Socialwg/2017-10-10-minutes

From W3C Wiki

Social Web Working Group Teleconference

10 Oct 2017

See also: IRC log

Attendees

Present
rhiaro, aaronpk, ajordan, sandro, eprodrom, cwebber, tantek
Regrets
Chair
eprodrom
Scribe
cwebber, ajordan

Contents



<cwebber> scribenick: cwebber

<eprodrom> chairnick: eprodrom

<eprodrom> https://www.w3.org/wiki/Socialwg/2017-09-19-minutes <-- review now

+1

<eprodrom> PROPOSED: approve https://www.w3.org/wiki/Socialwg/2017-09-19-minutes as minutes for 19 Sep 2017 meeting

+1

<rhiaro> +1

<sandro> +1

<eprodrom> +1

<ajordan> +1

<aaronpk> +1

<ajordan> have people ever -1'd minutes?

RESOLUTION: approve https://www.w3.org/wiki/Socialwg/2017-09-19-minutes as minutes for 19 Sep 2017 meeting

<ajordan> ahh

eprodrom: next question is November telcon meeting
... we have a meeting on the 10th and 24th; Tantek has asked about future meetings which would be 11-1 I believe?

sandro: No, October 31st

eprodrom: and then 11-14 I guess?

sandro: sorry it would be 10-24 and 11-7

eprodrom: how about we do the 14th and the 28th of November

<eprodrom> PROPOSED: Telecons on 14 Nov 2017 and 28 Nov 2017 unless others are necessary

<ajordan> +1

+1

<sandro> +1

<eprodrom> +1

<rhiaro> +1

<aaronpk> +1 works for me

RESOLUTION: Telecons on 14 Nov 2017 and 28 Nov 2017 unless others are necessary

eprodrom: we'll still have one more meeting before TPAC but while we're still doing admin stuff I wonder if people who are...

plans for TPAC

eprodrom: oh I think we discussed that people who are at TPAC will do a BOF?

<rhiaro> cwebber: you emailed about space for cg right?

cwebber: yes BOFs, and yes I emailed about CG but haven't heard definitively

sandro: CGs are happening, follow up

cwebber: ok

<Loqi> tantek: ajordan left you a message 1 week, 5 days ago: ugh, IWC NYC looks so fun! probably too late to get train tickets, etc., however I *did* announce it to the folks in the programming institute I was in over my gap year (Recurse Center, https://www.recurse.com/) so hopefully you get some alumni and in-batch folks who show up!

<tantek> I will be at TPAC

eprodrom: I assume among the three of you one of you will work out a BoF

cwebber2: yes

cwebber: yes

tantek: I can probably do one, the BOF rules may change this year... I'll explain more later

eprodrom: I will ask cwebber, are cg members able to join

tantek: CG would be cutting edge discussions, whereas BoF would be a bit more like what we had in Portugal: more demos of more specs taken to REC, etc... trying to gather interest for the CG

eprodrom: excellent, so there seem to be reasons to plan both

tantek: yes, and I'd like someone else to do the BoF

cwebber: I can probably organzie the BoF

Mozfest

eprodrom: I want to talk about another event at the end of the month

<rhiaro> ohhhh I'm back and forth on going to Mozfest

eprodrom: I'm facilitating a discussion about the AP network and where things are. dunno if anyone else is going to be there but wanted to get that out there

<rhiaro> maybe I wll then

eprodrom: tantek fyi we're doing nov 14th and 28th, is that ok?

tantek: no objections

ActivityPub

eprodrom: if there's any other admin issues, otherwise we'll move on to AP

<ajordan> scribenick: ajordan

<eprodrom> Can we get a temp scribe?

eprodrom: so cwebber where are we with uhhhh ActivityPub?

cwebber: main thing is the test suite and gathering implementation reports
... last week we decided I'd do the more prompty thing
... but also I was at Rebooting Web of Trust and had some client works
... but I have some time in the next couple weeks and the prompty thing should be pretty straightforward
... if I have time to refactor it and pretend to be a server great, but I think the priority is getting something that works
... questions? or I'll move on

sandro: I don't want to do the math about exact deadlines again but can we say it'll definitely be done by the meeting in two weeks?
... and if at say the halfway point if you think you won't get there you ask for help?

