#social 2017-05-31
2017-05-31 UTC
# 
@JefferyRPowers The latest De La Crème Marketing Daily! http://paper.li/JefferyRPowers/1425169611?edition_id=d1747ce0-4597-11e7-9df2-0cc47a0d1605 Thanks to @SocialWebWG #socialmedia #seo (twitter.com/_/status/869711700404166660)
fkleedorfer and ajordan joined the channel
# 
@csarven ? #LinkedData Notifications presentation #ESWC2017 http://csarven.ca/presentations/linked-data-notifications-eswc for article http://csarven.ca/linked-data-notifications about https://www.w3.org/TR/ldn/ (twitter.com/_/status/869855041741565953)
JanKusanagi, jankusanagi_ and fkleedorfer joined the channel
herco joined the channel
# 
saranix WTF? Google and Chrome dropped <keygen> support without replacing it with anything!? I'm reading the notifications for both and there's plenty of replies telling them NOT to do that and they still did it anyway!? WTF?
# saranix Haha... subconcious slip. Firefox and Chrome. essentially Google and Google anyway
# saranix so does anyone know of a way to do user enrollment without <keygen>? Keeping in mind that any kind of friction added whatsoever vs email/pass that the masses are used to will stop them from switching to ACTUALLY SECURE decentralized networks
# 
ben_thatmustbeme dropbs link here for those reasons Chrome had for removal https://groups.google.com/a/chromium.org/forum/#!topic/security-dev/pX5NbX0Xack
# saranix Yes and see rebuttals for why it needs a replacement first
# ben_thatmustbeme is not saying either way
# ben_thatmustbeme don't know enough
# ben_thatmustbeme just good to give links
# saranix WebID, a W3C spec, requires it to work. Dropping it as WHATWG did directly conflicts
# saranix see also the Linked Data specs from above, another dependency
# ben_thatmustbeme sort of getting out of social topic, can certainly argue that with whatwg
# saranix many many many social implementations require it
# saranix not at all off-topic
# ben_thatmustbeme "many many"?
# ben_thatmustbeme again, links?
# saranix AHEM. READ. Linked Data, just announced in this channel, AS RELATED CONTENT, requires
# saranix look up
# ben_thatmustbeme what? LDN?
# saranix yes
# saranix https://github.com/solid/solid-spec "Note: Several browser vendors (Chrome, Firefox) have removed support for the KEYGEN element, on which WebID-TLS relied for in-browser certificate generation."
# saranix That S in solid stands for Social ;-)
# saranix just so you know, it is related :-P
# ben_thatmustbeme I don't see that LDN depends on solid, thats not a w3c spec, but anyway, i wouldn't consider https://github.com/solid/solid-platform#servers as 'many many many'
# saranix ugh, you are so dismissive without even listening
# ben_thatmustbeme i know what it stands for, i was there when we made of the abbreviationj
# saranix My platform, and many others that have come out in the last 3-4 years, all do similar things. What our platforms DO, is what will ultimately end up being spec. I haven't gotten solid implemented yet, but it is merely a translation of what my app has been doing for years beforehand
# ben_thatmustbeme the other thing is, the socialwg is chartered to stay away from trying to define new auth methods, can only reference existing ones.
# saranix I'm asking in here, because others in this realm use WebID/ TLS client certs. I've seen it talked about before in here. If you don't have answers, then that's fine, don't feel compelled to throw mud at this just because you can
# ben_thatmustbeme i wasn't, i originally was just posting a link for background
# saranix WebID relies on it, it is existing
# ben_thatmustbeme i think the reasons listed by chrome team are pretty reasonable
# saranix I never ONCE claimed that the reasons the gave are not reasonable. What I said, and I'm getting sick of repeating myself, is that there are plenty of practicle reasons NOT to deprecate it that supersede it
# saranix from the same thread you linked
# saranix read the freaking thread
jaywink joined the channel
# saranix It's also very much a peeve of mine when people use the word "deprecate" when they mean "abandon". https://en.wiktionary.org/wiki/deprecate "To declare something obsolescent; to recommend against a function, technique, command, etc. that still works ***but has been replaced.***" It has not been replaced, that is the entire problem. People are using it, WebID depends on it, and it's gone.
# ben_thatmustbeme chill out, jesus, exactly what i'm saying, link to things and you inform instead of just gripe. Not sure i agree with their reasoning on how extensive the use is, the original poster did discuess that. also note that this was posted 2 years ago, so it pretty late in the game to be reacting strongly now
# saranix It took effect January 2017
# saranix I was unaware until I read a link from a link from a link from the LD link that was posted in here
# saranix You say to provide links but the stuff I'm talking about is in the link you provided, but didn't read, and are now dismissing out of hand
# saranix I'm more upset with YOU than the situation
# saranix You're being nihilist
# saranix taking anti-stance to ANYTHING I say
# saranix very annoying
# saranix Again, I mentioned it in here because I know others are using it. I'm checking with these peeps to see if any of them have come up with a workaround yet. If you don't have one, or anything constructive to say, then please just quit it.
