XSF Discussion - 2021-02-18

  1. emus

    arc: But the reason is to provide extra payment that Google does not?

  2. arc

    I do not believe that Google supports that. I was suggesting that we use our own funding for Outreachy

  3. arc

    Google has a specific relationship with students and I do not believe their legal structure allows us to commingle with that.

  4. arc

    But we, the XSF, can find Outreachy students working on xmpp projects. And that's where other FOSS organizations are shifting their energy as well

  5. emus

    Ah ok

  6. arc

    Over the last decade we haven't put very much energy into sponsorships. We could very easily secure recurring sponsorships to fund numerous Outreachy students every year.

  7. arc

    At the same time I'm not saying that we should stop doing GSoC. I think it will still have a great deal of benefit even at reduced funding.

  8. arc

    GSoC is specifically open to straight men, which Outreachy isn't. Outreachy targets women, LGBT, and in the US racial minorities. Where GSoC targets college students (or anyone smart enough to enroll for a single community college class) of any race, gender, or sexual identification.

  9. arc

    Given how incredibly male the XSF is, I think participating in Outreachy can do us a world of good.

  10. Daniel

    arc: do you need anything to get started?

  11. arc

    We can use Conversation tasks on https://wiki.xmpp.org/web/Summer_of_Code_2021

  12. Daniel

    Since GSoC has been tied to the sprint this year and sprint has the requirement of teams I can't participate this year

  13. arc

    The Sprint does not require teams. The project's list on the wiki is just for people looking to recruit new developers, vs people working on solo projects

  14. arc

    I am working on solo projects myself.

  15. arc

    And they're not really "tied", I just timed them so that the sprints can help in community engagement during the application period

  16. Daniel

    > Participating projects should be team-based with seasoned developers It's not?

  17. arc

    That defines the list on the wiki only, so that new developers are not left to figure out which are solo projects

  18. arc

    Would you like to reword that to be less confusing?

  19. arc

    The Sprint is for everyone who wants to participate to work on any xmpp project. The list is for people looking to recruit new developers

  20. marc

    Where is the list?

  21. Zash

    Does anyone remember why the avatar metadata node in XEP-0084 uses the hash of the image as item id, as opposed to "current" like most PEP things?

  22. Kev

    Not me.

  23. Zash

    Ho, it's "current" in https://xmpp.org/extensions/attic/jep-0084-0.5.html

  24. mathieui

    Zash, isn’t that so you can fetch the avatar data by itemid by using its hash? (and also prevent multiple publications of the same image with different ids)

  25. mathieui

    (but I haven’t seen that discussion so I am only speculating)

  26. Zash

    mathieui: for the data itself, sure. but the metadata carries that info in an attribute.

  27. mathieui

    ah yes

  28. mathieui

    maybe because it was somehow easier if they are the same? idk

  29. Zash

    so you end up duplicating that hash

  30. mathieui

    Zash, "(the match between the ItemID of the data node and metadata node is used by the subscriber to determine if a locally cached copy can be displayed)." from 3.2

  31. Zash

    I may have looked at this before.

  32. Zash

    > Subject: Protocol Action: 'Hypertext Transfer Protocol Version 3 (HTTP/3)' to Proposed Standard (draft-ietf-quic-http-34.txt) hmmm

  33. Zash

    mathieui: if you see the item id, you also see the item itself with the hash(es)

  34. Zash

    Could have made sense if it used notifications without payloads, but then it wouldn't have needed to split metadata and payload

  35. Kev

    *Maybe* some implementations back then lacked notification without payload, or some such argument. But I really don’t remember.

  36. Zash

    Multiple resolutions (which nobody does afaik) wouldn't work that way tho

  37. mathieui

    I hope I have not overstepped my boundaries by editing the applications wiki page to put 2021 instead of 2020, Alex :)

  38. SamWhited

    Is that actual date right? I mean, obviously it was meant to be 2021, but is it March 9th?

  39. mathieui

    Looking at previous years, I am not sure the dates for end of applications and the meeting follow a particular pattern

  40. mathieui

    but since the days are different from 2020, I think the dates are right, only the year is wrong

  41. SamWhited

    oh, good point

  42. Alex

    thanks mathieui‎ 👍

  43. lovetox

    some things in 0084 are just weird

  44. lovetox


  45. lovetox

    yeah so you want to publish a gif

  46. lovetox

    but you need the sha of a png to publish it

  47. Zash

    waaaaait what

  48. mathieui

    yeah that is suspicious

  49. lovetox

    yeah but it breaks nothing, and noone can actually verify the sha

  50. lovetox

    and it does not matter anyway

  51. lovetox

    because you are supposed to pull the gif from http

  52. lovetox

    so you can just use any id you want

  53. lovetox

    and just pretend its the sha of a non existing png

  54. L29Ah

    i wonder what was the reason behind this

  55. Zash

    If it's adding another format, shouldn't that publish look more like https://xmpp.org/extensions/xep-0084.html#example-10 ?

  56. lovetox

    no Zash, you are not forced to publish a png, you can just only use a http method

  57. lovetox

    according to the xep

  58. lovetox

    and in light of the multiple format thingy its even more useless that one chosen image sha is set as id

  59. Zash

    Anyone up for sitting down and rewriting this to do what we all think it's meant to say? :)

  60. lovetox

    i just dont know why its a MUST to set the sha to the item id

  61. lovetox

    the only thing it could be that people back then thought we send pubsub notifications without payload?

  62. Zash

    The metadata item id can just be "current", and contain 1 or more <info> refering to either blobs in the data node or something else.

  63. lovetox

    i know at least one impl that depends on the id beeing the sha 😋️

  64. lovetox

    rightfully so when you read the xep :)

  65. lovetox

    so i dont get my hopes up that this xep will be changed

  66. lovetox

    everybody uses this with one avatar anyway

  67. Zash

    Not me!

  68. arc

    marc: https://wiki.xmpp.org/web/Spring_Relay_2021