XSF Discussion - 2021-02-10


  1. arc

    > Why? Because students need to contribute code in addition to their application, so an event where they can get some mentorship in doing so is helpful

  2. Kev

    FWIW, as a person who wouldn’t have suggested it himself, it sounds like a solid idea to me.

  3. Kev

    Also has advantages in encouraging “XSF” students rather than “Arbitrary Project” students.

  4. Daniel

    i'm afraid that the list of XSF students wont be very long

  5. Daniel

    which doesn’t have to be a bad thing

  6. Daniel

    but we should be aware of that if we make any change that encourage that

  7. Kev

    Well, the ideal would be that we have only as many applications as we end up with slots, and each applicant is perfect, right? :)

  8. arc

    I have a similar concern, especially since the program is effectively halved this year. But also my experience as program admin for different groups has shown that group coding events during the application period greatly increased the number of applications

  9. arc

    That'd be ideal Kev 🙂

  10. arc

    Also you need to consider that it is not just about picking the best students and watching them work. We are a good part of that equation

  11. arc

    And part of this change from Google's side will also increase the number of thoughts overall

  12. arc

    https://wiki.xmpp.org/web/Spring_Relay_2021 still needs sprint topics

  13. SamWhited

    arc: added my project and myself, thanks. I'll probably work on some MIX stuff since I recently needed it for something.

  14. Daniel

    how is the sprint going to be advertised to 'outsiders'?

  15. arc

    I was going to post it with some groups that I know

  16. arc

    I also assume that participating projects would be advertising with their communities

  17. Kev

    And presumably it would be in our org description on GSoC, same as we always do for our teaser task requirements?

  18. Holger

    BTW, if anyone happens to be interested in discussing OX (as in OpenPGP for XMPP), we're going to have a meetup on the topic in half an hour: https://xmpp-meetup.in-berlin.de/4d69d3cc-f28f-465a-baf2-0696024dc9ef

  19. arc

    Yeah, in addition to the normal list

  20. arc

    https://wiki.xmpp.org/web/Summer_of_Code_2021 is for the main ideas page

  21. arc

    We have not decided if you want to apply as an organization yet

  22. Daniel

    why wouldn’t we

  23. Daniel

    We have volunteers to act as admin

  24. arc

    Well the last time I looked we didn't have any projects signed up on the wiki page

  25. SamWhited

    I didn't know there was a wiki page, will think of project ideas. I think I'd like to participate this year

  26. MattJ

    What's the deadline? If it's not before the end of next week I can probably add some stuff

  27. arc

    I'm somewhat distracted this morning though by baby chickens nipping at my beard https://photos.app.goo.gl/wLbBP14CzxATb6Ud8

  28. arc

    We need to get the application in by this time next week

  29. SamWhited

    daww, look at that baby face. Oh, and also there are some birds.

  30. arc

    Lol thanks

  31. arc

    What other topic should we cover in the Sprint?

  32. arc

    MIX is the big one for me too, it would be absolutely excellent to get make support extended

  33. SamWhited

    Oh weird, that does say MIX, somehow I read that as MAM. I was just looking at an issue regarding MAM so I guess my brain crosswired them.

  34. SamWhited

    I was thinking about working on a *MAM* implementation during the sprint and maybe a MIX implementation if I can get someone to do Google Summer of Code or what not

  35. wgreenhouse

    Zash, MattJ: does `clix` (and/or Zash's upstream review version) still require lua5.1?

  36. Zash

    In theory no. In practice, it's the least painful path.

  37. wgreenhouse

    Zash: okey, thx.

  38. arc

    SamWhited: why not add MAM to the list?

  39. SamWhited

    arc: still trying to decide if that's what I want to do, but I will if I figure it out.

  40. arc

    I have been working on xmpp microservices, and I don't have a MAM microservice yet so that would be interesting to me

  41. SamWhited

    Realistically the thing that would provide the most value is if I finished doing disco, but I get frustrated and stop working on it every time I try for <reasons>

  42. arc

    Lol yeah

  43. arc

    MAM is also part of MIX. Everything is inter-related

  44. SamWhited

    And if I work on MIX I'll have to work on pubsub, which will make me cry too. I don't think I've ever made it through that XEP.

