See mlemweb’s post here: https://octodon.social/@mlemweb/104145782400356981
We’ve scheduled a planning meeting for ActivityPub Conf 2020 next Saturday (5/16/20) at noon EST . We’ll be connecting via mumble at: https://www.w3.org/wiki/SocialCG#Joining_the_meeting and on irc in #apconf on freenode
Is there a Matrix bridge for the chat room? I havent used IRC in a long time.
You could use the official bridge with nodeJS
It can bridge any freenode channel.
The Logs of the first planning session can be found here :
–> https://redaktor.me/pad/p/apconf2020
How to join Mumble
Using the SocialCG meeting room:
https://www.w3.org/wiki/SocialCG#Joining_the_meeting
Agenda
Things we need:
-repository for collaborative coding (github/gitannex/other?)
-ConfTube instance for pre-recorded videos (same as last year)
-scheduled jitsi/mumble sessions for discussions of videos
-scheduled jitsi/mumble birds of a feather meetings
-participant driven unconference jisti/mumble sessions for specific topics
-jitsi/mumble meet + greet or ‘hallway’ track conversations
*I prefer Jitsi over mumble because there is the option for video, though participants can opt out, whereas mumble can only be voice. Video gives a closer aproximation of conference interaction.
General Timeline: (Proposed Sept 12-13 2020) * or Sept 19-20 2020
Now: Choose Platform(s) and structure of conference, delegate specific tasks to volunteers
3-4 months out (May 22): CFP
2-3 months out (June 12): Deadline for CFP
2-3 weeks out (Aug. 27): Deadline for pre-recorded talk submission
2-3 days before sessions (Sept 10): screened/edited talks uploaded to conf-tube instance
Day 1: Jitsi meet and greet; scheduled Jitsi meetings for questions/discussion for each talk, Unconf topic submissions
Day 2: Unconf planning session AM (Jitsi), scheduled Jitsi meetings for unconf sessions, Jitsi closing session (hallway track)
meeting logs
mlemweb: I think we can start the meeting, though we may be waiting
but this is to see who wants to help, delineate tasks
emacsen: If we do the synchronous / asynchronous model, I like that, it seems to simplify a ton of problems
also affords you the ability to do things over a longer period, reduces burnout
you can also do multiple sessions for one talk
the only things you really have to solidify is how do you host the videos, how you do the synchronous part
the one thing I’d propose, a little different than what I’ve seen proposed last week, I don’t think you should do the penguicon “choose your own…” thing
mlemweb: I agree, we were never plannign on doing that
emacsen: I’d suggest choosing one platform for each avenue of communication
it would be cool to have people watch the video together, and be able to take questions while the talk is going on so you can have questions
and then optionally break out for a BoF session
that’s my 2 cents
mlemweb: there are a couple of ways we can do this
one thing with the synchronous/async thing… I think we could post the videos at least 2 days before the scheduled jitsi meetings
and that way we have a buffer in case there’s tech difficulties, etc
so that way people can watch the videos more leisurely and don’t have to watch 10h late
as for watching together, we can always use the screen sharing method on Jitsi where the presenter plays the video from their screen and takes questions after that
but not sure how that would work with buffering
or we could schedule a time, have enough people watch the video and then do Q&A
emacsen: is this a paid conference?
mlemweb: no
emacsen: one thing you can do if it’s paid is say “anyone can watch the videos but only people have paid is do live Q&A”
mlemweb: that’s true and it’s a way to know how many and limit it
but, if we have people pay for the conference that’s a lot more for setup of the conference that I don’t want to deal with
and since this conference has no costs to run it I think it’s unnecessary
emacsen: it will have some but you’re right that it’s relatively low
but I’m really interested in hearing from others
mlemweb: doing introductions: I’m Morgan Lemmer-Webber, I ran apconf last year, presumably will be doing so again this year
cwebber: I worked on the activitypub spec, co-run the socialcg
eth01: I came across this on twitter the other day, I’m founder of fosshost.org; we’re relatively new open source project. We provide hosting services to open source projects. I’ve been involved in freenode for the last 10 years… we currently have a global platform, we’re in 6-? spaces across the USA. I’m not sure what can add directly to this conversation, but I want to build networking between projects to see how we can work together
sebi: may I suggest you join the activitypub forum? I can link it later
mlemweb: yes can you link it
emacsen: I’m Serge, have been in activitypub stuff for a while and would love to help apconf
sebi: Hello, I’m sebastian, I co-organized the last Prague ActivityPub conference, and I work on an ActivityPub CMS
mlemweb: the format we’re discussing is doing a synchronous/asynchronous conference with recorded videos of talks and then scheduled meetings via Some Platform ™ where people can ask questions / discuss talks and do hallway tracks / Birds of a Feather, to recreate the feeling of a conference
one objective of this call was to find out what platforms to use
we have a conftube peertube instance that sebastian helped with last year, can we use it again this year?
