#social 2021-01-09
2021-01-09 UTC
hellion0, jacky, ajordan, remy, Grishka, join_subline, Zakim and RRSAgent joined the channel; hellion0 left the channel
# 
RRSAgent logging to https://www.w3.org/2021/01/09-social-irc
# 
rhiaro Hi all, we're having a Social CG meeting in 1 hour, in jitsi and here, about the Fediverse Enhancement Proposals https://socialhub.activitypub.rocks/t/2021-01-09-socialcg-meeting-fep/1246
remy and jarofgreen joined the channel
# 
nightpool[m] Just to be clear, https://meet.jit.si/ScatteredConsequencesActRegardless is the jitsi link, correct?
# nightpool[m] +👌
# nightpool[m] present+
# cjslep[m] joining, 1 sec
# cjslep[m] present+
# jarofgreen present+
# Grishka present+
# nightpool[m] I can do scribe!
# nightpool[m] scribenick: nightpool[m]
# nightpool[m] cwebber2 (IRC): This group currently has 3 cochairs: myself, aaron perecki (sp?) and nightpool
# nightpool[m] ... you may notice the socialcg has still been struggling with respect to organization, and I talked to amy, and amy has a lot of ideas and is very knowledgeable and involved with the way the W3C is structured
paul joined the channel
# nightpool[m] ... so I talked to the three other co-chairs, and we all agreed
# nightpool[m] +1 !
# Grishka imagining chairs as pieces of furniture being thrown at the problem
# nightpool[m] is still warming up their fingers :D
# jarofgreen +1 (but i’m unsure about voting eligability)
# paul +1
# Grishka +1
# nightpool[m] Topic: The Fediverse Enhancement Proposal process
# nightpool[m] cwebber: we're proceeding with pukkamustard here, CJ, are you okay with representing it?
# nightpool[m] cjslep: Yes, I talked to pukkamustard before the meeting and we're okay to go ahead
# nightpool[m] cjslep: Going back to APConf last year, we had some BoF sessions where we wanted a lightweight way to discuss improvments to activitypub and communicate in a way that provides for cross-software or cross-developer communication
# nightpool[m] ... it's really design to be a lightweight way, so i don't know if it could be considered a "standards" process, but the name was inspired by Python's PEP process or the XMPPEP(?) process
# nightpool[m] ... I'm not very familiar with these processes, but I know there were some concerns raised due to the similarities wrt that process, but we wanted something that was much more lightweight then those processes generally are
# nightpool[m] ... We just set up a Gitea instance at random, and we had one proposal from Claire (mastodon developer), and that seemed to work very well, and it engendered a really robust discussion within the community
# nightpool[m] ... and we just wanted to open up the floor to feedback from members of the community
# nightpool[m] q+ to talk more about the mastodon side of Claire's proposal
# nightpool[m] CJ, anything of substance i'm missing here? I got a little behind at the end
# nightpool[m] cwebber: I like it, it definitely doesn't have the full stnadardization process that say a w3c standard would, but it's not intended to, and I like how it provides for a way of creating new ideas that we can then take into a new WG charter or something similar
# nightpool[m] s/stnadardization/standardization
# nightpool[m] rhiaro (IRC): I agree, I read through it earlier, and I thought it was very good. It's lightweight, easy to read.