# saranix FYI "Not sure i agree with
# saranix their reasoning on how extensive the use is, the original
# saranix poster did discuess that."
# saranix crap
# saranix anyway, the others in that thread very convincingly disputed the original post
# saranix additionally, I'm telling you that my entire business relies on it, and that when I did research of competitors and others doing similar things, that I found dozens of outfits also doing the same thing, and that I've heard other people talking in this very channel about it too, so I'm giving you testimony that it is in use, and you're contradicting it with your gut feeling. That's messed up.
# saranix Wish I could find the link where tim berners-lee is complaining about how it's an attack on the decentralized social web... but I lost track of it
# 
sandro saranix, there are many people who share your frustration. TimBL's SOLID project relied on it, too, and has been re-engineering to use a style of oauth in its place. https://gitter.im/solid/chat is probably the best place to talk about that.
# saranix thx sandra. OAUTH is a convoluted nightmare. I can't imagine how it could possibly be even close to a solution for this but I'll ask those folks
# saranix gah sandro sorry
# saranix oh gitter needs a github or twitter log in, I don't do silos. crud.
# saranix SMH. so many people talk about dogfooding but none of them do it.
# ben_thatmustbeme oh, nice aaronpk, hadn't see that before
# saranix definitely bookmarking that aaronpk
# ben_thatmustbeme aaronpk++
# saranix aaronpk++
# ben_thatmustbeme hah, quite the list of 'other mentions' going there
# saranix sandro: since I can't log into that chat, do you know of where maybe it was mentioned about their oauth-like alternative? I still can't imagine how oauth solves the problem with generating a client key /CSR in a userfriendly way for client TLS. Are they abandoning client TLS? I'm not seeing the connection.
# sandro saranix, maybe https://github.com/solid/solid/issues?q=is%3Aissue+is%3Aopen+label%3AAuthentication ? although I see those are fairly. You can't login to that gitter because you don't use github, or it has additional access control?
# saranix sandro: yeah I don't use github, and I'm sick of creating throwaway accounts for specific purposes
# saranix hmm. Yeah, remote auth seems to be a 'solved' problem, there is no shortage of sensible solutions. The problem is more with *local* auth, i.e. the software (website) that is hosting the identity confirming the browser in question is the one that owns that account [typically done with passwords, better with TLS client certs]
MMN-work joined the channel
# MMN-work Am I alone on Mumble? .)
# ben_thatmustbeme not anymore :)
# MMN-work I won't be able to talk today but I'll try to listen
# MMN-work So when it's my time to introduce myself someone else could just say the usual that I'm the GNU social maintainer which is an implementation of OStatus with ActivityStreams 1.0 etc.
# ben_thatmustbeme MMN-work: same here
# ben_thatmustbeme waves to rhiaro
# 
jaywink hey all. first time here in SWICG call, though IRC only at least this time. I've also been part of the SocialWG as an invited expert, though not that active most of the time. Independent federated social web hacker involved in own projects currently, aiming to implement activitypub in those. Have been relatively active in the github issues for activitypub.
JanKusanagi joined the channel
# MMN-work 15.00 UTC is now afaik
# MMN-work The previous meetings have been one hour from now though
# MMN-work (in my time zone)
# nightpool MMN-work: they've been 1 hour from now in everybodies timezone :P
# nightpool Does anyone have contact info for cwebber?
# MMN-work cwebber2 is offline on XMPP
# MMN-work in my xmpp zone!
RRSAgent joined the channel
# 
RRSAgent logging to http://www.w3.org/2017/05/31-social-irc
# nightpool Just to sum up what's happened so far in mumble, we're going to start the meeting
# nightpool Sandro called cwebber2, but got voicemail
Zakim joined the channel
# 
Loqi Benthatmustbeme made 1 edit to [[Socialwg/AccountDiscovery]] https://www.w3.org/wiki/index.php?diff=103077&oldid=102999
# nightpool I think someone needs to scribenick me?
# MMN-work present+
# ben_thatmustbeme nightpool, you can scribenick yourself
# nightpool present+
# ben_thatmustbeme present+
# nightpool (thanks ben_thatmustbeme)
# nightpool sandro: The Social Web Working Group hasn't had any particular changes from last week, still working on ActivityPub and WebSub.
