Re: MediaGoblin community meeting - 8am Sat May 16 UTC-4

[Christopher Lemmer Webber]

Here's the agenda and logs from the etherpad / irc:

  https://etherpad.wikimedia.org/p/mediagoblin-2020-05-16

MediaGoblin community meeting May 16 2020

Agenda

Intro/Housekeeping:
 - Chris scribe?
 - thanks for coming - big committment, sorry to Andrew and west coast US folks
 - 1 hour only!
 - around the group - who, where and and how you're coping (15 sec)
 - record future meetings? (requested by JP Badger)

Current situation:
 - release 0.10.0 "We're still here" is out, well done all!
    - helped several people upgrade via IRC
    - posts to mailing list, new Trac issues
    - a few mentions by tech bloggers
    - MediaGoblin still means a lot to many people
 - current accute issues:
    - After upgrade to 0.10.0 restarting marks all items as "failed" 
(https://issues.mediagoblin.org/ticket/5608 )
    - Upgrading with third-party plugins according to new upgrade docs causes 
missing plugin issues - disabling causes migration issues

Upcoming:
 - timing for 0.11 minor release "showing momentum" - July-ish?
 - nominations for priorities in 0.11.0
   - start culling Python 2 code?
   - new audio spectrograms module! (new contributor Fernando Gutierrez)
 - bug squashing day (Amelia Rose)
 - infrastructure
   - CI/Docker (Michael)
   - wiki (Ben)
   - IRC - are we excluding casual drop-ins by requiring registered nick?
   - new "forge" - no news yet, but existing systems improved
   - issues/Trac
      - still a teenie bit intermittent
      - new issues and discussion somewhat invisible to the mailing list?
   - Read the Docs
      - removed "latest" to avoid confusion (master/stable/v0.10.0 etc)
      - documentation needs to be built automatically
      - maybe drop "stable"?
  - packaging - chatted with Jonas Smedegaard regarding his Debian packaging 
work
 - Ben's own focus
   - providing a response on IRC and Trac, resolution may take longer
   - enabling and connecting contributors
   - docs - helping people help themselves
   - minor improvements to build momentum 

Long-term/strategic/reflection:
 - what are the problems we're trying to solve?
 - what does success look like for MediaGoblin?
 - what's the landscape looking like now? What are our opportunities for 
impact/collaboration?
   - allies: PeerTube, DTube, BigBlueButton
   - others: Mastodon, Wordpress, Skuttlebut, Open Broadcaster Software
   - deployment: Sandstorm, Docker, RasberryPi
   - browser APIs have significantly improved
 - what small-medium sized features would make MediaGoblin significantly more 
useful?
  - a more modern/immersive gallery (Shtrom) eg. 
https://www.thregr.org/~wavexx/software/fgallery/ or 
https://github.com/xieranmaya/blog/issues/6
   - batch uploads via web interface (@stigatle)
   - chunked uploads/progress bar
   - mobile image uploads/"share with MediaGoblin" (@stigatle)
   - oembed so you can drop a media item into eg. a WordPress post
   - multiple images per media item eg. two photos of same piece of art
   - private mode (Shtrom)
   
