XMPP Council - 2025-03-11


  1. daniel

    It is time

  2. daniel

    1) roll call

  3. goffi

    here

  4. daniel

    singpolyma, larma you around?

  5. daniel

    (we don’t have anything on the agenda so no worries if you aren’t)

  6. larma

    👋

  7. moparisthebest

    Note from peanut gallery copy/pasted from xsf@ > So https://xmpp.org/extensions/xep-0172.html is stable, but security considerations say "you SHOULD do it like https://xmpp.org/extensions/xep-0165.html" which is deferred/experimental :/ > > Might be something for council to look at

  8. daniel

    2) Agenda Bashing no agenda this week. just here to catch up and eventually prompt for AOB

  9. singpolyma

    Hi

  10. daniel

    3) Editors update

  11. daniel

    editor has published 503: spaces

  12. daniel

    4) Items for voting

  13. daniel

    none

  14. daniel

    5) Pending votes

  15. daniel

    none

  16. daniel

    6) Date of Next

  17. daniel

    I have to watch my timezones but +1w wfm

  18. goffi

    +1w wfm

  19. singpolyma

    +1w wfm

  20. daniel

    7) AOB

  21. daniel

    should we quickly discuss what moparisthebest said?

  22. larma

    Will be at IETF +1w

  23. daniel

    personally I think the security considerations of User Nick seem a bit strict?

  24. singpolyma

    Probably could be renamed to privacy consideration mostly

  25. daniel

    Conversations for example does a lot of what 165 suggests wrt to highlighting irregular unicode chars

  26. daniel

    but i don’t think we can reasonably apply the same strictness to nicknames which inherently aren’t unique

  27. daniel

    i mean I can highlight the weird 'a' in 'Daniel' but nobody can stop anyone from just copying the name without the weird 'a'

  28. goffi

    sorry, I'm lost here. What are you refering to?

  29. singpolyma

    Oh I see what you're saying. Right. I think if you show jid or something always alongside an untrusted nick it's ok

  30. daniel

    yes. and/or at least show the jid where it matters(tm)

  31. daniel

    but yes Conversations for example doesn’t use the nick for 'strangers' but always shows the jid

  32. moparisthebest

    For reference what made me look again was a new gajim commit to do that (display "$nick ($jid)") but no length limits on nick so it might end up hiding $jid with lots of whitespace

  33. moparisthebest

    It's an ever present footgun....

  34. daniel

    but that said I guess we can also LC 0165

  35. daniel

    goffi, see the quote from moparisthebest a few lines above

  36. daniel

    which in turn is a quote from the xsf room from a few days ago

  37. goffi

    Oh right, sorry I've several things happening at the same time today.

  38. daniel

    moparisthebest, sure but then "don’t show username for 'untrusted' chats' might still be a better security consideration than "highlight unicode chars"

  39. daniel

    but if council doesn’t have a concrete suggestion we could try a mailing list thread?

  40. moparisthebest

    > moparisthebest, sure but then "don’t show username for 'untrusted' chats' might still be a better security consideration than "highlight unicode chars" Agree

  41. moparisthebest

    I linked the conversations commit that did that and lovetox said he didn't see why contacts should be more trusted... so kinda hard to push

  42. daniel

    i mean the primary reason i’m not showing user nicknames in Conversations for untrusted chats is that I was worried someone would set their nick to 'daniel@gultsch.de' (instead of trying to clash on 'Daniel')

  43. moparisthebest

    yep

  44. daniel

    as I’m not seeing other council members jumping into the discussion I will start a thread on the mailing list

  45. daniel

    any other AOB?

  46. daniel

    doesn’t seem to be the case

  47. daniel

    8) Close

  48. daniel

    thank you all. see you next week

  49. goffi

    Thanks daniel, thanks all.

  50. daniel

    i mean security considerations aren’t normative, no? so lovetox could just say "i've considered these securities and did something else"

  51. moparisthebest

    sure, I think the concrete standards problem is a stable standard referring to an experimental one

  52. moparisthebest

    Certainly not the end of the world but feels odd

  53. Daniel

    Oh yes certainly. I assumed everyone was agreeing to that. I was trying to get a discussion started on how we fix this

  54. singpolyma

    I think replacing it with a statement about the actual consideration "consider that a user might try to spoof your UI by setting their nickname do a Jabber ID, especially for untrusted users" might be enough

  55. moparisthebest

    I tend to agree, could additionally note "beware Unicode lookalikes" but meh