# nightpool sandro: w3c advisory committee is still voting on whether to extend the SWWG charter
# nightpool sandro: To work on activitypub for longer, given recent interest
# nightpool sandro: may turn into community spec if charter doesn't get extended
# nightpool aaronpk: micropub has turned into a reccomendation since last week
# nightpool aaronpk: the websub testsuite is up to date, and at websub.rocks
# nightpool aaronpk: we're looking for implementation reports, especially from people running current websub impls
# nightpool aaronpk: like mastodon, etc.
# 
aaronpk details here https://websub.rocks/
timbl joined the channel
# nightpool nightpool: how many changes have their been to websub since PuSH?
# nightpool aaronpk: not many, we're aiming for compatability
# nightpool aaronpk: there's a changelog section in the guide.
# nightpool s/guide/spec
# nightpool aaronpk: we're looking at implementing the above issue
# MMN-work Good addition to the websub spec!
# nightpool aaronpk: because implmentations are doing it
# nightpool aaronpk: and it doesn't break interoperability with existing hubs
# nightpool (thanks!)
# nightpool TOPIC: GH issue #4 - Forwarding
# nightpool sandro: there's been a longstanding question in decentralized systems: what do you do when a server goes down?
# nightpool sandro: issue #1 talks about this a bit, and there's a lot of discussion there
# nightpool sandro: but when I was looking at setting up a w3c mastodon instance, and talking to the systems people
# nightpool sandro: they said that they didn't want to run mastodon, because eventually they would to shut it down, and users would be stranded
# nightpool sandro: But they'd be willing to run a forwarding server, and that seems to be the majority usecase
# nightpool sandro: where people want to shut things down orderly, do sunsets
# aaronpk websub section on redirects: https://w3c.github.io/websub/#subscription-migration
# nightpool sandro: But I wanted to make sure that we talked about it in the spec, and look into what current implementations do for redirects
# nightpool sandro: and behave gracefully in allowing users to move to different sites
# Loqi Benthatmustbeme made 1 edit to [[Socialwg/AccountDiscovery]] https://www.w3.org/wiki/index.php?diff=103078&oldid=103077
# nightpool aaronpk: what's the goal for this call?
# nightpool sandro: experience, mostly, looking at what people have tried
# nightpool sandro: or other feedback
# MMN-work q+ to discuss redirect and URI stuff
# nightpool sandro: I agree that's a concern, but I think that if you're being responsibile, you need to do redirects in perpetuity
# astronouth7303 present+
# nightpool sandro: URLs get printed in books, etc
# nightpool MMN-work: when you do redirects for your URLs, you still have to have an identifier, and you still have to have local databases
# nightpool MMN-work: and you have to figure out what to change in the local database
Denise_Henley__ joined the channel
# nightpool sandro: right, the question here is "when do you consider this permanent", and when do you change your users URL
# MMN-work The point I was trying to get across was that when you move servers the URL/URI for that account stored with other servers is going to have to be updated - if the URI is supposed to be a fully functional URL
# nightpool sandro: Gargron brought this up, that's he's worried about user account hijacking
# nightpool cwebber2: I'm sympathetic to this, although I feel like once you have access to somebodies account you have lots of opportunities to do damage
# nightpool ajordan: I think the problem here is that most things you mentioned can be undone, you can undo side-effects, undelete post, etc.
# MMN-work nightpool++
# MMN-work good summary of my rambling :)
# nightpool ajordan: but once you've been redirected to another server, there's no way to tell that server to "unredirect" you
# nightpool aaronpk: there's a non-technical solution for this too, and a post like "hey, i'm leaving X domain, go find me on this domain instead"
# nightpool aaronpk: and that only works while the servers up, but I'm reluctent to push a technical solution when a solution exists anyway
# nightpool cwebber2: this comes out of solutions like that--both on push.io and on mastodon, I've seen lots of people leave messages like that, which basically imitate the behavior we have here
# nightpool cwebber2: but you're right, people do work around it at the moment
# nightpool ajordan: I may be architecture astronauting, but I think it's really cool how standards have interpo with non-technical systems
# nightpool s/interpo/interop
# nightpool ajordan: but if you ??? you'll break interop
# ben_thatmustbeme s/push.io/pump.io/
# MMN-work aaronpk & sandro: I'm a proponent of making social systems requiring social interactions, including some things that are "too manual" .)
# ben_thatmustbeme ajordan anyone can correct the scribe
# nightpool sandro: Am I hearing any disagreement? Or do most people think that systems should handle redirects cleanly?
# nightpool nightpool: I think the 30 day heuristic aliviates my concerns
# nightpool cwebber2: appologies about missing the time
# MMN-work maloki?
# MMN-work Marie
# nightpool TOPIC: Using Discourse for activitypub
# nightpool cwebber2: Maloki and Gargron brought up hosting a activitypub category on the mastodon discourse forum
# MMN-work cwebber2: I am text-only todat
# nightpool cwebber2: I know that there's some existing people who don't feel comfortable using Github
# MMN-work today*
# MMN-work You can continue discussing
# MMN-work and I'll write something up
# nightpool cwebber2: And this could provide an alternative, although we would have to talk about bridging convos back and forth
# nightpool ajordan: chris I'm definitely with you on fragmented conversations, I find most things beside issue trackers hard to use
# nightpool ajordan: Is there a compelling reason to keep tracking for this group on Github, rather then switching to something like Gitlab?