   
<@dustyweb> (starting with intros to give people some time to show up)  [08:02]
<@dustyweb> sturm: I'm ben, I'm in australia, I'm an hour outside of melburne,
            we're doing pretty ok with the pandemic though I do find it
            tiring!  had a lovely sunny day, feeling positive
<@dustyweb> breton: Hey, I'm Boris, I mostly worked on video and audio and
            some other stuff on MediaGoblin... I started in one of the GSoCs,
            then mentored a few  [08:03]
<@dustyweb> breton: I currently am finishing my work on openstack, my company
            is moving from openstack to kubernetes
<@dustyweb> sturm: where are you based currently?
<@dustyweb> breton: german, but please don't try to speak german to me :)
*** seanrobert (~seanrober@2601:680:c001:72f0:55e0:4692:4481:1824) has joined
    channel #mediagoblin  [08:05]
<sturm1> chris: co-founder and technical lead for some time, worked on
         ActivityPub standand with Jessica, ActivityPub became a big thing and
         adopted by many projects, irony don't yet have ActivityPub support in
         Mediagoblin. Busy with standards side of things, next gen federation
         involving Spritely. Happy Ben has stepped in and Simon too, lots of
         people have contributed over time.  [08:06]
<@dustyweb> mtlynch: I'm michael lynch, I discovered mediagoblin... I was
            trying to create a private media server for my family... part of
            that involved getting that to run under Docker, has been kind of a
            long process because there have been breaks due to third party
            projects  [08:07]
<@dustyweb> mtlynch: but I've been working on docker support  [08:08]
<@dustyweb> mtlynch: also I'm in western mA!
<@dustyweb> chris: Oh I didn't know!
<@dustyweb> mtlynch: also I'm working on documenting on my setup, has a bunch
            of information, and it costs only 70 cents a month!
<@dustyweb> sturm: wow!
<@dustyweb> mtlynch: hope it catches some attention on HN/reddit/etc
<@dustyweb> rigelk: I'm RigelK (or at least that's how chris pronounces it).
            I'm part of the PeerTube project, I'm interested in what the
            future of mediagoblin looks like, especially the federation part
                                                                        [08:10]
<@dustyweb> rigelk: we've been easily anticipating that and connecting via
            federation  [08:11]
<@dustyweb> fergtm: I'm Fernando... don't have much to say, have been a
            MediaGoblin users for two years, just recently upgraded to a new
            version, saw a few things that I could help with  [08:13]
<@dustyweb> sturm: absolutely you did, for those who didn't see there was a
            patch to add back audio spectograms in python 3 and I didn't
            expect that to be back so soon
<@dustyweb> seanrobert: I am a Utahn. Just started looking at working on Media
            Goblin. I am still mostly just getting the lay of the land on the
            code  [08:14]
<@dustyweb> sturm: I spoke to Sean a couple of days ago in the chat, he just
            graduated and was looking at projects to get involved in, was
            seeing if this was a good project to get involved in
<@dustyweb> seanrobert: I thought I would tune into this meeting to see if I
            might learn something
<@dustyweb> sturm: thank you to everyone coming out this weekend, sorry to
            everyone who couldn't make it for timezone reasons
<@dustyweb> sturm: we're going to keep it to max 1 hour, no limits  [08:15]
<@dustyweb> sturm: jpbadger posted on the mailing list asked if we could
            record future meetings
<@dustyweb> sturm: does anyone feel whether we should or shouldn't record
            things?
<@dustyweb> q+
<@dustyweb> sturm: does anyone have strong feelings?
<@dustyweb> cwebber: suggested protocol: announce that we'd like to record
            audio in next meeting announcement, then gather consent at start
            of next meeting, and if anyone shows up late and doesn't consent
            ask them to be quiet because we've already agreed  [08:18]
<@dustyweb> sturm: so!  first topic: we got out our 0.10.0 release, and we've
            seen a lot of excitement since that  [08:19]
<@dustyweb> sturm: we've gotten some tech places catching our release and etc
            in their posts
<@dustyweb> sturm: and goes to show that mediagoblin means a lot to people in
            the sense of the software and moving towards an independent and
            accessible web for people that respects people