sebi: yes we can use it again, and we should prefer activitypub services where possible. Do we just do video recordings? because if we also do audio recordings we can do it on funkwhale… if anyone wants to do written pieces on writefreely or something… then we have unique hashtags for activitypub posts
and then I have an idea for pre-recorded talks that we have a CFP-like system, I like the idea of the CCC, they have it the “angel” system; anyone who is good at video recording or has technical background and say “I want to be a video angel” so anyone can ask questions in the IRC
cwebber: I think using ap-oriented things is a good idea
I think we should focus on just a few services to keep it simple
sebi: I think we could have a page that aggregates things
cwebber: sounds good basically the official things organizers are involved in
mlemweb: have a webpage collecting things
serge: yes submitting presentations, etc
emacsen: I agree with Chris, we want to avoid have attendees/submitters submit stuff through a lot of places… that could be a UX fail; I think that the major amount of work is going to be in 2 areas
first of all help people submit video talks… that’s based on going to other online events, many people aren’t set up to do good video and especially audio capturing
so it might be that either we offer handholding or not offer handholding
so either we offer it or don’t because it would be very time intensive
the realtime component could be labor intensive, and it will always have to have a moderator to remove disruptive people or to pull questions from audience to the speaker
there has to always have a helper on line to complete that task
mlemweb: yes and I think whoever sets up the call can be the moderator
cwebber: If someone’s willing to volunteer to host a jitsi meet event for a pre-recording a few days ahead in Jitsi, this way we have someone to capture the video and help
sebi: yes this is basically what I meant about the video angels, I think it’s a good idea has worked in several forms before
to emacsen I absolutely don’t disagree but the basic failure is usually about these services, most of them can’t render all of them… if peertube would allow for instance to allow long markdown texts below the video it could be solved already. The problem is that most of them are clones of major things in the web but it’s kind of an external problem
cwebber: I think we should do a test of the video moderating
mlemweb: we should combine chris’s two points; in my combined timeline I suggested that we have a deadline 2-3 weeks before the event for pre-recorded talk submissions, maybe 2-3 weeks before then we have a jitsi or etc meeting for the presenters to 1) test moderation 2) help anyone who isn’t able to record their own videos and 3) maybe make sure we can handle this format ahead of time and suss out any potential problems that we might run into. so just take the system for a test drive when it’s low stakes and not the conference itself
emacsen: so we’ve talked a lot in theory, these shouldn’t be considered set in stone but we should start breaking out into teams of people based on their interests of what they want to do… I’m happy to volunteer for looking at options for streaming synchronously and I’d be happy if I had another person to look at async options
mlemweb: so do we want to shop around for platforms? because we want to use FOSS… the two platforms suggested last week were jitsi or mumble, I’d prefer jitsi because it has a video option and feels more like a conference and those who don’t want to be on video can turn it off and still see others. so what other platforms are we considering?
emacsen: so I have a couple of ideas; one is big blue button which is designed for educational environments, was used by sean o’brien for his conference, they seem to have thought it’s ok. The other one is if the aim is… if it’s everyone feels like it’s one big room, except that it breaks down after a certain number of people. if it’s a large system we should aim for a chat system with one active presenter, and that can be anything including a DIY thing. Look at jitsi and big blue button, big blue button apparently taps out at around 100… if we think it’s going to be bigger than that we should see
mlemweb: we’ve had 15-20 people on jitsi
eth01: jitsi depends on where hosted, we support about 15 people, we’ve seen jitsi conferences with over 100 people
the datacenters we’re in, we’ve been allocating 30-40 gigs
cwebber: is fosshost willing to host a jitsi meeting
eth01: so right now we have a lot spare capacity because we’re relatively new… one reason I wanted to join this conference was to show we’re willing to join and help
so if you want us to provide hosting we’ve got a platform in chicago, some in LA, two platforms in the UK… one in amsterdam, one in poland
mlemweb: excellent… if we can set up something that’s more secure and able to host more than between 15-30 people that would be wonderful
eth01: interestingly, some of the volunteers on fosshost runs a coop in the UK, and he uses big blue button… so we have someone on board who’s familiar with that who’s pretty much an expert
mlemweb: so… emacsen if you’d like to do some research between bigbluebutton and jitsi and possibly a homebrewed thing you can do that and maybe get back to us?
emacsen: so maybe I should talk with eth01 if you’re going to host it?