# nightpool ajordan: Because that seems like it would solve both problems.
# nightpool cwebber2: this came up in the working group. The problems raised is that many people on the working group used Github, and they would have to create another account (federated identities...)
# nightpool cwebber2: and the w3c has some amount of tooling around it, although that's unclear at this point
# nightpool sandro: I'm not aware of any tooling, although I might be missing something?
# nightpool ???: I think it has to do with backing things up for legal purposes
# nightpool sandro: But we're not under the w3 org, so that wouldn't help us really much anyway
# nightpool cwebber2: Discourse does seem to walk a line between issue trackers and traditional forum software
# nightpool cwebber2: allows closing/resolving threads, etc.
# nightpool ajordan: as another note, gitlab has a "sign in with github" button, which may alleviate conerns about additional accounts?
# nightpool sandro: I just tried using Discourse this morning, I had never used it before and I liked it.
# nightpool sandro: I liked the forwarding discussion I saw there, and I linked one way but should like the other way
# nightpool sandro: If the idea here is to use Mastodon's discourse installation, would other projects feel excluded?
# nightpool cwebber2: It might be, I don't know. We would have to ask
# ben_thatmustbeme we need to make our own decentralized social system so we can design a decentralized social system
# nightpool cwebber2: I wouldn't feel that about my own projects, but I know that there might be some concerns there, and we would have to hear from other people
# nightpool cwebber2: another concern is what happens to our discussions if their installation goes down
# ben_thatmustbeme github had its downtime today
# nightpool cwebber2: And that's a concern for Github as well, big services (such as google code) have gone down in the past
# MMN-work I'll paste this:
# MMN-work I am personally not using GitHub because I cannot accept their TOS. I agree that using a separate discussion forum is much work (and I'm probably the only one currently to actively avoid GitHub).
# astronouth7303 q+
# saranix MMN-work: I actively avoid github too
# MMN-work Suddenly GitHub (or whatever third party we host through) could change TOS, forbid a certain participant to log in because of actions in some _other_ repository etc.
# MMN-work The bottom line is I don't think it should be hosted on a domain not controlled by the community of SWICG. Also "multiple accounts" is very little an issue if the tool used has third party logins (OpenID/OAuth)
# nightpool ajordan: I think my main concern is about how other people perceive this, we know that we just "happen" to be there, but the average observer would not
# nightpool ajordan: so I'm a tenative -1 between backup concerns and that.
# MMN-work cwebber2++
# saranix agree should be hosted under w3 control. Also agree with bad optics of Mastodon, even if control wasn't an issue. optics being that they are very silicon valley hipster of the day
# nightpool sandro: does gitlab solve this problem? or would we have to have an installation of gitlab or an installation of discourse somewhere else?
# nightpool ajordan: I think gitlab.com would solve these problems?
# nightpool aaronpk: gitlab.com seems like it would just be moving to another 3rd party service, which doesn't solve the conerns MMN-work raised.
# astronouth7303 aaronpk: gitlab can be self-hosted for free. It's FOSS+Premium.
tantek joined the channel
# nightpool sandro: I can't find evidence of it right now, but I think MIT may have a gitlab hosting thing? Which we may be able to use, given that w3c is somewhat part of MIT?
# MMN-work sandro aaronpk: Currently GitLab hosts git.gnu.io for us, so if we had a domain name to use I'm sure they could take care of the hosting part
# nightpool sandro: i would prefer it be w3c branded, but that may be another option
# MMN-work sandro aaronpk: us = GNU social
# nightpool cwebber2: It sounds like we don't want to use discourse, and we can continue this conversation on better venus later.
# nightpool s/venus/venues
# astronouth7303 q-
# nightpool aaronpk: we can continue this discussion on IRC or other forums.
# nightpool cwebber2: we have a little bit less people here then usual, do we want to get this shortname thing over with or postpone for another week?
# nightpool TOPIC: shortnaming
# nightpool aaronpk: naming is always a rabbithole, I would consider punting
# nightpool cwebber2: we can do this discussion async
# nightpool sandro: let's poll the channel on consensus for the name
# nightpool sandro: and maybe we can set it aside quickly
# saranix shortnaming?
# nightpool cwebber2: okay, 1 minute summary: our full name is set in stone, but we're considering two options for the short name
# astronouth7303 saranix: https://www.w3.org/wiki/SocialCG/2017-05-31
# nightpool cwebber2: one option is SWICG, which is hard to pronounce but keeps the "incubator" and "web" aspects, which may be important to you
# nightpool cwebber2: the other option is SocialCG, because it's more pronouncable and implies more continuity with SocialWG
# nightpool cwebber2: and we mention the incubator and web aspects on the wiki page heavily
# MMN-work tantek: No problem, we could register TheReal$shortname
# nightpool aaronpk: Just to mention where this would be used--it's the namespace for the wiki page, it's the account for social media, and on the w3 url for the group.