cwebber: yeah I'm hoping to get it done by the end of the week
... will check in with the group on IRC next Tuesday when we don't meet

sandro: not sure everyone's on IRC but... I probably will be so it'll be okay

cwebber: yeah feel free to ping me

<cwebber> https://github.com/w3c/activitypub/issues/239

<Loqi> [puckipedia] #239 mediaUpload: only post to outbox if it's a Create activity?

cwebber: spec's pretty good but there are two issues we need to resolve
... I hate to say it but I think we should drop mediaUpload endpoint, it's at risk
... just given our timeframe
... obviously this needs to happen for my sake for MediaGoblin so it won't be dropped on the floor entirely
... this is just AP proper
... does anyone have opinions? should I make this a proposal?

eprodrom: if we took out the uploads endpoint would we be able to upload things using literal data ??? or something?

cwebber: what you could still do (and I have done this) is you could point... tantek are you typing?

<tantek> sorry. I was going to ask about existing implementations of mediaUpload

cwebber: what you can do and I've done this on my own server when you slot in the URL for C2S video, you can put *any* URL in there
... you can use an external host as a separate step

tantek: my biggest question is, where are we with impls of it? client and server?
... just in your estimation, nevermind tested

cwebber: I've got an impl, puckipedia has an impl, but we both discussed wanting to change it

tantek: in a non backwards-compatible way?

cwebber: yes

tantek: this is exactly what CR is for, try to implement it, see if there are problems
... based on your experience it seems like the spec needs to change and we're running out of time

cwebber: that's exactly why we marked this at risk too
... it's the least tested part of the spec and if it's not in there the protocol still works just fine

tantek: that's a wise way to make sure the spec is in shape

cwebber: obviously I'm not proposing this since it's the most exciting for me
... but I feel confident we can do it in an extension

tantek: and certainly better than prematurely specifying something that's wrong

cwebber: yep
... should I type up a proposal?

tantek: I'll leave the floor open if anyone else wants to provide opinions

  • silence*

scribe: if no one else has opinions go ahead

eprodrom: yeah why don't you write it up cwebber

<cwebber> PROPOSED: Resolve https://github.com/w3c/activitypub/issues/239 by removing mediaUpload and specified behavior from ActivityPub spec proper, move to extension via SocialCG process

<Loqi> [puckipedia] #239 mediaUpload: only post to outbox if it's a Create activity?

<cwebber> +1

<ajordan> +1

<eprodrom> +0

<aaronpk> +0

<eprodrom> actually -0

<sandro> +0

cwebber: eprodrom I'm guessing your -0 is "it sucks to not have this in the spec"?

eprodrom: yeah

cwebber: I feel the same way even though I +1'd

<aaronpk> same

<tantek> +1 it's the right thing to do to make the spec more solid, and avoid implementations coding the wrong thing

RESOLUTION: Resolve https://github.com/w3c/activitypub/issues/239 by removing mediaUpload and specified behavior from ActivityPub spec proper, move to extension via SocialCG process

<Loqi> [puckipedia] #239 mediaUpload: only post to outbox if it's a Create activity?

eprodrom: yeah I understand the reasons it just seems like an important thing we didn't manage to get in
... okay, looks like we're there

cwebber: I'm gonna copy it to the issue and deal with the rest of it later
... okay so the next one is...

eprodrom: cwebber could I ask that you start a new document with that content ASAP?

cwebber: yeah I'll even put it as a topic for the SocialCG which meets tomorrow
... lemme put that on the issue too so I don't forget it
... okay
... alright great, ummm, as great as you can get

<cwebber> https://github.com/w3c/activitypub/issues/260

<Loqi> [yvolk] #260 Relation between Actors and Users of servers is undefined

cwebber: I'd like to talk about issue 260
... it's about actors and users and their relationship
... the author of it is the author of an important impl, AndStatus, so we do want this feedback
... there's a part of this I think is correct and we should address
... there's also a part of this that is conflating things
... the correct thing is that in the non-normative intro to the spec we say "here's a user" and that's incorrect because there's not necessarily a 1-1 mapping between actor/user
... actor could be a service, a user could have more than one actor
... there might not be a user for a given actor
... but the spec is also talking about there being... they raised oh well we don't have a way for a person to move between multiple servers

<cwebber> https://github.com/swicg/general/issues/1

<Loqi> [sandhawke] #1 Follower Migration

cwebber: we've talked about this and we have an issue for follower migration in the SocialCG

