jdev - 2024-04-22


  1. rom1dep

    Sorry if that's the wrong chan for that, I was thinking about message replies, is there any provision for replying to only part of a message? If not, would it be advisable (although ugly) for clients to split multi lines messages so that distinct lines can be replied individually?

  2. lovetox

    I think that train left the station. Yes it would be horrible I'd you send a message for each line, only for the chance that somebody wants to reply to a specific line

  3. lovetox

    Simply use quoting as always

  4. MSavoritias (fae,ve)

    i dont think there is any app that allows you to select a specific thing to reply to

  5. MSavoritias (fae,ve)

    with replies that is, not quotes

  6. debacle

    Please register yourself to the Berlin XMPP sprint in July, or ask someone with wiki access to do so for you: https://wiki.xmpp.org/web/Sprints/2024-07_Berlin

  7. rom1dep

    > Yes it would be horrible I'd you send a message for each line, only for the chance that somebody wants to reply to a specific line Not necessarily, the displaying could be that of a single message, while the quoting UX lets you do it at paragraph-level

  8. rom1dep

    > Simply use quoting as always Wouldn't that break the relationship/"jump back" to the quoted message?

  9. rom1dep

    If not, I'm all for it

  10. MSavoritias (fae,ve)

    what is the differnce between quoting and replies?

  11. MSavoritias (fae,ve)

    in xmpp client level

  12. lovetox

    rom1dep: it's not worth the effort for a jump to functionality

  13. lovetox

    Cost/Benefit does not check out for me

  14. rom1dep

    lovetox: it just looks inconsistent, then (you end-up with both replies and quotes, which are basically the same thing conceptually from a user perspective, but have vastly different user experiences/expectations and underpinnings)

  15. rom1dep

    not that I want to die on that hill, I just think there's a case for that

  16. lovetox

    Hm it only the same on the very surface. Reply is not a quote for me. It just references a message. So it adds a context

  17. lovetox

    While a quote focuses on a specific part of a message

  18. MSavoritias (fae,ve)

    i mean there is no reason replies cant be quotes no?

  19. rom1dep

    > i mean there is no reason replies cant be quotes no? I thought that was the whole point :)

  20. lovetox

    MSavoritias (fae,ve): yes there is. It is not and it's published as standard.

  21. lovetox

    Of course you can do a new thing with a different bame

  22. lovetox

    Of course you can do a new thing with a different name

  23. MSavoritias (fae,ve)

    or i can update the replies xep :)

  24. lovetox

    MSavoritias (fae,ve): you can try. I doubt the author will follow your argument.

  25. MSavoritias (fae,ve)

    didnt know the author had all the power :/ well that sucks

  26. MSavoritias (fae,ve)

    i hope i dont need to change many xeps then

  27. nicoco

    About replies, if we're being practical, from what I see from non-techies in walled garden, they use rich replies *a lot*, notably because it has the nice property of "pinging" the person being replied to. And I've never seen any body be interested in partial quotes. Look at emails, besides a few weirdos (like me and I guess a lot of in people in the MUC), who quotes part of the original email and reply below?

  28. Kev

    Replies that point to the original message rather than quoting it are definitely widely used, not least because that's what Slack, Discord, Messages, etc. de.

  29. Kev

    Replies that point to the original message rather than quoting it are definitely widely used, not least because that's what Slack, Discord, Messages, etc. do.

  30. Kev

    (Well, Slack's slightly odd about how it does it, to be fair)

  31. nicoco

    And if we don't want to be practical and yakshave, we should think more about how replies and threads interact. In some popular walled gardens, they're unified as a single concept for instance, ie, replying to a message starts a new thread from this message.

  32. MSavoritias (fae,ve)

    i mean most of the time i quote because the client i am using shows the *whole* message and i dont want it to because it takes space

  33. MSavoritias (fae,ve)

    > And if we don't want to be practical and yakshave, we should think more about how replies and threads interact. In some popular walled gardens, they're unified as a single concept for instance, ie, replying to a message starts a new thread from this message. +1

  34. MSavoritias (fae,ve)

    ( i dont have reactions)

  35. rom1dep

    > yup, the current status-quo (well, besides XMPP, which still mostly lacks replies) is a 95% solution, but even non-techies sometimes write long messages with multiple ideas/concepts defeating "reply to whole message" > If we could have replies with start_char_ix/end_char_ix it could be neat (but yet again the UX wouldn't be trivial to come up with) yup, the current status-quo (well, besides XMPP, which still mostly lacks replies) is a 95% solution, but even non-techies sometimes write long messages with multiple ideas/concepts defeating "reply to whole message" If we could have replies with start_char_ix/end_char_ix it could be neat (but yet again the UX wouldn't be trivial to come up with)

  36. lovetox

    rom1dep: you can reply and quote at the same time

  37. lovetox

    Of course you probably say then I want to reply to multiple people in the same message

  38. rom1dep

    > Of course you probably say then I want to reply to multiple people in the same message Isn't that permitted with the current reply spec? I see that used very often on Teams

  39. lovetox

    No it's not permitted

  40. lovetox

    The problem is the reply spec has no notion of where a replY should be included inside the text

  41. lovetox

    You could add multiple replies, but the receiving client cannot assign specific text to each reply

  42. lovetox

    And even if you add a possibility to specify start end etc

  43. lovetox

    This breaks the fallback indication xep

  44. lovetox

    So all in all the reply XEP is a very simple WhatsApp reply kind of spec

  45. lovetox

    And that works for most people I would say

  46. lovetox

    If we want something more advanced, a new spec should be made in my opinion

  47. lovetox

    But this is beside the discussion because Ms teams also does not allow you to reply to parts of text

  48. lovetox

    And I'm not sure we need to solve that. This is IM. People don't write 100 word messages

  49. MSavoritias (fae,ve)

    heh you are not in the rooms i am in :P

  50. cal0pteryx

    Clients don't have to display the whole message which has been replied to. That's where you click the referred message to scroll to that position.