# MMN-work SWICG
# nightpool SocialCG
# MMN-work (don't really mind though)
geppy joined the channel
# geppy Sorry I'm late.
# ben_thatmustbeme walked away for a little bit
# geppy + SocialCG
# ben_thatmustbeme SWICG++
# saranix SWICG++
# astronouth7303 +1 SocialCG, if only because it looks wordish, not just a jumble of letters.
# nightpool cwebber2: Looks like 8-3, which isn't a complete landslide victory, and it doesn't capture everybody
# nightpool cwebber2: but it does seem to be the leaning here
# nightpool sandro: that's what we're currently using on the wiki, right?
# nightpool (rhiaro, sorry, that was my fault: cwebber called for votes but I didn't transcribe that part)
# nightpool cwebber2: this feels extremely bikesheddy, and we do need to choose one
# nightpool cwebber2: does anyone feel strongly that they would be upset if one of these was chosen?
# geppy I'm fine with SWICG being used for the wiki or whatever, and I'm fine with CWICG being used, but I feel like it's clearer especially to outsiders if I talk about SocialCG
# nightpool cwebber2: we're seeing some preferences, and maybe some strong preferences
# nightpool cwebber2: and it looks like we might not be in trouble from the people here if we choose socialcg
# nightpool cwebber2: we might want to throw this out to the rest of the world, or decide on this once and for all here
# saranix could go both ways when talking to outsiders. They could get confused about diff between CG and WG, also could be missing the Incubator and Web apsects. OTOH, SWICG is opaque at first glance as well.
# nightpool cwebber2: aaronpk do you have a preference whether we decide now or not?
# nightpool aaronpk: I see arguments for both sides, but we should probably close this discussion sooner rather then later
# nightpool cwebber2: should we do a resolution?
# nightpool sandro: that would be nice
# nightpool s/nice/useful/
# nightpool +1
# geppy +1
# MMN-work +0
# MMN-work That's how you do it right? .)
# geppy tantek: Someone registered @SocialWebCG on Twitter in May of 2017, I'm guessing that means right now someone here grabbed it.
# MMN-work Cool, thanks for the explanation.
# astronouth7303 +1
# saranix -0
# nightpool cwebber2: we don't have any -1s, and we did do this based off of the poll
# nightpool cwebber2: it's a mix of positive and wishywashy, so I feel we can probably close this
# nightpool aaronpk: we are at the top of the hour, and that's the end of our scheduled time
# nightpool aaronpk: we still have three or four things on the agenda
# nightpool aaronpk: but we can probably wait until next week?
# MMN-work agreed
# nightpool cwebber2: yes, meetings are weekly now
# nightpool aaronpk: everyone is welcome to continue chatting on IRC or on the call, it just won't be part of the official minutes
# geppy Thanks, all, sorry I missed everything.
# nightpool cwebber2: we oficially finished painting a bikeshed!
# ben_thatmustbeme nightpool++ for minuting
# nightpool do I do generate minutes, or should someone else?
# RRSAgent I have made the request to generate http://www.w3.org/2017/05/31-social-minutes.html trackbot
# aaronpk agenda for next week is up https://www.w3.org/wiki/SocialCG/2017-06-07
# nightpool sure, no problem!
# nightpool I don't believe I can access the drafted minutes though?
# nightpool Oh, right, I never fixed my wike login :(
# nightpool s/wike/wiki
# nightpool no, I think it's fine, I just need to go change my password at w3c
# geppy I didn't set an alarm because I normally wake up very early in the morning. I'll start setting an alarm for these. I've been finding out about them based on sandro's GitHub posts, but can we get some consumable schedule? Perhaps an ActivityStream?
# nightpool what was the limit on wiki passwords? 19 chars?
# geppy is interested
# ben_thatmustbeme knowing aaronpk, he'll have a new system for this written by next week, lol
# astronouth7303 on discourse usability: I don't think mixing communities in a single discourse instance is a good idea. Discourse doesn't present that well.
# astronouth7303 although i missed the bit where people don't like discourse.
# nightpool astronouth7303: I think the concerns were mostly around that, and concerns around backup/longevity of the mastodon discourse instance
# astronouth7303 if there's going to be discussion on hosting our own, i figure we should put discourse back on the table for a moment
# nightpool aaronpk: Discourse does have a "mailing list" mode, which I've had good experience with
# geppy The thing I like about GitHub is that (I think) you're massively expanding the pool of people who might: find what you're doing, observe, or participate.
# nightpool aaronpk: No, I mean, you wouldn't use the JS interface
# nightpool you'd just use email to interact with it
# saranix geppy: that's what they want you to think. Also, it's no coincidence that they set themselves up to suddenly be the host of all non-profit codes. It gives them big data insights.
