XSF Discussion - 2019-05-02

  1. pdurbin

    I just posted this: https://discourse.opensourcedesign.net/t/restoring-chat-functionality-collaboration-with-xmpp-community/901/14

  2. Ge0rG

    pdurbin: đź‘Ť

  3. (◠_◠✿)Demon(◕‿◕✿)


  4. Zash

    Good morning to you too

  5. pdurbin

    The main question in my post above: Where would visual design and user experience discussions about XMPP clients take place?

  6. Zash

    I would guess that here or xmpp:jdev@conference.jabber.org?join or xmpp:modernxmpp@rooms.modernxmpp.org?join would work.

  7. Zash

    Tho here is more meant for protocol discussion than implementation things.

  8. Zash

    It's pretty much the same set of people regardless of where you, so maybe it doesn't matter that much :)

  9. pdurbin

    Here is fine with me. :)

  10. Ge0rG

    Zash: some people want to keep this place for protocol discussion only, because they tend to read up their backlogs. So it wouldbe good to separate it that way.

  11. Zash

    The modernxmpp room is probably closest in focus, altho more towards UX than graphical/visual design I guess.

  12. jonas’

    Graphics/Visual Design \subset UX

  13. Zash

    I don't know how to read that

  14. Zash

    I'd imagine most kinds of design have some overlap with each other.

  15. jonas’

    Graphics/Visual Design ⊂ UX

  16. pdurbin

    I'd prefer this channel because it's logged.

  17. Ge0rG

    Let's draw Venn diagrams all day long.

  18. Zash

    jonas’: Neat. Tho I don't have that much of a math/logic background to read that either.

  19. Zash

    Something something UX of Unicode

  20. jonas’

    Zash, "is a subset of"

  21. jonas’

    or "is contained in"

  22. Zash

    And here I say something about UX without graphics or something

  23. edhelas

    there was a job board planned for XMPP no ?

  24. Ge0rG

    edhelas: yes

  25. MattJ

    pdurbin, ah, turning on logging for the modernxmpp chat is on my short-term todo, I just haven't had much time for such admin tasks recently

  26. pdurbin

    MattJ: that would be great!

  27. pdurbin

    MattJ: want me to create a GitHub issue?

  28. MattJ

    That would be great, thanks

  29. pdurbin

    MattJ: done: https://github.com/modernxmpp/modernxmpp/issues/20

  30. MattJ

    Thanks :)

  31. pdurbin

    sure :)

  32. pdurbin

    I just created this "Document how to set up a Gitter-like user experience" issue as well: https://github.com/modernxmpp/modernxmpp/issues/21

  33. Wiktor

    Is this only about room logs pdurbin? I'd love to see xmpp web client with such high quality / UX as gitter...

  34. Zash

    The room log thing is a single purpose read-only room log view.

  35. Zash

    pdurbin: Check out https://chat.prosody.im/ or https://xmpp.org/chat

  36. MattJ

    Zash, maybe we could make the links buttons

  37. Zash


  38. Zash

    Can a button be an xmpp: link tho?

  39. MattJ

    Just give the <a> round borders and a background :)

  40. Zash

    Go for it

  41. MattJ

    Guus, iirc you said you couldn't make the meeting this week, but you're the only one here. Did we agree to skip or what?

  42. Guus

    I actually apologized for last week.

  43. Guus

    This week, I'm simply late 🙂

  44. Guus

    but if it's just the two of us...

  45. MattJ

    Seve, ping

  46. MattJ

    Too many nicks :)

  47. Guus

    Ralph is in a plane, and nyco also apologised iirc

  48. MattJ

    Ok, we can skip then I think

  49. Guus


  50. Zash

    So you have free time to fiddle with that CSS then? ;)

  51. MattJ

    Free time?

  52. MattJ

    I'd like to hear more

  53. Guus

    Hah! Yesterday, I overheard two guys discuss prices of GPU-based cloud servers (and that they're insanely expensive).

  54. Guus

    I asked why they wanted that power. One guy laughed, pointed at the other guy, and said: "he wants to rotate a css object"

  55. Guus

    I stopped asking questions at that point.

  56. Zash

    And backed away slowly?

  57. MattJ


  58. edhelas

    "he wants to run a Matrix instance for 100 users"

  59. eevvoor

    You have weekly meetings here?

  60. Zash


  61. Zash

    The XSF board has meetings here and the council has weekly meetings in xmpp:council@muc.xmpp.org?join

  62. eevvoor

    As a member should I participate in some meetings? I've applied for membership for Q2.

