thought.

  29. psa

    My feeling is that the Board needs to approve SIGs, but the Council should provide technical guidance where needed.

  30. Lance

    That sounds reasonable to me

  31. Dave Cridland

    OK, I'll note that to the Board.

  32. Lance

    thanks Dave

  33. Lance

    AOAOB?

  34. Dave Cridland

    NOAOB.

  35. Lance bangs gavel

  36. Lance

    thanks all

  37. MattJ

    Oops, I meant to ask about http://xmpp.org/extensions/inbox/message-deletion.html

  38. psa

    That was approved and never published, right?

  39. Lance

    not approved iirc, there were some concerns that needed to be fixed

  40. Dave Cridland

    I looked at it earlier, actually, thanks to your reminder yesterday, MattJ - made me uneasy to look at it. Gut feeling is that it's never actually going to delete a message, and it's one of those cases where I think the naming is significant.

  41. Zash

    Tombstone ?

  42. Dave Cridland

    "retract" is the term we've used before.

  43. MattJ

    Yeah, I'd prefer that too

  44. Lance

    Right. It's one step stronger than a message correction that sets the body to empty string

  45. Dave Cridland

    If this were MIX, it *would* be a retract, I suppose.

  46. MattJ

    So what I would like is a defined thing to replace deleted/retracted/whatever messages in a MAM archive

  47. MattJ

    Since we can't actually remove the message without breaking sync

  48. Zash

    Like a tombstone?

  49. Zash

    Or retract in the past tense

  50. MattJ

    Like a tombstone

  51. Lance

    is that necessary at the protocol level, or something needed internally for a server?

  52. MattJ

    Necessary at the protocol level

  53. MattJ

    Well, strongly desired at least

  54. MattJ

    The problem is if you delete a message whose id a client was using as a sync point - the next time the client comes online and requests for all messages since that one, the server would find it didn't exist and send the whole archive (as per the spec)

  55. Lance

    so something like replacing the contents of the stanza with <retracted by='...' xmlns='...' /> ?

  56. MattJ

    Yes

  57. Lance

    and still sending that in the history results?

  58. SamWhited

    That makes good sense to me; it's consistent with how message editing does it and means you can always get a valid snapshot of history at any given time (whereas if you actually delete the messaeg and then try to sync "all history up to 5 minutes ago" you won't see the message even if it did exist 5 minutes ago

  59. SamWhited appears and catches up on history.

  60. psa

    SamWhited: agreed - making it consistent with message editing is especially appealing

  61. Zash

    Do we need something separate from a message correction?

  62. MattJ

    Zash, I think so

  63. SamWhited

    It means that if clients support editing, but not retraction they can advertise that; that's and it's just nice semantically to make them distinct, but I'm not sure that's important

  64. Zash

    I think we do care about semantics.

  65. SamWhited

    Sounds good to me :) it certainly "feels" right to me, I just don't have a good objective argument for why it's important.

  66. Lance

    i think the main distinction is that this is primarily meant to be a moderation and 'cleanup' tool. think a muc room getting spammed with dozens of messages. corrections alone don't quite convey the right intent

  67. Lance

    especially since (iirc) corrections doesn't cover the case of someone else (room admin/moderator) sending a correction for someone else's message

  68. SamWhited

    ah, good point

  69. SamWhited

    Added a card I forgot about for next time; I'm not entirely sure this needs council approval, but I never remember… https://trello.com/c/Jedp9SQX/119-vote-on-https-github-com-xsf-xeps-pull-250

  70. Lance

    SamWhited that looks all editorial, nothing needed from council here

  71. SamWhited

    Sounds good; I'll go ahead and archive that card then. Sorry for the noise.

  72. Lance

    no worries. noise is good :)

  73. ralphm

    Without noise, there's no signal

  74. psa

    such wisdom from ralphm

  75. Lance

    ok, I just sent a PR to update the message deletion protoxep based on the conversation earlier

  76. SamWhited

    Ralphm is the zen master of online communications :)

  77. ralphm

    Heh