# nightpool cwebber2: can you see the minutes? They're not showing up for me.
# nightpool Not sure if that's just because they're still not done, or b/c I don't have access.
# geppy saranix: I realize that. But another WG I participated in had a really well-informed discussion about this, and that's what we ended up concluding.
# astronouth7303 if we go with gitlab, i'm pretty sure we'd want a self-hosted Gitlab Community Edition
RRSAgent joined the channel
# RRSAgent logging to http://www.w3.org/2017/05/31-social-irc
# RRSAgent I have made the request to generate http://www.w3.org/2017/05/31-social-minutes.html sandro
# astronouth7303 gogs vs gitlab depends on features. Gogs is simple, but gitlab has soooo much more
# geppy For what it's worth, that WG was an inclusivity working group, so "how do we increase diversity" was an area of expertise for the members.
# astronouth7303 if we have a gitlab, would we be moving the repos and publication to it as well?
# astronouth7303 i can see us making use of the CI and Pages
# saranix geppy: A bit unique that this group is underneath it all about decentralization, and leveraging centralization for inclusivity would be hypocritical, it would be declaring that you can't be inclusive if you are decentralized, which is absolutely false
# astronouth7303 seeing as we produce documentation in source code, which gets turned into HTML and thrown at a website.
# geppy And GitHub sucks at notifications, but there are tools on top of it like https://github.com/octobox/octobox
# geppy saranix: I mean, I'd love it if we selfdogfooded on this. We could even have a service/bot that posted activity to+from GitHub.
# astronouth7303 ... do we even have an idea of how ActivityPub is applied to group discussions?
# astronouth7303 just checking
# geppy ajordan: You're welcome!
# saranix astronouth7303: I haven't gotten to that stage yet but I just assumed that it would be like zot where certain id's simply behave as forums, and forward posts to interested parties.
# MMN-work ajordan cwebber2: it replicates the behaviour of mailing lists (but everyone apparently dislike mailing lists for some reason)
# MMN-work enjoys mailing lists
# astronouth7303 list management is a thing, which i think eventually ties into cross-provider authn
# astronouth7303 ie, how does someone add themselves to a group?
# saranix astronouth7303: well for my impl (and zot), they connect, like connecting to a person
# saranix 'friending'
# geppy I'd be interested in following an as:tag
# astronouth7303 this breaks LD signatures and e2e, doesn't it?
# MMN-work ajordan: Yeah, the UI integration with issue tracking is what mails don't have. The amount of "unsubscribe" failures to your average 'mailman' list shows all clients haven't caught up yet :]
# saranix astronouth7303: it does when federating with diaspora, but that's because of the enveloping. It's solvable.
# MMN-work (though Thunderbird and Evolution I think understand the mail headers which indicate how to unsubscribe etc.)
# astronouth7303 a forwarding agent (list, group) can do one of two things: forward the message unmodified (possibly leaving providers confused about who should get it or running afoul of spam prevention), or resolve the collection in the destination fields (modifying the message and potentially invalidating signatures applied).
# MMN-work ajordan: I think the main issue is that issues are always centralised and thus have implementation issues on decentralised, federated systems :] (because why should your "solved" mean that I can't continue discussion via _my_ mail account in that thread? .D)
# astronouth7303 Both are impossible with e2e without public or semi-public membership lists
# astronouth7303 and I'd have to look at the spec again to apply it to LD Signatures
# saranix astronouth7303: not true. each subsriber has a one<->one relationship with the forwarding agent. This handles spam validation and keeps signatures intact.
# MMN-work ajordan: (I wrote that with the premise that everything could be solved with more mail headers, indicating whenever a thread is solved etc)
# MMN-work a thread/issue
# astronouth7303 so basically the forwarding agent just "retweets" everyone?
# astronouth7303 needs to read the specs again
# saranix yep
# MMN-work ajordan: I
# saranix err, not sure in relation to ActivityPub, I'm speaking on the protocols I work with
# MMN-work ajordan: I'm getting a bit off-topic, I'll stop writing distracted thoughts from my head now. .)
# astronouth7303 saranix: what protocols do you work with?
# saranix but I assume it's the same in ActivityPub
# saranix or at least declarable
# saranix I'm an optimist :-)
# astronouth7303 I think attributed reshares would be under AS2, which i have not read through yet
# saranix zot and *redacted* ;-)
# astronouth7303 s/reshares/republication/
# saranix yeah not reshare
# astronouth7303 we're going to have spec terms and UI terms, and it's going to be less good
# nightpool oh, nice.
# ben_thatmustbeme good call aaronpk
# nightpool sandro: also, wasn't able to mention this during the call, but I did find your github post useful (I thought I +1'd it?)
