XSF Discussion - 2019-10-19


  1. Daniel

    question to the people who are longer involved with the XSF then I am; why are XEPs (usually) created by individuals instead of working groups?

  2. Daniel

    i mean some (few) XEPs are written by more than one person; but we don’t seem to have the tradition of working groups like the ietf has

  3. fippo

    most successful IETF work starts from individual drafts too

  4. fippo

    the closest the XSF has to a WG is a SIG and we haven't really had them for a while

  5. Daniel

    i have a vague feeling (without knowing a lot about how the IETF works) that working groups could be more focused (regular meetings) and also more "democratic" as it would be a group of multiple people who have to have a rough consensus on the spec instead of just one author who has ultimate say in how things are written down (and then council basically just being able to say yes or no when it moves on)

  6. Daniel

    but i might might be wrong. i just noticed that the xsf process mirrors a lot of the itef process; except that part

  7. Daniel

    and i'm wondering if there was a reason for that

  8. flow

    IETF working groups also usually have (more or less) a single person who tries to determine the consensus and writes it down in the I-D. That job is not easy, at least sometimes, just look at the OpenPGP WG and 4880bis. It is also what I would expect from a XEP author, but as you mentioned, I also have the feeling that this is not how it currently works in XSF

  9. flow

    When I worked on OX (XEP-0373, -0374), I set up a public monthly meeting and announced it. There was also a git where public could send patches to, which got regulary build so that everybody could view the current state of the ProtoXEP

  10. flow

    I wonder if we should do the same with message fastening

  11. Link Mauve

    “18:16:46 Zash> C) Steal Matrix thing where every server downloads and re-hosts every uploaded file?”, that’s prosody-modules’s mod_bob.

  12. Link Mauve

    I’m still not sure how much of that is a good idea.

  13. Link Mauve

    Ah, you mentioned it afterwards.

  14. Zash

    Link Mauve: It being in-band is limiting

  15. Zash

    Fine for stickers I imagine tho

  16. Link Mauve

    Yes.

  17. Link Mauve

    For a further iteration of bob, we might want to hook into 0261/0047 at least, to support more than one stanza.

  18. Link Mauve

    “12:06:48 flow> pubsub/pep experts: Is it normal that I receive a notification twice on publication: http://paste.debian.net/1107636/ ?”, I have experienced a similar bug with Ejabberd when I have multiple resources subscribed to a PEP node, Holger told me it’s a known bug.

  19. pep.

    Daniel, flow: I think that's something we could improve yes. I don't especially like that only one person ultimately has a say and don't accept your change because they haven't had coffee yet, or more realistically because they have their own agenda and don't care about the rest

  20. Kev

    One thing to note is that the authors don't 'own' the XEPs, the XSF does. So the authors don't hold a XEP hostage.

  21. pep.

    Yes technically the council could "take over", but I'm not sure how well that would be accepted socially

  22. pep.

    I haven't seen this happen during my short time being involved, apart for clearly abandonned XEPs, and then it was really "can we really do this?"

  23. MattJ

    The counter to this is that it's also good to have one person ultimately responsible for getting the doc done and resolving any disputes

  24. jonas’

    I see my role as a XEP author mainly as proxy between the community and the editor, once it has been accepted.

  25. jonas’

    I see my role as a XEP author mainly as proxy between the community and the editor, once it has been accepted as Experimental.

  26. jonas’

    This means however that I select which voices from the community I listen to though

  27. led

    If the council votes, do they give any kind of feedback about why the have took their position?

  28. jonas’

    led, when veto-ing, a council member is obliged to give a rationale

  29. jonas’

    those should be included in the council minutes

  30. jonas’

    and they are also included in the voting summaries Tedd Sterr sends

  31. led

    Ok that's good then.

  32. jonas’

    and of course you can always ask

  33. jonas’

    we’re not like the City Council, we don’t devour citizens.

  34. flow

    MattJ, the counter to what exactly?

  35. flow

    As far as I understood it, nobody disputes that there should, and maybe even has to be, one person chairing the process. I just think we can do better…

  36. Ge0rG

    We hardly have enough persons to just write down what they think is best, never mind to have a democratic XEP design process

  37. pep.

    https://xmpp.org/extensions/xep-0002.html what does it mean to be a SIG leader

  38. Zash

    pep.: Like a chair I guess?

  39. pep.

    so...?

  40. pep.

    XEP-0001 defines "Approving body Chair", well it doesn't. It only defines what a chair should do

  41. pep.

    "The precise rules for polls, including meetings and timeouts, are determined by the Chair of the Approving Body."

  42. Zash

    pep.: I was looking for https://xmpp.org/extensions/xep-0019.html

  43. pep.

    ta

  44. pep.

    So what, does that prevents anybody from creating SIGs anymore?

  45. pep.

    Do we need a XEP to revert this change

  46. pep.

    Do we need a XEP to revert this change?

  47. Daniel

    > pep.: I was looking for https://xmpp.org/extensions/xep-0019.html That is a good read. That kinda explains my earlier question

  48. Zash

    So, nothing prevents you from forming your own, as '19 says, small ad-hoc group to work on a proposal.

  49. pep.

    Sure that people can already do.. it's just to be transparent and stuff

  50. pep.

    Also https://xmpp.org/extensions/xep-0381.html

  51. pep.

    And it's still proposed, orz..

  52. pep.

    "The Last Call ends on 2017-03-14"

  53. Zash

    Heh

  54. Zash

    xeps$ grep status.Proposed xep-????.xml xep-0292.xml: <status>Proposed</status> xep-0300.xml: <status>Proposed</status> xep-0345.xml: <status>Proposed</status> xep-0353.xml: <status>Proposed</status> xep-0363.xml: <status>Proposed</status> xep-0381.xml: <status>Proposed</status>

  55. !XSF_Martin

    345 got an extra blank?

  56. Zash

    Indentation moves in mysterious ways.

  57. DebXWoody

    Groups can also be build based on components. .e.g. security, mobile, processes, performance,... maybe a kind of XEP Security Group which are working on security,... Just an idea. But it is important that there is one person, which is taking care about the activities with a group.

  58. pep.

    <xs:enumeration value='Organizational'/> < apparently there's no type Procedural.

  59. pep.

    Good thing schemas are not normatives..

  60. Zash

    Ops

  61. Zash

    Heheh

  62. pep.

    That's really sad

  63. pep.

    https://github.com/xsf/xeps/pull/846 here.

  64. pep.

    Technically schemas are not normative, that's editor land.

  65. pep.

    Oh, and revision block.