XMPP Council - 2017-09-27

  1. Tobias

    It's about time

  2. Tobias

    1) Roll call

  3. Tobias

    daniel, SamWhited, ping

  4. SamWhited


  5. Tobias

    Link Mauve is excused, Dave just missing

  6. daniel


  7. Tobias


  8. Tobias

    2) Minute taker

  9. Tobias

    jcbrand, are you available?

  10. jcbrand

    Yes I'm available

  11. Tobias


  12. SamWhited

    jcbrand: Thanks! It's nice having an editor that can take minutes on occasion again :)

  13. jcbrand

    I'm happy to help, just keep on pinging me please :)

  14. Tobias

    3) Vote on "XEP-0060: Add pubsub#multi-items in Publish-Subscribe features #500" https://github.com/xsf/xeps/pull/500

  15. Tobias

    I'll vote on list

  16. daniel

    On list

  17. SamWhited

    On list

  18. Tobias

    4) Advancing "XEP-0286: Mobile Considerations for LTE Networks"

  19. Tobias

    do we want to have the editors issue a last call?

  20. dwd

    Ooops, sorry I'm late.

  21. Tobias

    i'd be in favour

  22. dwd

    I'd be in favour of advancing XEP-0286.

  23. SamWhited

    I am obviously biased, but I think we should issue a LC and at worst I'll get feedback from it even if it doesn't advance :)

  24. dwd

    SamWhited, I'm biased too.

  25. SamWhited


  26. Tobias

    daniel, any opinion on this?

  27. daniel

    No. I don't. Let's issue a last call and see what comes up

  28. Tobias


  29. Tobias

    5) Discuss renaming "Draft" to "Stable" and make a recommendation to the board

  30. dwd

    The recommendation can be a PR to XEP-0001, incidentally. I'd note it'll mean regenerating all the XEPs, too.

  31. SamWhited

    All XEPs are regenerated twice a day or something now, so regenerating XEPs shouldn't be a problem.

  32. Tobias

    Stable/Stablish sounds quite similar to Final

  33. SamWhited

    (Just in case it wasn't clear, I was joking about "Stablish")

  34. dwd

    Tobias, In fairness, Draft *is* quite similar to Final.

  35. SamWhited

    For me anyways that's the idea. When I mention that I work on XMPP to people, the first thing I always get at meetups and the like is "Oh, we tried that, but every feature we wanted to use was just a draft"

  36. SamWhited

    So making them sound similar might help.

  37. dwd

    My problem with us discussing it is by the time we've got to Council we (should) understand what these things mean, and are no longer in the mindset of those who would benefit from a change in name.

  38. daniel

    I'm not sure what this change will actually accomplish. I have never personally had the experience SamWhited describes where someone didn't implement it because it was just a draft

  39. daniel

    But if that's a real thing than I don't have any issues with this change

  40. Tobias

    yeah...i'm not going to block this change, but other things might help more with this

  41. dwd

    I think Board might have a better chance of having a valid opinion than we do.

  42. daniel

    It's just a cosmetic thing anyway. It doesn't change any real problems we might have like not advancing xeps quickly enough and so on

  43. SamWhited

    Agreed; it is no substitute for fixing our process.

  44. Kev

    I think before doing something disruptive like this, there should be some real investigation into who it'll help, and who it'll hurt.

  45. Tobias

    maybe add a standardized table to each XEP of known implementations or such, if people see it's implemented X times, it might be more attractive even though it's calle Draft/Experimental

  46. SamWhited

    That sounds like a lot of work to keep up. Have we ever been known to keep client lists, wiki pages, etc. up to date?

  47. Tobias

    SamWhited, IIRC there were community fplk asking on how they could help, not?

  48. dwd

    (I still think the best marketing idea I've had for XMPP is to say it has a React-like wire format instead of saying XML)

  49. SamWhited

    Tobias: Yes, but communities ask how they can help and then dissapear after two weeks and it becomes neglected again.

  50. Tobias

    also there are different issues at play here: a) having XEPs early accepted so they can be easily referenced and discussed, b) having quite stable XEPs (MUC) idle around in draft state for 10 years, c) name doesn't sound attractive

  51. Kev

    To me "Oh, we didn't use XMPP because the specs are Draft" doesn't sound particularly convincing. It might actually be true that someone read just one word in a XEP, and none of the others, but it seems a stretch.

  52. jonasw

    I’d like to throw in that Link Mauve has proposed an XML format to describe XMPP software, including XEP support

  53. jonasw

    we also already have the requirement for software we list on the website to renew yearly

  54. Zash

    Reactive Messaging and Presence Protocol?

  55. jonasw

    we can step by step increase the requirements

  56. jonasw

    and require them to list XEPs they implement

  57. jonasw

    that’d give us the nice table Tobias wants (and which I think would be amazing to have)

  58. Tobias

    jonasw, yeah but that's a different issue iI think?

  59. Tobias

    jonasw, yeah but that's a different issue i think?