# nightpool I edited the meeting minutes, but I'm still having trouble with my mediawiki account
# nightpool If I post a gist, could someone copy it into https://www.w3.org/wiki/SocialCG/2017-05-31-minutes ?
# nightpool aaronpk: thanks!
# aaronpk nightpool: done https://www.w3.org/wiki/SocialCG/2017-05-31-minutes
# nightpool yeah, kind of the purpose of reactions is to replace people commenting "+1", because it was clogging up notifications
Coco joined the channel
# astronouth7303 tantek: i still think it's better than a lot of the tools available prior
# astronouth7303 have you ever used phpBB?
# astronouth7303 yeah, we heard you the first time
# saranix what's the right question?
# saranix woah. I have to hard disagree on that one. See my comments in scrollback.
# ben_thatmustbeme wow, phpBB, that goes back a ways, haven't heard anyone speak its name in years astronouth7303
# nightpool there that it is seriously self-defeating to put an *open standards effort* someplace else
# saranix w3.org is not 'wherever'
# nightpool whoops
# nightpool sorry, I did not mean to send that
# astronouth7303 and the members that _are_ being dogmatic about it?
# saranix This entire CG is a failure if it has to centralize to work
# astronouth7303 there's a difference between "intrinsically needing something" and "have put in the engineering effort to do something"
# saranix The point of this group is to create stuff that "just works". All we have to do is do it
# saranix I also think we are way past the bootstrapping phase
# saranix this stuff has been discussed for years
# 
tantek e.g. self-hosted repos (mirrored to github) http://indieweb.org/git#IndieWeb_Examples
# astronouth7303 so what's concretely required to happen in order to move discussions to a decentralized platform?
# tantek and then self-hosted issues, comments etc, copied to GitHub: http://indieweb.org/GitHub#POSSE_to_GitHub
# ben_thatmustbeme so, the point of this group is to experiment with things, thats why it is an "Incubator" community group. Just pick something and move on.
# astronouth7303 we have mastodon and gnu social, and i believe they're interoperating.
# saranix tantek: you are contradicting yourself
# astronouth7303 so we have some platform
# astronouth7303 what's concretely required to happen in order to move discussions to a decentralized platform?
# astronouth7303 define "platform"
# saranix yes, but my "just works" comment was WRT your "just works" comment about why github
# tantek astronouth7303: http://indieweb.org/platform
# astronouth7303 we have protocols and software that implement them.
# astronouth7303 not all the protocols have complete implementations.
# astronouth7303 there's a bunch of test suites
# astronouth7303 and i think it was said that pump.io implements forwarding agents?
# astronouth7303 no, none if it is complete, but it feels complete enough to move towards dog-fooding
# astronouth7303 a forwarding agent is the thing that would actually implement lists/groups
# saranix jaywink: old way of thinking. ActivityPub and Linked Data is set to reverse that
# astronouth7303 if you just mean archiving for posterity, you could make a webhook that stashes stuff to files in an afternoon
# saranix jaywink: s/could/should
# astronouth7303 tantek: yes, everything we're discussing would involve technical effort.
# astronouth7303 .... you're just being pedantic now.
# tantek cwebber2 the folks doing http://indieweb.org/git#IndieWeb_Examples and http://indieweb.org/GitHub#POSSE_to_GitHub
# nightpool aaronpk: I tried working through some of the websub tests, but it seems that (as far as I can tell) mastodon passes none of them. Which ones were you expecting it to pass?
# saranix I abhor POSSE. I feel like it is an afront to what we're trying to do here.
# tantek cwebber2 http://indieweb.org/git#IndieWeb_Examples are doing federated pull requests
# tantek and http://indieweb.org/GitHub#POSSE_to_GitHub are doing federated issues and comments
# saranix tantek, to me, that's like trying to lose weight with portion control alone
# saranix tantek, terrible analogy, I'll admit. I'm trying to convey that just be reducing dependency on evil does not free you from evil
# astronouth7303 a VCS+PM extension to AS2 would be interesting, though
# jaywink I would think of federated git a bit like a traditional social media (decentralized) site. Like Diaspora, as an example. My profile would contain several git repos. WHen I post a PR it would be pushed to my followers, same for comments and such. Any anyone can subscribe to one of my repos, getting updates about. The repo is till located on my server.
# saranix tantek, but still allows silos to be a source of YOUR content, they still get aggregation benefits. The shortest distance between 2 points is a straight line and not a zig-zag. Allowing silos to be part of the end game means you will never get to the end. They are diametricly opposed to the goal.
# nightpool Pull Requests are just basically a Note with a specialized @-mention, right?
# nightpool they are "I request you pull this code from me"
# nightpool with threaded discussion
# saranix tanktek, if you admit they are not the end, then you can't get to the end by passing through them
# astronouth7303 @nightpool there can be additional data attached. GitHub/Gitlab actually attach commits to PRs
# saranix you have convinced yourself that they will eventually disappear but if you keep feeding them content then they never will.