<@dustyweb> sturm: so congrats to everyone for the work they've done so far
<@dustyweb> \o/
<@dustyweb> sturm: I've noted a couple of current issues that are of interest
                                                                        [08:20]
<sturm1> https://issues.mediagoblin.org/ticket/5608
<@dustyweb> sturm: one here is there's an issue where after the upgrade that
            there are issues with marking items as failed
<@dustyweb> sturm: I hit a similar-ish issue, I'm not sure what the
            commonality is  [08:21]
<@dustyweb> sturm: second issue is I've modified upgrade instructions to
            bridge the gap that we've had in terms of time since the last
            major upgrade, I wanted to make it as reliable as possible, so the
            instructions I've given on the release are remove your virtualenv
            and re-run your configure-make   [08:22]
<@dustyweb> sturm: works ok but obviously blows away any third party
            mediagoblin packages installed... that's to say that's not a
            perfect way to upgrade yet
<@dustyweb> q+
<sturm1> Chris: idea has always been to move people towards system packages -
         moving towards that  [08:24]
<@dustyweb> shtrom: hellow, showed up a bit late  [08:25]
<@dustyweb> sturm: so for 0.11.0 release, I want to do something not too
            complicated but another point in time so that there's confidence
            the project is moving consistently  [08:26]
<@dustyweb> sturm: so that's what I'm thinking is gradual progression release
<@dustyweb> sturm: thoughts?
<@dustyweb> breton: I agree with the priorities on the etherpad
<@dustyweb> sturm: I just added a couple of priorities there, one thing I
            added considering was start killing python2 code, and add back
            spectrograms  [08:27]
<@dustyweb> sturm: I'm not sure about whether or not we remove python 2?
<@dustyweb> q+
<@dustyweb> oops
<@dustyweb> breton: I think we can just leave it  [08:28]
<@dustyweb> cwebber: say in the next release that the one after will deprecate
            python 2  [08:29]
<@dustyweb> sturm: yes we already announced that the next release won't
            support py2  [08:30]
<@dustyweb> breton: I think we shouldn't spend any time on py2, but don't need
            to prioritize removing code
<@dustyweb> sturm: yes and I spent a long time making sure py2 worked
<@dustyweb> shtrom: yes I think we shouldn't spend time on py2... I think we
            should remove test suites and test scripts and etc, we should cull
            those so we don't give the illusion that py2 is supported  [08:31]
<@dustyweb> sturm: you're thinking like the configure with py2
<@dustyweb> shtrom: yes exactly, anything user-visible we should cut off
<@dustyweb> shtrom: we just don't want to give the wrong impression
<@dustyweb> sturm: yes I suppose that includes docs as well
<@dustyweb> mtlynch: I think having py2 can be discouraging to new devs
                                                                        [08:32]
<@dustyweb> (or the reverse: that py2 specific code may be worked on and be
            time wasted?)
<@dustyweb> shtrom: I think stuff like the unicode stuff, etc.. that stuff
            runs in py3 or?
<@dustyweb> sturm: there's a lot of switch conditionals.. I think our sense is
            that we can leave it as is and not remove immediately but remove
            soon  [08:33]
<@dustyweb> shtrom: maybe you can remove stuff if you're working on it but
            don't spend time on the py2 stuff
<@dustyweb> sturm: ok good, anything else to address in the next release?
<@dustyweb> sturm: I've included a bunch of other things at the bottom
<@dustyweb> sturm: batch uploads, more responsive gallery interface, etc
                                                                        [08:34]
<@dustyweb> mtlynch: so my pet future is CI and docker, it's not particularly
            tied to a release but I'd love to see that addressed as soon as
            possible
<@dustyweb> sturm: I'm in on the CI stuff michael
<sturm1> R13ose: Are you with us for the meeting?  [08:35]
<@dustyweb> cwebber: I think maybe we could consider in scope of the release
            is thinking about infrastructure changes  [08:36]
<@dustyweb> sturm: yes, chris simon and I had a call about infrastructure and
            how to get that stuff back up
<@dustyweb> sturm: we were in a bad place, even turning away new contributors
            if they were even trying   [08:37]