# nightpool[m] ... one way we can connect it to the CG is when things come down to the final stage, they can be published as a community group report
# nightpool[m] ... that way it gets a w3 URL and it makes it much easier to transition it into a W3 WG spec down the line if we want to
# nightpool[m] ... Although we may want to make that optional, in case people don't want to join the CG
sandro joined the channel
# Grishka btw if you need a second implementation of that FEP, I might try implementing it in Smithereen
# cjslep[m] q+
# 
cwebber2 nightpool[m]: my understanding is we still haven't merged that code for a few reasons, one is open questions on our side, but to our side it feels like we haven't gone through the process. from that perspective I think we were hoping for more consensus than we got from the process, where it seems like it's more to reflect the processes of what one application is doing, which is valuable but maybe not what we're looking for in this
# cwebber2 nightpool[m]: I guess the only other thing I have is... it's good for people to get their implementations in front of us so we can find more spaces of commonality... on the other hand if we rubber stamp everything that comes through as a lightweight process if we'll come to a point where everything does slightly different things in slightly different ways which might make it harder to converge on a future spec. not sure there's a good
# nightpool[m] cjslep: I definitely agree with nightpool, and I understand how the FEP process can be disappointing in that regard
# nightpool[m] ... because it wasn't designed to give that kind of feedback
# nightpool[m] ... it was designed so you can push your things out and let you communicate about it without necessarily coming to a consensus
# nightpool[m] ... and the CG could look at a wide range of FEPs when starting a potential standardization proccess
# nightpool[m] ... I don't have any good solutions either but i think it was intentional for it to be that way
# nightpool[m] cwebber: It strikes me that a lightweight process is useful even if it's not a full process
# nightpool[m] ... because a lot of things are happening, and we don't even have them documented
# nightpool[m] ... and that's a much worse solution then when lots of things are happening and we do have them documented
# nightpool[m] ... To another end, I think this ties in to something else that we were talking aobut with the SocialCG. Now that we're using Jitsi, which supports video calls
# nightpool[m] ... We've had problems with doing things regularly, and not having enough things to work on on a rigorous schedule, and not having enough meetings to cover the topics when not doing a rigorous schedule.
# Grishka "here are groups in Smithereen, here's how they work…. oops they don't" :D
# nightpool[m] ... We've talked about doing a more "show-and-tell" sort of process, where people come and present on what they're working on, and maybe it would be good to couple that with a FEP, so people come and present their projects and what they're working on
# nightpool[m] ... alongside an FEP
# nightpool[m] cwebber2 (IRC): It might be useful to have a proposal at this time, let me type something out
# cjslep[m] q+
# nightpool[m] cwebber2 (IRC): There are a couple of things not captured here, including governance for the FEP process, as well as "graduation" or anything that comes after the FEP process
# nightpool[m] ... not sure whether we should be adding anything on top of that
# nightpool[m] cjslep: Wanted to highlight here that we currently have an "editor" role in the FEP process, which is curerntly me, pukkamustard (IRC), and lain_soykaf (IRC)
# Grishka +1
# paul +1
# nightpool[m] didn't quite catch what rhiaro (IRC) was saying, pls feel free to reiterate here!
# cjslep[m] qq: am i allowed to vote?
# cjslep[m] +1
# nightpool[m] +1 if we also widen the group of editors or change that process in some way to make it more clear what editors are doing
# nightpool[m] cwebber2 (IRC): woo! 2 proposals!
# nightpool[m] ... Let's see if there's anything else on the waitlist we can knock out easily
# Grishka rhiaro: I had a similar problem, some email providers don't accept those emails for some reason
# Grishka had this with APConf as well
# nightpool[m] cwebber2 (IRC): I think we've made good progress on the activitypub.rocks page this week, it's been updated with things people have been requesting and points to test.activityrocks.dev, which is cjslep's test suite instead of the broken test.activitypub.rocks
# nightpool[m] ... round of applause to cjslep for making sure the fediverse has a working test suite!
# nightpool[m] rhiaro (IRC): I took the waitlist and turned it into a meeting schedule
humanetech joined the channel
# nightpool[m] rhiaro (IRC): I just arbitrarily assigned them days and times, alternating fridays and saturdays, so people can see if they can make the times for topics they're interested in
# nightpool[m] rhiaro (IRC): We've been alternating between Friday's and Saturday's to accommodate people who can't come due to work or due to the weekend, we can continue doing that for now and move topics around to accommodate people's availability.
# nightpool[m] cwebber2 (IRC): Since we've already gotten through the stuff on our agenda for the week, is there anything else we should work on for the rest of the meeting?
# nightpool[m] rhiaro (IRC): Neither, I'd prefer to do a "any other business" section to open the floor to anything else that we haven't planned to address
# nightpool[m] q+ for the Policy group thing
# nightpool[m] sandro (IRC): If I understand correctly, over the last 2 days all major platforms have banned Trump and possibly some amount of his followers (possibly around 500k accounts removed?)