<rhiaro> Thaaaat sounds like an implementation detail

cwebber: I think the SocialCG is the right place to discuss that, I don't think we should add that to AP
... the other thing is they suggest we should model in the relationship between users/actors
... I'm pretty -1 on that
... we don't need it in the spec and there's a reason there's a difference
... options:
... a) ???
... b) take it to the CG
... c) add(?) something to the spec but we don't have a lot of time and it'll be very complex

eprodrom: we have a queue

<cwebber> cwebber: a) clarify that users and actors are not one to one relationship

rhiaro: the issue of follower migration is definitely an impl detail
... no amount of data modeling will make it easier or harder
... in summary I agree with everyting cwebber said basically
... we had this whole discussion in ActivityStreams on how to model this

<cwebber> cwebber: c) don't add complex User and Actor modeling which will muddy up the spec

rhiaro: I don't think the spec in any way constrains implementations in the way this guy is suggesting
... but maybe a sentence noting that there is *no* 1-1 mapping

eprodrom: I see one possibility talking about data portability

<eprodrom> data portability story = register your own domain

eprodrom: probably the easiest one is register your own domain

<tantek> eprodrom++ :)

<Loqi> eprodrom has 48 karma in this channel (49 overall)

eprodrom: that is clearly the easiest way to get data portability
... it also happens to be the wayt o get data portability in many other systems
... may be worth calling that out
... telling implementors to allow users to use their own domain

<eprodrom> 1 user -> 1 person

eprodrom: the other thing is yes, it's not necessarily the case where one user would be one person but that's often the case

<eprodrom> 1 user -> 1 organisation

<eprodrom> multiple user -> 1 organisation

<eprodrom> 1 user -> multiple personae

eprodrom: I think those are the most common formats
... it gets pretty crazy in there but it may be worthwile saying "these are some comon ones, the spec doesn't define what other possibilities there are"

aaronpk: just wanted to say that there's so many reasons tying to real people is a terrible idea
... I'm all for doing whatever it takes to keep an AP user as an AP user without any idea of what that means in a real-world space
... I like eprodrom'd idea to just get a domain and use it as an id
... that simplifies the identity aspect without muddying it with real people's identities

cwebber: we have an empty queue, should I follow it up with a proposal?

eprodrom: SURE

cwebber: mkay, gonna type one up

<cwebber> PROPOSED: Resolve https://github.com/w3c/activitypub/issues/260 by clarifying that there is no specific mapping of one user to one actor (and there can be many configurations) in the spec; leave follower migration to SocialCG, and do not add extra modeling of mapping real-world Users to Actors

<Loqi> [yvolk] #260 Relation between Actors and Users of servers is undefined

<eprodrom> +1

<cwebber> +1

+1

<aaronpk> +1

<sandro> +1

<ajordan> and I guess whether to add a list of "common" mappings falls under editor's discretion

RESOLUTION: Resolve https://github.com/w3c/activitypub/issues/260 by clarifying that there is no specific mapping of one user to one actor (and there can be many configurations) in the spec; leave follower migration to SocialCG, and do not add extra modeling of mapping real-world Users to Actors

<Loqi> [yvolk] #260 Relation between Actors and Users of servers is undefined

<tantek> +1

eprodrom: alright let's mark this as resolved
... that was your second issue cwebber, is there anything else on AP today?

cwebber: nope we got through all of it

eprodrom: great
... so we're at 40 minutes but we've got a 90 minute schedule today
... let's move on and talk about WebSub

<cwebber> scribenick: cwebber

WebSub

<eprodrom> ajordan++

<Loqi> ajordan has 20 karma in this channel (21 overall)

eprodrom: aaronpk where are we at with WebSub

aaronpk: PR was published Tuesday, which is great news... "what's next" is my question

sandro: I can take a look at the survey but I think the main thing is to remind people to vote

<sandro> AC Reps should vote at https://www.w3.org/2002/09/wbs/33280/websub-cr/

<rhiaro> we have older votes from CR right?

tantek: if someone could reach out to Google implementors that would be good

sandro: also Mozilla could vote... anyone else

<tantek> nothing new on PTD, striking from agenda

sandro: no issues, right aaronpk ?

aaronpk: nothing new

eprodrom: are we looking for implementations or looking for them?

aaronpk: we hit the bar but more is always better

<eprodrom> ben_thatmustbeme: ?

eprodrom: I think tantek removed PTD from the discussion, so next is JF2 if ben is here?

