Socialwg/2017-08-29-minutes

From W3C Wiki

Social Web Working Group Teleconference: 29 Aug 2017

See also: IRC log

Attendees

Present
sandro, tantek, rhiaro, ajordan, cwebber, tsyesika, eprodrom (partial)
Regrets
aaron, julien
Chair
tantek
Scribe
sandro

Contents



+sandro

<scribe> scribe: sandro

<tantek> https://www.w3.org/wiki/Socialwg/2017-08-15-minutes

<cwebber2> +1

+1

<cwebber2> PROPOSED: accept https://www.w3.org/wiki/Socialwg/2017-08-15-minutes as last week's minutes

<cwebber2> +1

<rhiaro> +1

<cwebber2> last week as in the last week we had a call ;)

<tsyesika> +1

RESOLUTION: accept https://www.w3.org/wiki/Socialwg/2017-08-15-minutes as last week's minutes

<Loqi> Cwebber2 made 2 edits to Socialwg/2017-08-29 https://www.w3.org/wiki/index.php?diff=104106&oldid=104101

<Loqi> Sandro made 1 edit to Socialwg/2017-08-29 https://www.w3.org/wiki/index.php?diff=104105&oldid=104104

<ajordan> +1

September meetings

tantek: question 1 - every week or every other week. question 2 - 60 or 90 minute meetings

<rhiaro> I can't make next week, but don't let me stop anyone else

ajordan: It seems to me we've getting more done in 90 minute meetings. I'd rather have more meetings now and relax later, then be squeezing things in before the deadline

aj: so I suppose I'm suggesting weekly 90 mins

<ajordan> I mean even this week the agenda is super long

sandro: Sounds okay to me, but I'm not paged in on agenda

cwebber2: I have mixed feelings. I think biweekly 90 minute is pretty good. I wouldn't want biweekly 60.

tantek: the two weeks blocks seemed productive, that was my impression
... Did anyone feel like we would have gotten more done if we met every week?
... From chairs perspective, every two weeks felt more focused, less tedious

<ajordan> I know I suggested weekly but I don't feel strongly against biweekly

<ajordan> +1 timeboxing

cwebber2: Lets do 90 minute bi-weekly, but if run low on time, we can fill in the intermediate meeting

<tsyesika> +1 on chris' suggestion :)

tantek: Didn't hear anything from Evan about this.
... But let's meet next week for websub

<cwebber2> that's good with me

tantek: so biweekly starting the 6tg

PROPOSED: Biweekly up-to-90-minute meetings, starting Sept 6, with the option of filling in the missing weeks if we have lots to talk about

<cwebber2> +1

+1

<ajordan> +1

<rhiaro> +1

RESOLUTION: Biweekly up-to-90-minute meetings, starting Sept 6, with the option of filling in the missing weeks if we have lots to talk about

Inviting more implementors

rhiaro: It kind of looks better to have impls from outside the group

cwebber2: I don't actually know if it would speed the process

tantek: I worry a little that these meeting details might dampen their energy

AS2

cwebber2: I wanted to make sure the spec for adding extensions ...... assuming the CG voted to add something ....
... Do we talk to Amy to get it committed?

<rhiaro> The process should not be tied to a named person!

<ajordan> second question: what do CGs even have authority to publish? Notes? nothing REC-track I'm assuming

<rhiaro> ajordan: correct

<ajordan> rhiaro: thanks

sandro: two obvious options: w3c staff or github repo
... slight concern if/when Amy and I leave the staff

tantek: governance for repo

<rhiaro> I was gonna vaguely handwave what sandro is saying

<ajordan> if we go with the GitHub repo it would make sense to me to set Evan/Jason as the gatekeepers? since they're the editors actually listed on the spec

sandro: third option, maybe webreq can do it -- I think that's policy

<scribe> ACTION: sandro find out if webreq can handle namespace document, or if there's a better solution [recorded in http://www.w3.org/2017/08/29-social-minutes.html#action01]

<trackbot> Created ACTION-91 - Find out if webreq can handle namespace document, or if there's a better solution [on Sandro Hawke - due 2017-09-05].