  63. Ge0rG

    eevvoor: as a member you only must attend the quarterly membership meetings (or vote in advance), but you are allowed to listen in on any meetings

  64. Ge0rG

    we are looking for volunteers to write the meeting minutes

  65. Zash

    Members vote on membership applications, as well as applications to the board and to council.

  66. eevvoor

    You write minutes of a chat? wow. Yes I'll attend end of may of course. Thus no more duties concerning meetings, I see.

  67. pep.

    "but you are allowed to listen in on any meetings" you don't actually have to be a member for that

  68. eevvoor

    pep. yes ;)

  69. eevvoor

    What are your expections for a member?

  70. Zash

    Vote on the things. :)

  71. Zash

    Being an active participant in the community and standards process doesn't hurt of course, but you don't really need to be a member for that.

  72. pep.

    You do if you want to be part of some teams that board defined I guess, but yeah otherwise you mostly don't

  73. Zash

    Do you even need to be a member for those?

  74. pep.

    You do vote for council and board, so you get to shape the future of the foundation :)

  75. pep.

    dunno, I'd think so

  76. Zash

    There are some kind of focus groups that can be created where I only think the chair needs to be a member, or something like that

  77. Zash

    Might be different from teams tho

  78. eevvoor

    I meant it the other way round: are there situiations where you don't accept the membership application?

  79. Zash

    there are some rules for when they should be rejected

  80. Zash

    like if there are too many from the same organization already accepted, or if it's of the wrong form

  81. MattJ

    It's very rare that someone doesn't get accepted

  82. jonas’

    Zash, you need to be a member to be an editor for example

  83. jonas’

    I think at least

  84. MattJ

    Is there any indication of consensus on references using bytes or... glyphs or codepoints or...?

  85. MattJ

    Just had possibly our first interop problem due to this 'TODO' :)

  86. MattJ

    (I'm really not a fan of accepting XEPs with 'TODO' comments FWIW)

  87. jonas’

    MattJ, I strongly suggest codepoints

  88. jonas’

    Sam strongly suggests bytes

  89. Zash

    I'd go with codepoints

  90. jonas’

    given that we operate on top of XML, codepoints (or XML’s version of that) seems like the only sensible thing to do

  91. Zash

    It would just be painful to count bytes (of UTF-8?) if your internal representation of the data is in UTF-16 or something, which isn't uncommon

  92. jonas’

    yeah, think javascript

  93. jonas’

    although javascript will have a pain anyways because UTF-16

  94. jonas’

    adn I think we can all agree that counting UTF-16 things is the worst of all worlds

  95. MattJ

    If you have a Python background then codepoints obviously make sense, if you have a C background bytes make sense, if you have a Lua background the whole thing is irrelevant because the index should start at 1

  96. Zash

    Counting UTF8 bytes when you have UTF-16 seems ... not fun

  97. Zash


  98. MattJ

    It looks like Converse.js is using codepoints

  99. Ge0rG

    A world of pain, and then there is "đź’©".length === 2

  100. Zash

    Well if you have UTF-16 then you get to deal with that fun

  101. dwd

    If references is counting codepoints (or even bytes) should we be normalizing the text first?

  102. jonas’

    text as on the wire

  103. jonas’

    otherwise you’re up for lots of fun with unicode versions

  104. lovetox

    please what ever i can determine with len(string) in python, thanks

  105. jonas’

    lovetox, codepoints

  106. lovetox

    and bytes makes not a lot of sense to me, because the xml comes encoded, to deal with it in any way, i have to decode it, so now i dont have bytes anymore

  107. moparisthebest

    I propose "non-ascii unsupported" >:)

  108. lovetox

    then i pick the body, than re encode the body to utf8 then count?

  109. lovetox

    sounds shitty

  110. jonas’

    lovetox, agreed

  111. jonas’

    lovetox, but XML implementations in C and other lower level languages will typically give you UTF-8

  112. MattJ

    Your problems are alien to me, but I can imagine how that might be a pain

  113. jonas’

    so there’s an argument for that, too

  114. lovetox

    yeah, sorry python noob here :/

  115. Ge0rG

    MattJ: your language does pattern matching based on individual bytes, you are our

  116. Ge0rG

    MattJ: your language does pattern matching based on individual bytes, you are out

  117. Zash

    Maybe one day we'll have utf8 aware pattern matching too

  118. moparisthebest

    I think rust has it

  119. jonas’

    Zash, ITYM Unicode-aware? :(

  120. Zash

    No, I did not

  121. Zash

    Lua has some utf8-specific stuff in recent versions, but no other unicode things.