<ajordan> eprodrom: I'm queued

eprodrom: looks like ben_thatmustbeme is not here, so next item is Social Web Protocols

<rhiaro> Nooothing new

<tantek> SWP needs an update with WebSub PR :)

<rhiaro> SWP needs many things.

ajordan: I can say I submitted a lot of issues to JF2 which has been going through implementing some of my patches

JF2

tantek: working draft out there is pretty old, if there are improvements we should turn the crank to publish another update

ajordan: you have to ask ben_thatmustbeme but afaict it's consistent enough to publish a draft, but don't want to speak for ben

tantek: this may be an instance where if folks want to take a look...

<ajordan> https://dissolve.github.io/jf2/

ajordan: looking at the commit history, I don't think there's anything really substantial there; it's mostly just editorial stuff. the one thing is there's a bunch of normative references that were informative before. I guess that's a fix but it was probably obvious

<Loqi> Tantekelik made 1 edit to Socialwg/2017-10-10 https://www.w3.org/wiki/index.php?diff=104644&oldid=104637

tantek: that's worth it, last official publication was from ~ 2 months ago so that's worth updating
... ok AJ I leave it to you to review this document and see if it's good as a WD update
... I'd like to include your updates/patches/etc
... let's let folks take a look at it then we could have a proposal at the next telecon

ajordan: I could also ping Ben

SWP

<rhiaro> Nothing for SWP

<ajordan> !tell ben_thatmustbeme hey, we discussed publishing a new JF2 WD on the telecon with the patches I submitted. AFAICT it's in a publishable state but didn't want to speak for you? what do you think?

<Loqi> Ok, I'll tell them that when I see them next

<rhiaro> Needs updating in general..

<tantek> I thought we resolved to update SWP whenever we had a document status change?

<rhiaro> I don't remember from last meeting, and I scribed

<ajordan> was it bridging?

<ajordan> oops I should write up some text for that

eprodrom: you had some briding ideas I believe AJ?

ajordan: yes

<rhiaro> there's already bridging stuff in SWP

ajordan: this was from a while ago

tantek: not only that but people have put those ideas into practice in the last few weeks. Ryan aka (snarfed) of bridgy fame has put forward bridging between various indieweb protocols and activitypub

<ajordan> tantek: link?

tantek: he's using fed.bridgy to converse on Mastodon

<tantek> fed.brid.gy

ajordan: that's phenomenal

tantek: happy to see evidence of it working, pretty cool
... that may be worth adding the bridging section to SWP, but it is tying our work together

ajordan: we could also reach out to mastodon to futher improve interop... I have to check but I think all they need to do is add a config option to add a link to a webmention (?) and that should in theory interop with all indieweb sites

tantek: wow ok

<rhiaro> Sure

<rhiaro> issues, PRs

<rhiaro> but please read the existing bridging stuff first :)

tantek: should we file issues / PRs?

ajordan: yes I filed an issue about it but then I never followed up

tantek: that's definitely worth a big update

eprodrom: anything else on SWP?

<sandro> +1 the meetings are fun when folks show up

cwebber: SocialCG meeting tomorrow, please show up!

eprodrom: with that I think we can close it up

<tantek> cwebber++ for scribing!

<Loqi> cwebber has 26 karma

<eprodrom> trackbot, end meeting

<ajordan> cwebber++

<Loqi> cwebber has 27 karma

Summary of Action Items

Summary of Resolutions

  1. approve https://www.w3.org/wiki/Socialwg/2017-09-19-minutes as minutes for 19 Sep 2017 meeting
  2. Telecons on 14 Nov 2017 and 28 Nov 2017 unless others are necessary
  3. Resolve https://github.com/w3c/activitypub/issues/239 by removing mediaUpload and specified behavior from ActivityPub spec proper, move to extension via SocialCG process
  4. Resolve https://github.com/w3c/activitypub/issues/260 by clarifying that there is no specific mapping of one user to one actor (and there can be many configurations) in the spec; leave follower migration to SocialCG, and do not add extra modeling of mapping real-world Users to Actors

[End of minutes]