  51. cal0pteryx

    And yes, sure you can reply to a message (which references it), then quote parts of it manually via ">" message styling, plus pinging people by mentioning them in your reply

  52. singpolyma

    > Sorry if that's the wrong chan for that, > I was thinking about message replies, is there any provision for replying to only part of a message? If not, would it be advisable (although ugly) for clients to split multi lines messages so that distinct lines can be replied individually? Reply with a non-fallback quote

  53. MSavoritias (fae,ve)

    i mean reading the xep its very bare bones and doesnt mention anything about parts of a message so wouldnt it be possible?

  54. MSavoritias (fae,ve)

    to add it i mean

  55. MSavoritias (fae,ve)

    it doesnt even mention what to do with the body at all

  56. MSavoritias (fae,ve)

    the 0461 that is

  57. MSavoritias (fae,ve)

    also second nicoco that not taking threads into account is non-starter and severily limits the feature of this xep

  58. singpolyma

    Cheogram Android defaults to reply starts a thread but allows the user to change that before sending

  59. MSavoritias (fae,ve)

    sure. thats great. it would be nice to mention something in the xep about it :)

  60. nicoco

    > also second nicoco that not taking threads into account is non-starter and severily limits the feature of this xep wow I never said it's a non-starter ;) I don't even have a clear idea of what should be mentioned regarding threads. xep0461 does not mention threads at all, but does not forbid using both a reply and a thread "orthogonally". what cheogram does sounds reasonable to me. with the slight issue that you start a thread from a message that does not have a thread, so you end up in this weird state where the first message of a thread is a reply to a message that is outside the thread. maybe that's not a problem in practice, I don't really know tbh.

  61. MSavoritias (fae,ve)

    i would love to go the zulip route there

  62. MSavoritias (fae,ve)

    everything is a thread by default

  63. MSavoritias (fae,ve)

    its just most messages are the "main thread" and there is no ui for it

  64. MSavoritias (fae,ve)

    that way there is always a thread to refer to

  65. MSavoritias (fae,ve)

    at least thats the way i plan to implement it

  66. singpolyma

    > > also second nicoco that not taking threads into account is non-starter and severily limits the feature of this xep > wow I never said it's a non-starter ;) I don't even have a clear idea of what should be mentioned regarding threads. xep0461 does not mention threads at all, but does not forbid using both a reply and a thread "orthogonally". what cheogram does sounds reasonable to me. with the slight issue that you start a thread from a message that does not have a thread, so you end up in this weird state where the first message of a thread is a reply to a message that is outside the thread. maybe that's not a problem in practice, I don't really know tbh. yes, for clients which prefer a UI showing a thread as "forked off" from a message this is what you want. In Cheogram Android we simply include such message in the single-thread view

  67. rom1dep

    > zulip strikes me as "more noise, for the sake of it", I don't get it nor see the benefits (the UX is pretty much that of creating many rooms, that's not pretty, that's not searchable, that's "in your face", etc) zulip strikes me as "more noise, for the sake of it", I don't get it nor see the benefits (the UX is pretty much that of creating many rooms, that's not pretty, that's not searchable, that's "in your face", etc)

  68. MSavoritias (fae,ve)

    i didnt say to copy the UX :)

  69. cal0pteryx

    Also that's UX, not protocol

  70. MSavoritias (fae,ve)

    you dont have to expose the thread if you dont want to. so its client logic really not UX

  71. cal0pteryx

    Exactly. Nobody forbids replies + threads

  72. rom1dep

    sorry for jumping the guns there, there was no upper bound to what "the zulipe route" might look like :) that said, I'd be cautious with "everything is a thread by default", people not minding threads (whom I expect to be the majority) might too easily break threads accidentally at the expense of those who care

  73. lovetox

    rom1dep: from which client are you writing these messages

  74. lovetox

    I see weird behaviour in C

  75. lovetox

    Seems like you quoting yourself

  76. rom1dep

    lovetox: yours, WS's message replies branch

  77. lovetox

    That's like work I'm progress, just so you know it looks weird in other clients

  78. rom1dep

    oh, wow, I now see how bad it looks from dino 🙂

  79. cal0pteryx

    may I see a screenshot? :)

  80. lovetox

    it sends itself as fallback instead the replied message

  81. cal0pteryx

    that's certainly confusing :)

  82. rom1dep

    https://upload.tamytro.org/upload/a3849b367f4dd4cedd5ad6f64afa7645b7d4f834/ZUf47z19WHFNu6hVyo9N97JzrA9A4xSjcQgGE0tX/b79b5982-a78b-4e4d-9405-e2f25924a1c3.png

  83. rom1dep

    https://upload.tamytro.org/upload/a3849b367f4dd4cedd5ad6f64afa7645b7d4f834/ordxU6RYJq2V90RZrzzhWp5XocRTN0oRwr0r2W9b/2c6bf72a-1d6d-4ba8-a8fc-d455a1f736a6.png

  84. rom1dep

    huh, how come my image paste (from both dino & gajim) appear as urls in the other client instead of as "embeddings"?

  85. cal0pteryx

    room policy

  86. cal0pteryx

    thanks, rom1dep :)

  87. Zash

    how come it doesn't appear that way to you?!

  88. rom1dep

    > thanks, rom1dep :) breaking gajim in novel ways? Sign me in!