sandro: in any of the three cases, I see the chairs as formally recognizing consensus and informing github or staff

ajordan: so it almost doesnt matter which mechanism we use

sandro: right

AS2 context for sensitive tag

cwebber2: Mastodon is rolling out AS2 with sensitive tag
... they could use their own term, but it's going to rolled out, so should it be in AS2?
... we'll be talking about this in CG tomorrow
... so I guess this is a Head Up
... more of a CG discussion
... also about Hashtag object

tantek: If you're looking for input, the CG is responsible for this kind of decision, so you the CG needs to come up with how to evaluate extensions, maybe, instead of making lots of one-off decisions
... if you decide on some criteria, I dont think the CG is bound by WG requirements, but COULD consider things like 2-implementations
... I realize that's kind of chicken-and-egg
... I leave that as something for CG to resolve
... that's my input

<ajordan> I would suggest discussing this directly on the CG call, even if the chairs make the final decision about criteria

cwebber2: Maybe I should reach out to jasnell directly re hashtag

tantek: w3c process to seek out dissenting opinion and try to incorporate it

<cwebber2> https://github.com/w3c/activitypub/issues/254

<Loqi> [cwebber] #254 Should "endpoints" be dropped?

ActivityPub clarifying scope of endpoints (#254)

cwebber2: question was does anyone like this design
... lots of endpoints
... having raised that, some people like tsyesika explained there's a good reason to keep it
... it's very messy to have all these properties not really about actor be on the actor
... there are two ways to resolve messyness
... (1) better documentation
... (2) (from puck) move sharedInbox out of there, make it client-to-server only
... so with (1) endpoints are shared across on whole domain, or with (2) it's just in c2s
... so now I'm with (1), just clarify things
... This would affect implementations, and be a new-CR change, fwiw

<cwebber2> PROPOSAL: rephrase description of endpoints to clarify that its scope is for endpoints that tend to be shared on a domain/server

cwebber2: this would not be a normative change

<cwebber2> +1

<ajordan> +0

+0 sounds reasonable, but don't really know issues

<ajordan> ^^^ same

<tsyesika> +1

tantek: me too, mild support, anyone else?

RESOLUTION: rephrase description of endpoints to clarify that its scope is for endpoints that tend to be shared on a domain/server

<saranix> I thought it was just other collections that the actor had?

https://github.com/w3c/activitypub/issues/242

<Loqi> [cwebber] #242 sharedInbox / siteInbox type endpoint (publicInbox, but not just for public posts)

Accept/Reject Follow (#244)

<tantek> Hi saranix, we are in the middle of a Social Web WG telcon - can you wait til after or CG call tomorrow?

ActivityPub -- Accept/Reject Follow (#244)

cwebber2: We now have implementation support, which we'd made this contingent on
... so I'm just informing the group
... Mastodon is rolling this out in next release

sharedInbox (#242)

ActivityPub sharedInbox #242

cwebber2: oh, I got these backwards. sharedInbox was the one we made contingent on implementations!

<cwebber2> https://github.com/w3c/activitypub/issues/242

<Loqi> [cwebber] #242 sharedInbox / siteInbox type endpoint (publicInbox, but not just for public posts)

cwebber2: messages to followers on big instances; being rolled out.

Accept/Reject Follow (#244)

Back to Accept/Reject follow

cwebber2: Mastodon is rolling this out, too
... but waiting for our new CR

<cwebber2> https://github.com/w3c/activitypub/commit/08cbb048bf35df52567431d39e993f3d4a7c60ac

<cwebber2> https://w3c.github.io/activitypub/#follow-activity-inbox

<cwebber2> > Servers MAY choose to not explicitly send a Reject in response to a Follow if there is a good privacy reason not to do so, though implementors ought to be aware that the server sending the request may be left in an intermediate state.

tantek: any thoughts on this change....?

cwebber2: If there's a good privacy reason you don't want to let people know, ... we left that option

tantek: good case for "for example"

cwebber2: the may is an escape hatch, but then the SHOULD would generate an accept...

tantek: no no, keep the strong wording, just in the may-choose-to-not ....

<cwebber2> https://github.com/w3c/activitypub/commit/08cbb048bf35df52567431d39e993f3d4a7c60ac

<cwebber2> https://w3c.github.io/activitypub/#follow-activity-inbox

<eprodrom> sandro, thanks

<eprodrom> I've been on the call for a few minute

<eprodrom> s

<cwebber2> so maybe: Servers MAY choose to not explicitly send a Reject in response to a Follow, though implementors ought to be aware that the server sending the request may be left in an intermediate state. For example, a server might not send a Reject to protect a user's privacy.