  60. Kev

    I'm opposed to opening up XEPs for marketing purposes, FWIW.

  61. SamWhited

    Yah, let's discuss one at a time unless someone has a reason these issues are related and need to be discussed together?

  62. Tobias

    SamWhited, +1

  63. daniel

    anyway let's not drift away from the topic. so regarding the recommendation on changing draft to stable I suggest "we don't have one. let the list discussion play out for a little while longer. and/or let board investige real world benefits of changing the name"

  64. jonasw

    indeed, it’s not related to the name change. but Tobias brought that table up, so I wanted to bring the thing back into mind. nevermind me.

  65. Tobias

    daniel, +1 for continuing the list discussion

  66. SamWhited

    That sounds sensible to me; let's let the board discuss and see if they can think of a way to investigate further

  67. Tobias

    jonasw, yeah...could be integrated dynamically on XEPs, like you open a XEP and can expand a table that'll show existing implementations, but yeah...something for another time

  68. Kev

    This is a marketing question, and so Council probably isn't the right place anyway (although their opinion should certainly be noted).

  69. Tobias

    naming XEP states is a marketing question?

  70. Tobias

    6) Date of next

  71. dwd

    Tobias, XEP-0001 is a Board XEP, so it's not Council.

  72. Kev

    The reason for the rename is based on the premise that XMPP is hard to market because of the name of one state, yes.

  73. Tobias

    dwd, ahh..alright

  74. Tobias

    well..same time next week?

  75. Kev

    I don't think anyone is going to claim that a XEP would be easier to implement with the name of the state changed.

  76. SamWhited


  77. Tobias

    7) AOB

  78. jonasw

    I’d like to throw in that there are still pending list votes from last week :-)

  79. SamWhited

    We have two outstanding things on our Trello board, ODR/OMEMO and XEP-0280. I don't know the status of either, but should we discuss them?

  80. jonasw

    (regarding the LC proposals and the two ProtoXEPs)

  81. daniel

    odr/omemo can be removed

  82. daniel

    that's a list thing

  83. daniel

    not a council thing

  84. Tobias

    There are a bunch of old probosed agenda items in Trello, which I think are all stale or at least in an unknown state. It would be great to have them updated.

  85. dwd

    jonasw, I'm no objection on colors and jet, FWIW. I'll do the rest on list (and these two if you don't get to them)

  86. Tobias


  87. Tobias bangs the gavel

  88. Tobias

    thanks everybody

  89. jonasw

    then I’ll see that colors and jet get accepted soon-ish.

  90. jonasw

    thanks, dwd & all

  91. Ge0rG

    I think 0280 was in LC and got delayed/deferred because of the rule complexity

  92. Ge0rG

    And because of issues with <no-copy/> vs. <private/> that never quite got sorted out.

  93. daniel

    Ge0rG: the <transient/> annotation did have consensus thought, didn't it?

  94. daniel

    One marker to rule them all

  95. daniel


  96. Ge0rG

    daniel: I don't think so.

  97. daniel

    Mhh. it's hard to keep up when the discussion happens in three different channels

  98. Ge0rG

    My agenda was to create a set of rules that work for IM, but underway I realized that I'm not competent enough to enumerate all affected user stories and XEPs.

  99. Tobias

    daniel, indeed

  100. Tobias

    in the past it was just ML + chat room, now it's ML + chat room + PR

  101. Ge0rG

    I'm in the process of refining the problem statement, as discussed in xsf@ MUC today, in the hope to provide food for a Summit discussion.

  102. Ge0rG

    My preferred venue would be my private blog, making #4 on the list of things to follow.

  103. jonasw

    Tobias, I’m working on avoiding PR

  104. Tobias

    jonasw, ta

  105. jonasw

    i.e. directing technical discussions to the mailing list

  106. daniel

    the muc still doesn't have history, does it?

  107. Ge0rG

    Because if I write long posts to standards@, they are either ignored or individual points are picked out and get discussed at length :>

  108. Tobias

    daniel, not yet

  109. Ge0rG

    I'm not sure how often I've brought up Carbons on the ML so far.

  110. daniel

    and there is stuff on the wiki

  111. Tobias

    iteam wants dockerization but not keys in prosody docker...maybe dwd's Metre can help there

  112. daniel

    finding the right place to discuss how to route messages is even more complicated than to acutally route messages

  113. Ge0rG

    daniel: sad but true

  114. SamWhited

    Tobias: Isn't that what environment variables are for?

  115. Tobias

    SamWhited, ?

  116. Ge0rG

    I'm open for proposals, though.

  117. Tobias

    SamWhited, passing keys through envvars?

  118. SamWhited

    Tobias: For not putting secrets directly in a config file (a Dockerfile in this case)

  119. Tobias

    IIUC other iteams didn't want the prosody docker to have access to the LE private key at all, might have misunderstood them though

  120. Kev

    Yes, they can have access to the key.

  121. Kev

    Not sure how Prosody's going to work over TLS without it.