  45. arc

    I know that feeling. But if we just focus on one thing every week, with focused discussion on that topic during that week, we can get a lot done

  46. arc

    Part of the reason I'm working on the microservices model is we have a lot of software that is only partially supporting the annual compliance stack

  47. arc

    But, is relatively simple to add support for microservices. So even a traditional XMPP server should be able to just add support for the microservice protocol to add major extensions

  48. Zash

    What microservices protocol? Like XEP-0355?

  49. arc

    Just browsing over that xep, looks very similar

  50. arc

    I will have to read further and get back to you on that.

  51. arc

    But I have been working on this since 2009

  52. arc

    It has become a fundamental part of the concordance stack when I moved the s2s stanza router out of the Python domain. Back then the python code ran as a subprocess akin to cgi

  53. arc

    By 2014-2015 it went further with cloud-ready microservices handling client and server connections, authentication, registration, etc. And the stanza router largely a dumb service that just routes stanzas between the microservices https://drive.google.com/file/d/0B-mQgeagTilHZWN4M0ZlaEwyNVU/view?usp=drivesdk

  54. arc

    That's when EXI became part of the stack

  55. arc

    That was when Atlassian hired me for Hip chat, Holden picked me up to work on his xmpp startup, and grindr was head hunting me for a bit. Dodged all of those bullets.

  56. arc

    But then I also realized the exi work that I was putting into all this was worth a thousand times what the xmpp work was worth, so the xmpp work became largely a side project

  57. SamWhited

    I'm jealous that you can make any of your XMPP work (or tangentially related to XMPP work) worth anything at all

  58. arc

    I have sold way more licenses for exi over the last 5 years than anything else.

  59. arc

    Mostly Asian IoT.

  60. arc

    But to be honest, I'm fairly certain that is because I don't release the source code openly. I bundle it with an integration service. If I were to release under the AGPL like I intended, I'm fairly certain I would have an impossible task of suing the thousands of IoT manufacturers who really don't care about the license.

  61. arc

    And I don't make enough money to hire legal firm.

  62. jonas’

    ah hm, that explains why I hear a lot about exi from you but never found a library to use

  63. arc

    Why would I give up my source of perpetual rent money? 😆

  64. arc

    And truthfully there isn't "a" library. All the hard part is in the schema compiler, which by design generates a library. EXI is simple enough that you don't really need a library once you understand it.

  65. moparisthebest

    ha, sounds like famous last words, X isn't hard once you understand it :)

  66. arc

    If they're were a library that you fed the schema/grammar into and give you a dynamic API to use it, thousands of these fly by night companies would use it without even giving you credit. They already don't give xmpp credit or see that they are using exi

  67. moparisthebest

    substitute what you will for X, TLS, TCP, x86 bootloaders

  68. arc

    That is very true. But you have to realize a lot of these companies only exist for a single production run of a few million devices. Then another company's name is put on it, usually without giving that company the source code, and is sold under their brand.

  69. arc

    The entire economy is a massive shell game of fly by night companies and abandonware.

  70. moparisthebest

    software companies in general right?

  71. arc

    American startups tend to act a little more responsible. I'm really talking about the Chinese, Japanese, and South Korean tech manufacturing companies.

  72. arc

    They specifically set themselves up like this in order to protect themselves against consumer lawsuits. But it really protects him against all forms of lawsuits. Their devices could burn down homes and offices, explode in people's laps, or leech poisons into their food, and it's nearly possible to find anyone who's legally responsible. So software licensing is the least of their concern.

  73. moparisthebest

    please send me your money, here have a look at my whitepaper.pdf

  74. arc

    It would be really nice if they offered a white paper 😅

  75. arc

    Without naming names, I was briefly working for a startup in San Francisco that was aiming to sell turn-key IoT websites for these kinds of companies. A service specifically tailored for fly by night manufacturers. XMPP was one of the protocols they wanted to support, along with all the usual suspects. And the service was a per-device one time fee that the manufacturer could pay up front. It was sketchy enough to drive away American investors.

  76. arc

    But the good part was that they offered documentation. Even though the turnkey websites were embedded in the device firmware.

  77. arc

    I don't think it serves us as a community for users to even know that they're cheap generic brand wifi light they bought from Amazon runs xnpp. I suspect that the shoddy mobile app they installed and then immediately deleted to connect the devices to their Amazon Echo or Google Home has some almost-hidden menu option for the licensing and protocol documentation.

  78. emus

    Hi Kev, I wanted to ask for your email contact to request the access for SamW. to Twitter