# nightpool[m] q+ to talk to gab
# nightpool[m] ... They were thinking of moving to Parler, although it's banned from the app store and we don't know much about if they're going to be able to stay there, with it removed from the app store
# nightpool[m] ... so I think the fediverse at some point is going to need to deal with them maybe coming here, and figuring with what comes next
# nightpool[m] .. and there's some opportunity here for us, with expertise, to help the world manage that problem
# nightpool[m] cwebber2 (IRC): Sandro, you might not be privy to what's happened to gab over the last year, are you?
# nightpool[m] ... They left the fediverse, and said f*%! activitypub, basically.
# nightpool[m] ... I linked a video where that explains kind of what happened, where they made a big deal about leaving the fediverse
# cjslep[m] q+
# nightpool[m] cwebber2 (IRC): As a note, I have been a vocal skeptic that the allow/denylist approach will work long term
# nightpool[m] ... but I think what surprised me over the last year was how well it did work to dissuade gab
# nightpool[m] cjslep: One thing I saw consistently from voices on the fediverse, including playvicious.social, was calls for improvement for tooling for handling these kind of cases
# nightpool[m] ... one thing I think may help personally is having more interoperability with moderated actions, possibly arrived at with a FEP like process
# nightpool[m] sandro (IRC): Yeah, I've been on the "we need better moderator" tooling thread for a long time, but i'm realizing it might not have a lot of benefits here when you're talking about 8 million people
# nightpool[m] .. and maybe there's something we can do with software licensing.
# Grishka privacy settings!
# nightpool[m] cwebber2 (IRC): I would disagree with going down the software licensing route, but I think that if there's something we can do here to make sure users who don't want to be exposed to these type of things aren't exposed to them, i think that's a big goal here
# Grishka q+
# nightpool[m] ... because historically, we've had private nazi forums for as long as we've had the internet, and a lot of this organizing has happened literally behind closed doors
# nightpool[m] ... and a lot of the storming the capital planning happened in closed fora
# nightpool[m] ... so I think we need to acknowledge our limits and scope this to what we're practically able to achieve.
# nightpool[m] ... since we cannot prevent other people from using protocols to communicate
# nightpool[m] Grishka (IRC): privacy settings on Interaction is something that can help solve this
# nightpool[m] ... for example, if you set up "only ___ can reply to my post" or "I only see notifications from people I follow"
# nightpool[m] ... you can protect yourself from some of these types of content
# nightpool[m] cwebber2 (IRC): I will note that OCapPub addresses some of these interaction concerns, but it's not tested or implemented yet
# nightpool[m] ... I suggest we spread some of these discussions out over the next several meetings, since I don't think we have a solution here but this conversation isn't going away
# Grishka services like Facebook do the opposite btw — they make it nigh impossible to guard yourself from content outside of your circles
# nightpool[m] ... I don't think it's going away in the following way: this organizing happened in public, and it's already leading to calls for banning end-to-end encryption
# nightpool[m] cwebber2 (IRC): we've had a few informal socialcg meetings about it, as well as a few meetings with policy groups from the EU about it
# Grishka Facebook would do to AP what Gmail did to SMTP
# Grishka *if* Facebook chooses AP
# nightpool[m] ... And discussions about the proposed EU interoperability law, which doesn't say much yet except for the word "gatekeepers" repeated over and over throughout the document
# nightpool[m] ... and the concern was around organizations adopting the protocol and then closing it off, like we saw with Google and Facebook and XMPP
# nightpool[m] ... and then we moved to discussions about how legislatures aren't aware that non-centralized approaches exist
# nightpool[m] ... and how to make sure that the legislation doesn't harm decentralized approaches and make sure that regulators are aware that these alternatives exist
# cwebber2 nightpool[m]: that makes a lot of sense... I guess I'm wondering from the technical land of what this group is good at and has an expertise in, what's the best way for us to approach it that's not necessarily presenting ourselves as a policy or political body which I don't think we have the setup or expertise or chops to be. but I think there's a lot of good goals here and we should do our best to make sure that we're providing the
# nightpool[m] cwebber2 (IRC): I would like to note as Amy mentioned that we did have some of those people here, and that was the goal as of the place we left it last time, and that most of the people in the Policy group are connected to these kind of issues, and
# nightpool[m] ... it's a different composition then the people who show up to this group
# RRSAgent I have made the request to generate https://www.w3.org/2021/01/09-social-minutes.html rhiaro
# nightpool[m] cwebber2 (IRC): and that will be the topic of our meeting next time! next Friday
# nightpool[m] rhiaro (IRC)++
# nightpool[m] cjslep++
# nightpool[m] rhiaro++
# RRSAgent I have made the request to generate https://www.w3.org/2021/01/09-social-minutes.html Zakim
Zakim left the channel
# nightpool[m] haha, I can't believe it picked up my emoji reaction as a present + !
# Grishka I like how two bots talk to each other
# cjslep[m] rhiaro (IRC): RE the gitea instance emails: I know the emails are prone to being marked as spam since gitea is just using basic sendmail to issue the emails. I'm not a wizard at email configuration. :( I can manually mark your account as activated if you'd like, however this means that you still probably wouldn't receive the notification emails that the instance sends.