  122. daniel

    i wonder if discourse has support for patch files. if so we could run on discourse only and get rid of github and mailman

  123. dwd

    Tobias, No. Well, sort of but we shouldn't, we should be using something like an SHCM.

  124. Kev

    Just that the key should live outside the container, and be linked in via a mounted volume.

  125. Ge0rG

    daniel: and of XMPP.

  126. Tobias

    dwd, SHCM?

  127. SamWhited

    daniel: I've never been able to convince discourse to work well as a mailing list. I'm sure it's nicer to maintain or nicer if you use it as forums via the web UI, but unless you know a secret I don't it's terrible for mailing lists as far as I can tell.

  128. SamWhited

    Every few months I try to sign up for the Rust lists again (they're on discourse) and it's always a rough experience.

  129. daniel

    SamWhited, oh i don't have personal experience with it. all i can say is that mailman doesn't work at all

  130. SamWhited

    yah, mailman is pretty terrible.

  131. daniel

    because it splits threads into month and doesn't provide a search

  132. jonasw

    maybe mailman2 is a compromise?

  133. jonasw

    I heard it has a flashy web interface

  134. Ge0rG

    daniel: you can search your offline copy ;)

  135. daniel

    yes i can. but than i cant link to it

  136. daniel

    yes i can. but then i cant link to it

  137. Ge0rG

    daniel: but you can open the web view of the respective month and manually find the message there :P

  138. SamWhited

    I can't remember what it's called, but someone told me that Fastmail has a hosted mailing lists product recently. I would probably be willing to just pay for that and donate it to the XSF.

  139. daniel

    that's my workflow. search offline; and then find the mail online

  140. jonasw

    I agree with daniel that that workflow is annoying

  141. Ge0rG

    daniel: that's my workflow as well.

  142. SamWhited

    (I have no idea if it's any good or not, but their email is great so I assume their other products are too)

  143. jonasw

    but my discourse experience (also with rust) as mailing list is wrose.

  144. Kev

    My experience with discourse is that it's pretty good as long as you dedicate a good portion of every day to it, but any less than that and it's a non-starter.

  145. Kev

    (Through the web interface)

  146. Ge0rG

    the workflow is annoying, but at least you can search the ML archive locally when offline, and you don't need to interact with a horribly bloated Web 2.0 application.

  147. jonasw

    Kev, that was my feeling too when rust moved to discourse

  148. jonasw

    it is just another web forum, really

  149. Ge0rG

    IIRC somebody proposed to provide IMAP read-only access with search to our MLs.

  150. SamWhited

    It seems fine if you want forums, but not if you want mailing lists.

  151. jonasw


  152. Tobias

    https://lists.mailman3.org/archives/list/mailman-users@mailman3.org/ this doesn't look that bad

  153. daniel

    personally i could live with not having mail access. if that's the main issue. but i understand that people like their mutt or what ever

  154. jonasw

    I need the push of email

  155. jonasw

    I don’t want to poll a website every day

  156. jonasw

    or every few hours

  157. SamWhited

    Alternative: Google Groups works well as a mailing list and is free.

  158. SamWhited

    And has a web UI.

  159. SamWhited

    Or Yahoo or whatever, I don't know the difference.

  160. jonasw


  161. daniel

    i see. i personally prefer pull for these things. i don't like it interupting what ever else i'm working on

  162. jonasw

    LTIC I couldn’t post to a google group because google doesn’t like the network my mailserver is in.

  163. jonasw

    even though the prosody group worked fine back then

  164. Tobias

    https://lists.mailman3.org/archives/list/mailman-users@mailman3.org/thread/YWYSIBYMNRZC5KLJHNKJEYPIXKXYNEXA/ also lloks much nicer, but they decided for unreadable URLs apparently

  165. jonasw

    so maybe it’s an option

  166. Ge0rG

    mailman3 looks like a social network to me. But at least it allows permalinking to individual items.

  167. Tobias

    Ge0rG, perma but not readable

  168. jonasw

    Tobias, you can only have one, I think ;-)

  169. daniel

    maybe we can bring back google wave

  170. Ge0rG

    Tobias: apparently what you are linking to is not mm3 but HyperKitty, "a Django-based application providing a web interface to access GNU Mailman v3 archives, and interact with the lists"

  171. daniel

    i feel like that was just ahead of it's time

  172. jonasw

    Ge0rG, it is the more or less official mailman3 web interfae

  173. SamWhited

    I think it was just a marketing problem… it tried to do every damn thing and I never could figure out what the point of it was.

  174. Tobias

    jonasw, you can put both unique and human reable data in URLs

  175. Ge0rG

    daniel: I think it was created right at the end of Google's "don't be evil" and "support decentralized services" era and thus was killed quickly.

  176. SamWhited

    It was vaguely kind of cool, but I never could figure out a workflow or how I was supposed to use it.

  177. SamWhited

    > Drafts XEP on Spam prevention:

  178. SamWhited

    *lists only deferred XEPs*

  179. SamWhited


  180. SamWhited

    This is why "Draft" status is confusing.