I guess the other big question is “what’s our approx attendee numbers”?
if it looks like prague where it was
mlemweb: 40, that was the capacity
emacsen: so 50 vs 100 vs 200 are very different needs, so we should either cap what we allow and say “we can only allow so many people into a live event” (because we only care about that for the async) or if we can get “some idea” of how many people want to attend a synchronous event at a time
mlemweb: I would also like to track that down but it could be useful for an online event… last year in Prague we had a cap because of the venue and also travel… we hit our registration cap and had a substantial waiting list last year, and that was with a cap and travel… so if we have a registration cap we can probably manage that
emacsen: I was going to suggest exactly that, you don’t want to charge I get that, but I don’t think of any major issue with saying “we will require you to register to get the link to the talk”
so maybe we have a webpage that we only put out to people who registered. obviously people can share it, but if we do that we have some idea of how many people we’re serving
mlemweb: yes, and since the videos are synchronous it’s mostly a participation cap on the live Q&A
cwebber: if we do simultanious talks, it should diffuse the participation in each talk, but it might take away from the feel of the conference
mlemweb: we can combine these two ideas, do registration for specific jitsi talks
as we discovered last year we had a few things in the unconference last year at the same as chris’s talk, everyone went to chris’s talk
so if we have registration for individual sessions we can guage it better
emacsen: I don’t see any reason we should have overlapping talks if we’re running a virtual event… I don’t believe in artificial scarcity as a means of crowd control. Having to register to specific talks doesn’t make a lot of sense unless we have software specifically for it… it would bring about a lot of complexity. If someone wants to write that program, great! go ahead and write it… but that sounds complicated for volunteers and attendees
how many volunteers will we have? right now if we only have everyone in this channel that’s an enormous amount of work… that seems like a lot. On the other hand if we have double this amount, that would be pretty easy and nice. What I think we’ll need to do if we do want to make this easier on everyone is increase the volunteers
mlemweb: going off of that the number of talks will determine the number of days we have to do
last year we were able to select all the talks but we had a very short CFP time…
partially how many volunteers we need depends on how many papers we have and whether we want to select everyone or have a selection committee
cwebber: registering for individual talks/sessions helps know numbers of attendees, but adds levels of complexity (individual links to individual sessions etc) requires more work and moderation. Easiest thing to do might be: no overlapping talks, per-day registration, couple of days of talks (2 days), couple of days of BOFs (2 day), depends on interest
sebi: I agree with 2 days + 2 days, but how about collaborative coding… we can('t) do it as the same 2 days… so we need a timeframe after for code together / hackathon
also I like the idea of overlapping talks but we need to identify overlapping talks, but there’s a big timezone difference
I already identified some timeframes that may be reasonable
l.a. 06 - 10.30 & 13 - 17.30
ber. 15 - 19.30 & 22 - 02.30
tokyo 22 - 02.30 & 05 - 09.30
2x 4.5h // that gives each timezone min. reasonable 6.5h / day
mlemweb: as long as we have a buffer so that there isn’t one timezone that’s screwed over for every session
sebi: exactly
mlemweb: so since it’s online we don’t want it to be US/EU centric, we want it to be global
emacsen: I mean, one of the nice things also that one can do with a global conference is that if someone is willing they can do their talk again… not saying people should feel obligated but there’s no constraint we normally have… we can say X talk is at 4 Y talk is at 8… no limitations other than speakers… we can offer that as well as BoFs that could happen at multiple times… I think until we all switch to (universal decimal time) we have to deal with timezones
mlemweb: since we have pre-recorded talks it’s just 2 discussion sessions
and this time maybe we should give speakers more input as to time
whereas last year it was everyone in a room for 10 hours so I decided on a schedule
cwebber: We have organizers spending time choosing specific talks for asynchronous talks, but everything else can be an unconf (including the discussion sessions). Have online meetings at the same time that we had an unconference planning in an etherpad where we solidify the schedule.