# cwebber2 wow the meeting logs recently got much fancier https://www.w3.org/2021/01/09-social-minutes.html
# cjslep[m] With your permission, I can also manually send a test email to your account, rhiaro (IRC).
# cjslep[m] I have issued the test email.
# humanetech cjslep[m]: you might also configure a different more reputable mail provider, maybe?
# cjslep[m] I will add it to my list of things to do. :)
# nightpool[m] wow I'm really impressed by that CSS! I took a look at the scribe code last year when I was figuring out the bot ropes and it was so incredibly complicated
# cjslep[m] git.activitypub.dev
# cjslep[m] Or that server's IP.
# humanetech ps. if the wave *does* come with, say, 20m users, even moderated away to a separate space, then it is 4m against 20m.. bye bye fedi-as-we-know-it: a public image problem arises.
# humanetech in that light even public discussion might trigger ideas with the wrong ppl
# 
@elidourado ↩️ Better protocols than RSS exist specifically for social data https://www.w3.org/TR/social-web-protocols/ (twitter.com/_/status/1347952401102331906)
# Grishka actually, I don't think it's really a big issue
# Grishka mere existence of bad actors IS a big issue on centralized services because those have recommendations
# Grishka and the very purpose of those it to forcefully expose you to content that you didn't consent to be exposed to
# Grishka on fediverse, you can't just randomly see posts outside of your network — you gotta specifically browse those profiles, or they have to actively interact with you by e.g. mentioning you or replying to you
# Grishka and privacy settings I was talking about would give the users control they need to essentially preemptively block those they don't want to see
dmitriz joined the channel
# nightpool[m] grishka: I think part of what humanetech was talking about was that even if the fediverse users aren't impacted, the public perception of the fediverse would change dramatically if we get a flood of nazi users
# Grishka of course it's as important as ever, PR-wise, to emphasize that there is no single entity running the whole thing and it should be thought of as many interconnected communities
# humanetech Yes, I meant the PR / public opinion aspect. Like fedi being the big bad dark web, very easily depicted as such by other parties, and easy too as it sticks in the minds of ppl who hadn't heard of it before.
# humanetech Growing in numbers, Fediverse name recognition and such, 'branding' of the fedi is very important in that regard, imho. Most of the grassroots growth so far came from tech circles, but more is needed to become prepared for the masses.
# humanetech Same with EU Policy thing..
# humanetech *iif* a big tech platform were to adopt AP we need to be able to handle the influx.
# Grishka the explanation that I found works best is "like email but social media"
# Grishka because email is a federated system and it's familiar to literally anyone who has ever used the internet
# humanetech Yes, but you also mentioned the problem here. That in the mind of most ppl email means Gmail, or maybe Hotmail or whatever. And mail from your small-time mail server ends up in Spam box at the Gmail end.
# Grishka you then proceed to explain that each server is independently owned and operated, and has its own rules it's moderated in accordance with, and if you don't agree with existing servers, you can always make your own
# Grishka spam will inevitably be a problem in the fediverse as it grows, it's just that it's now possible to deal with it manually
# Grishka this is the kind of tradeoff you get when you make the whole system open to anyone
# humanetech Agree. But this explanation is already sorta technical. Your grandma or teen daughter doesn't care about this. I referred to the spam in the way that your small mailserver is less effective in being 'part of the whole'. It it the big shots that have taken over.
# Grishka it's either centralized and you have to solve captchas, upload IDs and do other kinds of humiliating procedures when you're suspected as being a spammer, or it's decentralized and you have to filter the spam on the receiving end
# humanetech Also agree. I guess it is more a question when Fedi is ready to handle these scenario's. Are we mature enough at this moment to do so?