<@dustyweb> sturm: I pushed ahead with fixing trac and the wiki as they were
<@dustyweb> sturm: and we spoke to ian from the fsf about a replacement for
            savannah
*** R13ose (uid312103@gateway/web/irccloud.com/x-soohnzjysakpkxqt) has joined
    channel #mediagoblin  [08:38]
<@dustyweb> sturm: they are looking for a replacement for that, are going
            through a lot of options, so my current feeling is stick with what
            we've got until we hear what their plan is for a new collaboration
            tool... but I agree we're in a strange direction where we've got a
            mailing list where people can respond to the mailing list, and
            there's an issue tracker, but it's kind of invisible to many
            people... so I don't think we need any immediate
<@dustyweb> action but if anyone had particular thoughts
<R13ose> Hi  [08:39]
<@dustyweb> mtlynch: so I really don't like Trac... have had a bad experience
            with it.  I noticed sturm that you had your own gitlab fork, is
            that possible, a way of contributing there?
<R13ose> Sorry, I am here
<@dustyweb> mtlynch: as a contributor it's hard to even find the issues I've
            contributed to
<@dustyweb> mtlynch: having close integration between issues filed and merge
            requests related to the issue
<@dustyweb> mtlynch: trac feels very disconnected
<@dustyweb> sturm: as for the q about could we use gitlab for the time
            being... one reason is I didn't want to personally move around
            between things  [08:40]
<@dustyweb> sturm: I think we don't want to move and then move again
<@dustyweb> sturm: if you're asking if we could use gitlab as a
            merge-review-process, maybe we could
<@dustyweb> sturm: I think the main gitlab wariness is that it's mostly free
            software except where it's not
<@dustyweb> sturm: so there's just some tension there in jumping straight to
            gitlab  [08:41]
<@dustyweb> sturm: we have some momentum but I don't want to waste it all on
            extra infrastructure
<@dustyweb> fergtm: have you considered https://gitea.io/ ?  [08:42]
<@dustyweb> sturm: we haven't but the fsf has
<R13ose> I agree moving twice is no good.  Moving once is key if that is what
         everyone wants to do
<@dustyweb> cwebber: I agree that moving twice is a problem  [08:43]
<@dustyweb> sturm: would you like to talk about the CI?
<@dustyweb> mtlynch: we have a working CI for gitlab, it'll build the current
            code in docker
<sturm1> R13ose: I can't see you in Mumble? Are you in there?  [08:44]
<@dustyweb> mtlynch: the previous docker images referred back to the savannah
            server, the changes I've been working on with ben pulls in the
            current workspace and excludes those which are part of your local
            build.  If we can get that on CI
<R13ose> Mumble is not working.
<@dustyweb> mtlynch: I'm looking at cypress (?) which does modern end to end
            browser integration
<sturm1> Ok, that's a shame, next time!
<@dustyweb> mtlynch: I think that docker and CI would help a lot so that every
            time a change is made we know things still work  [08:45]
<@dustyweb> sturm: for the wiki, my major thought is to eliminate the wiki
            entirely.  I've culled a few obviously mostly blank pages just to
            make sure we get rid of all the low hanging fruit
<@dustyweb> sturm: there are ones I'm trying to pull into the documentation
            like the hacking guid  [08:46]
<@dustyweb> e
<@dustyweb> sturm: I saw someone get kicked because of the registration
            requirement on IRC?  That inhibits casual drop ins
<@dustyweb> cwebber: I'll give you irc ops and we can change it  [08:47]
<@dustyweb> sturm: trac issue, need to work with simon
<@dustyweb> sturm: it's going down every now and then but comes back pretty
            quickly
<@dustyweb> sturm: readthedocs was my next thing, I've removed the "latest"
            build on readthedocs because it wasn't clear
<@dustyweb> sturm: I think it's important that we remove the manual build
                                                                        [08:48]
<@dustyweb> sturm: would be nice if we had master updates too
<@dustyweb> sturm: I'm thinking of dropping the stable one too
<@dustyweb> cwebber: the history there is that to let people run from a git
            checkout without switching between git tags  [08:50]
<@dustyweb> cwebber: and not using git master
<@dustyweb> sturm: sounds like we should keep talkinga bout it