mlemweb: I like that idea, last time sebastian suggested we recreate the post-it note route, I think the etherpad route
emacsen: I thought that’s where we were converging on was letting speakers select their own slots… it’s a structured unconf where speakers select time
it may require some level of moderation, dont’ know if it will or not… my biggest fear is that someone comes in intending to disrupt the conference and there’s no way to remove that individual(s)
but I agree that allowing presenters… you have support and guidelines
you get 2 slots, but you also need to run it on this platform… you can’t say “everyone go use zoom or etc” because we want to make that experience is unified for attendees
but we should make that experience as uniform as possible
mlemweb: so basically the presentation you’re suggesting but limited to presenters and volunteers?
emacsen: I see two different kinds of sessions; one session is sessions around a talk… only presenters can do that. but also people really interested in a theoretical talk or something they haven’t presented on… eg “VR on the Fediverse”… nobody has it but don’t know how to do it. but we might say within those 3 days here’s the time for those ad-hoc sessions. I agree with sebastian that here’s where collaborative coding would be awesome, but I have a lot of questions about how to do collaborative coding.
so 3 aspects: presenters, BoFs, and collaborative coding
sebi: I have 1 more question: we’ve talked a lot about scope right now with emacsen just identified, but do we also create a scope for end users which is to invite all the platforms and activitypub projects to present in a place what they’re currently doing and something to present to the end users and which is more apart from all these things
mlemweb: so basically a “state of the project” for individual projects?
sebi: yes and for smaller ones people don’t know about yet
so it can be anything
what pukkamustard just did (?)
mlemweb: so I don’t know that we need to distinguish between that and other CFPs… we can solicit the missions for state of the projects from different projects… when we scheudle we cluster them instead of inviting people to talk, except maybe keynotes
sebi: I think it must not be talks, but also online demos written things etc… like anyone is completely new or mastodon user what else can I do, with AP or fediverse and etc… where we put the webpage for end users etc
mlemweb: I think that may be broadening the scope
cwebber: I’m hearing some convergence. 3-4 core aspects: Presenters, BoFs, collaborative coding, lightning talks. We have a CFP for talks, BoF topics, then set up a meeting with accpeted presenters or BoF moderators for scheduling. At end of conference we do a live lightning talk session. Through the conference we keep the IRC open for collaborative coding and can present at lightning talks at end
mlemweb: that makes sense, and if that’s happening on irc and etc that removes that infrastructure burden… people can use what they already typically use. not sure if that meets the ideal sebastian was thinking of
emacsen: I agree with chris’s structure, I think sebi was suggesting a 5th category for new users, I’m afraid if we make that another track that’s increasing the scope a lot… we’re already hitting scope creep for the number of volunteers as we have. As a volunteer I want to limit the structure. I think we want to limit talks etc, and then we can support sebastian’s ideas as BoFs.
mlemweb: I think that’s a great strategy, plus then if that’s a BoF we could do that as veterans and newcomers both able to set their levels
sebi: I agree to all of that and what Chris mentioned about lightnigng talks was exactly what I was looking for but I forgot that term… I think that thing is that it might be too much was also my concern which is why I made a question for BoF onboarding… if we have one week for collab coding afterwards we can have a group of people willing to work on onboarding
mlemweb: so my question for conference organizers, do we need to set that up for week after… is that something we set up but encourage after the conf to let people set their own tools etc
sebi: absolutely agree, should be up to the people
PROPOSED: conference uses the following structure:
There are 4 categories:
talks (synchronous, but then with Q&A slots)
BoFs
Hackathon that runs alongside the conference, and then extended week after (but mostly on IRC)
Lightning talks at the end of the BoF session times
Planning like so:
We do a CFP for talks & BoFs
Once talks & BoFs accepted, we have “slot allocation” for the 4 days: send an etherpad link and do a mumble (or jitsi) call where we live allocate talks for those slots with the speaker, BoF organizer, and volunteers present
Conference like so:
Videos released ahead of time (maybe a couple of days before)
Then 2-4 days of BoFs and Q&As (depends on CFP volume)
Hackathon runs in parallel and then next week
Last day of BoF’ish days has lightning talk slots, also to show off hackathon results (for first section, continuing into next week)
END PROPOSAL
Consensus check:
cwebber: +1
sebastian: +1
mlemweb: +1
serge: +1 (to structure)
eth01: +1
Group acknowledgement: we still need to solidify dates
We now have a “Big Blue Button” instance ready at fosshost.org where we could meet at this nice machine so we can check if the software meets the requirements for our conference and how we can improve it with AP or customize it …
As soon as @eth01 is ready with BBB I will setup Discourse integration:
Is there no CFP yet?
Hey! Ho! @eth01 just setup the SSL certificate on the BBB, so we can test it (only for desktop now):