# Grishka humanetech: are you suggesting that there will be centralization like there was with email?
# humanetech I am full in on decentralized web.
# humanetech There might be. Might play out a bit differently than with email.
# Grishka uh, Smithereen doesn't implement any sort of blocking as of right now… :D
# Grishka not even individual account blocking
# humanetech Do you intend to add that still?
# Grishka (are there proper message replies on IRC and it's just that my client sucks?)
# Grishka of course I do, it's an important feature to make it actually usable
# humanetech Ah, just asking. Maybe it was intentional :D
# humanetech Suppose FB was forced to be interoperable and they embraced AP for that..
# humanetech What would they do?
# nightpool[m] for which?
# Grishka they'd introduce a lot of vocabulary I presume
# Grishka like the exact sort of thing I'm doing right now
# humanetech Yes, and maybe they would also break up their centralized server in tons of virtual instances.
# humanetech And too they would take the lead in protocol development
# Grishka and btw I'm going to write a FEP about collections that anyone can add objects to
# Grishka that's an important construct and I'm surprised it's not part of the AP spec
# humanetech That's cool.
# Grishka it's useful for many things that aren't microblogging
# Grishka walls are this, photo albums are this
# humanetech Yeah, I'd like to see different domains being added. Lots of them..
# Grishka especially photo albums in groups, where many people can upload photos to one album
# humanetech I'd like to see groups span multiple instance (or instance-independant)
# Grishka what do you mean?
# Grishka the groups I'm making are VK-style, in that you join them (from any instance obviously) and then see various updates in them in your feed
# humanetech Maybe this is already possible. Dunno. But if you have a community, it is not bound to the instance where it was created.
# Grishka can't imagine how this could work
# humanetech Probably can't atm.
# Grishka AP objects have identifiers, and those are URLs, and the domain of the URL, by definition, is the server that has an absolute authority over that object
# Grishka so there has to be one server that hosts the group
# nightpool[m] Matrix does a similar thing with how it sets up rooms, but to me it seems a little wrong-headed.... you're going to almost always want to know which instance is "authoritative" for any given group
# humanetech Yeah, for Matrix I am also curious. I set some stuff up on matrix.org but want to move to feneas.org now.
# Grishka so for Smithereen, as groups can't exist by themselves, the group is hosted on the same server as its creator
# Grishka as in, groups aren't accounts in their own right
# nightpool[m] part of the open-world model AP operates under is that, unlike IRC, things are namespaced and have an authority associated with them. so if groups were "portable" between servers, it would be kind of a step back in that regard
# humanetech The URI thing is in current spec, but might be extend in later versions. How would it work in P2P AP for instance?
# humanetech Where e.g. my mobile phone is my personal actor.
# humanetech Can't host a group.
# nightpool[m] in a theoretical P2P ap, it would be a URI with a different authority scheme, for example something like TOR's public key onion router host names instead of DNS
# Grishka there's no such thing as P2P AP
# humanetech No, but there's talk about it for the future..
# humanetech Would be great if there was a mixed federated + p2p model.
# Grishka but yes, you can obviously run AP in an overlay network, I think Mastodon actually supports running through TOR
# nightpool[m] yes, although i'm just using that as a example of a non-dns based authority scheme
# Grishka the problem with p2p is the user experience — the setup is usually somewhat intimidating, and it can't run in a web browser, you need a client
# humanetech pukkamustard I believe intends to research p2p. Was part of openEngiadina future research, but now moved to https://dream.public.cat/ (I think)
# Grishka if you remove the requirement of TLS, you could, in principle, run AP without domains, using just IP addresses
# Grishka so that's already one possibility
# humanetech And cj is thinking of implementing a gossiping facility in Go
# Grishka and speaking of p2p stuff, I want global search that actually works
# humanetech very important.
humanetech left the channel
# Grishka my basic idea is to use a DHT, this would work for looking up actors by various keys, like centralized account IDs, because that's what hash tables do, distributed or not
# Grishka the bigger problem, though, that I know no solution to, is storing lists to which everyone could append
# Grishka those, obviously, would be extremely useful for things like hashtags
# Grishka nevertheless, the DHT idea solves the dreaded "I signed up for a fediverse account, now what" problem
# Grishka because you would then connect your centralized accounts, or import GDPR exports, or something else, to find the people you already know, wherever they are