<@dustyweb> sturm: I looked at the debian build stuff that jonas has done for
            debian packaging in 2017  [08:51]
<@dustyweb> sturm: I caught up with him we had a good chat, he's not in a
            position at the moment to pick up packaging work, but he's given
            us some good ideas
<@dustyweb> sturm: so that's encouraging, he's around so it's maybe possible
            to convince him or other debian devs to pick up his woork
<@dustyweb> shtrom: so is there any debian packaging code at the moment?
<@dustyweb> sturm: there's no mediagoblin in debian stable or testing, but
            it's in experimental  [08:52]
<@dustyweb> sturm: he's waiting for a couple of issues in our trac, he didn't
            make it super clear it was holding back his work... which it's a
            shame we didn't make that connection, possibly we could have
            treated it with more priority
<@dustyweb> shtrom: that is one things that I'd like to package things in
            debian, and an openbsd package
<@dustyweb> shtrom: it would be good to have pointers to the work  [08:53]
<@dustyweb> shtrom: I could probably help out on that front too
<@dustyweb> sturm: I'll point to that on the mailing list
<@dustyweb> sturm: there's lots of energy but not in places that are
            connected, so I'm trying to do a lot of work to connect people up
<@dustyweb> sturm: so what I'm working on at the moment is trying to get
            people responsible in irc and trac  [08:54]
<@dustyweb> sturm: see if I can help with a quick documentation and code
            changes
<@dustyweb> shtrom: one last thing on debian packaging
<@dustyweb> shtrom: being able to split data from ?? to the package
<@dustyweb> shtrom: this is one of the recurring issues where I wasn't able to
            install the package somewhere and have the data be somewhere else
                                                                        [08:55]
<@dustyweb> shtrom: i think that would be a really good thing from
            documentation
<@dustyweb> shtrom: I've looked at configure again, there's the base
            path... it's not entirely clear
<@dustyweb> shtrom: some clear documentation about that would help
<@dustyweb> sturm: I'll make that a priority for myself
<@dustyweb> sturm: I'd like to have a brainstorm of where we are, last 5 years
            or so, much has changed in our allies in peertube, big blue
            button, etc etc etc etc  [08:56]
<@dustyweb> sturm: new apis and etc, not a good time in our last 4 minutes
<@dustyweb> sturm: just watched Deb's talk from linux conf 2018, and the story
            about aiming for non-technical people and artists in particular
            resonates with me  [08:57]
<@dustyweb> mtlynch: one long-standing question I've had is "who is the target
            for mediagoblin"?
<@dustyweb> sturm: let's hold that discussion over for next time
<@dustyweb> cwebber: I did some work some time ago on mediagoblin in guix,
            happy to guide someone thorugh where things are  [08:59]
<@dustyweb> sturm: I asked a few people on twitter who mentioned on
            mediagoblin, etc.. asked what more people "what would make
            mediagoblin more useful"
<@dustyweb> sturm: batch uploads has been on our list  [09:00]
<@dustyweb> sturm: chunked uploads where things don't freeze
<@dustyweb> sturm: mobile uploads
<@dustyweb> sturm: oembed
<@dustyweb> sturm: [... goes quiet and disconnects, reconnects ...]  [09:01]
*** shtrom (~shtrom@2001:44b8:5155:e100:b132:754d:ea69:77eb) has joined
    channel #mediagoblin
<@dustyweb> sturm: oops I set up my computer to disconnect from the internet
            so I could go to bed  [09:02]
<@dustyweb> shtrom: I'm actually working on a private mode thing at the moment
<@dustyweb> sturm: last minute things?  [09:03]
<shtrom> ah, that's where the scribing is (:
<@dustyweb> breton: mobile clients, I feel that we need clients for mobile
            phones
<@dustyweb> cwebber: it's a real comfort to see everyone springing up again
                                                                        [09:04]
<@dustyweb> rigelk: I'll hopefully show up at the next one
<@dustyweb> sturm: we'd like to engage more, you can reach me on the fediverse
            as well  [09:05]
<@dustyweb> shtrom: echo what chris was saying, nice to see activity
<@dustyweb> sturm: with that we'll push ahead and look forward to a new
            release in the future (july?)
<@dustyweb> sturm: it's appreciated
<@dustyweb> ==== MEETING LOGGING ENDS ====  [09:06]