# tantek saranix, "they never will [disappear]" also false, by evidence: https://indieweb.org/site-deaths
# tantek I mean goodness sakes: https://indieweb.org/site-deaths#Google_Code
# astronouth7303 cwebber2: be sure to look in to some of the deeper features (eg, CI integration)
# astronouth7303 collaborative documents (ie, wiki, snippets) might be out of scope
# astronouth7303 possibly?
# saranix tantek, none of what you say rectifies that in order to put anything on github, one has to decide that they do not care about the evil that github is doing with that participation metadata. Since I'm someone who cares, I'm left out. How is that 'reaching more people'... it isn't.
# astronouth7303 interestingly, gitlab and github attach CI data to commits, not PRs
# astronouth7303 (note: gitlab calls them merge requests)
# astronouth7303 so a CI status could be an activity in reply to a commit activity?
# saranix So if I'm someone who cares about decentralized end goal, and I want to participate, but I can't because a central "trade-off" is required, and I don't want to make that trade off, then that is an impediment to the end goal. Do you get what I'm saying yet?
# astronouth7303 i was thinking that pull request was the more accurate in a decentralized system? :shrug:
# astronouth7303 so i guess a PR activity would need CommitAttach activity to actually link it to commit activities?
# saranix I believe that the cost to the individual is a great cost to the community (and it's principals), that's where we diverge I guess
# astronouth7303 since commit activities are attached to branches, if anything
# saranix What I find really strange about your reasoning, is the idea that people use github like they use facebook... like they're on there everyday surfing for projects. To me it doesn't matter if the link is to github.com, w3.org, or heylookat.me... I see no network effect.
# astronouth7303 jaywink: under current implementations, but the data schemas don't require that
# astronouth7303 fair
# astronouth7303 not intrinsically?
# astronouth7303 you have commits before the PR is made
# astronouth7303 and you're assuming that all implementations want to immediately attach commits to active PRs
# astronouth7303 i work a _lot_ with github/gitlab
# astronouth7303 oh god, i hate --force
# astronouth7303 it ruins everything
# astronouth7303 but yeah, you can't have a commit in a pr/mr without the parents up to the common ancestor
# astronouth7303 holds the opinion that you should not fight git's immutable history
# saranix tantek++
# astronouth7303 i disagree, and surprise feature rebases have bitten me before
# astronouth7303 (but yes, lots of people follow rebasing and/or squashing workflows)
# astronouth7303 (i just think they're wrong)
# astronouth7303 although note: from a code stand point, `master` isn't special
# saranix tantek, appreciated
# astronouth7303 assuming you're tying your extensions to git
# astronouth7303 (fun times: make it VCS generic)
# astronouth7303 so is the project a forwarding agent, or do you break up a project into several agents?
# nightpool I don't think that pushing commits into the social sphere is a good idea
# nightpool git is already decentralized
# saranix I think it's worth mentioning too that the 'absolutists' that insist the most on avoid silos are likely the most passionate contributors to the cause, and by requiring silos you are excluding the most passionate
# nightpool The only thing that should be represented as AS2 objects is the conversation
# nightpool which DOES include inline comments but does NOT include commits
# astronouth7303 nightpool: i think there's too much referring to commits to not have them represented at all. At minimum, you need a URL scheme for commits.
# astronouth7303 and CI requires a pubsub mechanism to be notified of activity
# astronouth7303 (ie, pushes)
# sandro When you say URL scheme, you mean a real new URI scheme, like gitcommit:xxxxxx... or just a mapping to working URLs like https://host.example/commit/xxxxxxx.... ?
# saranix tantek: You look at it as optimizing for the most passionate vs pragmatists, I look at it as making an explicit judgement call to *exclude* the passionate.
# astronouth7303 it needs to map to JSON-LD, so if it's HTTP/S, it needs to be LD. If it's `git+`, then you can say "oh, here's how to map it to git's protocols"
# astronouth7303 and the LD representation is virtual, then
# astronouth7303 but then you can't tie, say, a commit author to a AP agent
# astronouth7303 (or hg+, bzr+, mtn+, ...)
# astronouth7303 (ok, you could have a distributed table mapping git emails to agent URLs, but ugggggggggggggg)
# astronouth7303 so yeah, having actual commit AS2 objects would be extremely verbose, but I suspect that they become a necessity.
# ben_thatmustbeme lol, sandro, you just noticed that, loqi only says that to Zakim, I think there is some bot rivalry there
# ben_thatmustbeme or at least that i have seen
# astronouth7303 i just assumed it was hard-coded, because some developer was bored
# 
ajordan will read later but in the meantime I thought I'd just leave this here: https://gitlab.com/gitlab-org/gitlab-ce/issues/4013#note_27767988
# astronouth7303 problem is i don't think there's anyone really comparable to gitlab, so i don't think interoperability is real high on the list
# astronouth7303 *shrug* probably to be expected to some degree, and we should probably take the strangelove approach to it.
KjetilK joined the channel