<eprodrom> No

<ajordan> OK

<rhiaro> +1 what tantek is saying and chris's text

<cwebber2> PROPOSED: Accept Accept/Reject on Follow language from Editor's Draft with adjustments to example discussed on this call to resolve #244.

<cwebber2> +1

<tantek> +1

<rhiaro> +1

+1

<eprodrom> +1

<tsyesika> +1

RESOLUTION: Accept Accept/Reject on Follow language from Editor's Draft with adjustments to example discussed on this call to resolve #244.

<ajordan> +1

<eprodrom> Great

cwebber2: Evan, okay if I ping jasnell directly?

Issue normative CR on AP ?

cwebber2: Mastodon is holding up release, waiting on this.
... I'll do changelog now, while meeting talks about other things.

WebSub

sandro: regrets from Aaron and Julien

tantek: They resolved 119 amongst themselves

<tantek> https://github.com/w3c/websub/issues/119

<Loqi> [marten-de-vries] #119 'the hub terminates the subscription'

sandro: it looks to me like a normative change, so I'm looking for some argument that I'm wrong about that

tantek: can't resolve without more info -- hopefully next week

Post Type Discovery

tantek: nothing new

<rhiaro> Sorry, nothing new

nothing new on JF2 or SWP

tantek: AOB? (other than AP CR)

ajordan: I've been adding AS2 to pump.io and it's coming along nicelyt

rhiaro: Does it make sense to schedule anything for TPAC

<cwebber2> o/

<tantek> who is going to be at TPAC

<cwebber2> o/ <- will be at tpac

<tantek> o/

<ajordan> I'm not sure

not me

<ajordan> if I was I would only be there for us

<tsyesika> not me either

+1 schedule a CG meetup

if it's not too late etc

<cwebber2> whew, changelog: https://w3c.github.io/activitypub/#changes-22-august-to-present

-)

<cwebber2> staying in school is how you stay cool

<eprodrom> Fine with me

tantek: Evan, we made a plan to have 90 minute meetings tomorrow and every other week after that, unless there's need for me

<ajordan> cwebber2: exactly

<eprodrom> 19th

<rhiaro> s/6th/5th

tantek: I'll chair Sep 5 and Evan will chair Sep 19

<cwebber2> https://w3c.github.io/activitypub/#changes-22-august-to-present

<tantek> I see two here https://github.com/w3c/activitypub/tree/gh-pages/implementation-reports so this is just a link change?

<tantek> cwebber2, https://activitypub.rocks/implementation-report/ should have items for the impl reports in https://github.com/w3c/activitypub/tree/gh-pages/implementation-reports

sandro: Can we get the implementation report before publication?
... Can we clarify Owen's role, maybe as "Original Author"

tantek: So this was based on ActivityPump which was written by Owen. Maybe say that?
... Looking for cases where one spec came from another

<tantek> https://www.w3.org/TR/webmention/

<Loqi> [Aaron Parecki] Webmention

tantek: Do like the Author's Note in wb, explaining this is based on apump from owen, etc

<ajordan> someone run a git blame

+1

<cwebber2> PROPOSED: Issue new ActivityPub CR with changes as noted in changelog at https://w3c.github.io/activitypub/#changes-22-august-to-present

<rhiaro> +1

<cwebber2> +1

+1

<ajordan> +1

<tsyesika> +1

RESOLUTION: Issue new ActivityPub CR with changes as noted in changelog at https://w3c.github.io/activitypub/#changes-22-august-to-present

<ajordan> whoohooo!

sandro: so publicInbox will remain as deprecated, with an explanation, so people know what it means if they see that in some data or code somewhere.

<eprodrom> +1

tantek: AOB?

ADJOURNED until next week

<eprodrom> Thanks Tantek!

tantek: Thanks everyone

<eprodrom> tantek++

<Loqi> tantek has 69 karma in this channel (381 overall)

<cwebber2> sandro++

<Loqi> sandro has 49 karma in this channel (56 overall)

<cwebber2> tantek++

<Loqi> tantek has 70 karma in this channel (382 overall)

<tantek> sandro++ for scribing!

<Loqi> slow down!

<tantek> trackbot, end meeting

Summary of Action Items

[NEW] ACTION: sandro find out if webreq can handle namespace document, or if there's a better solution [recorded in http://www.w3.org/2017/08/29-social-minutes.html#action01]

Summary of Resolutions

  1. accept https://www.w3.org/wiki/Socialwg/2017-08-15-minutes as last week's minutes
  2. Biweekly up-to-90-minute meetings, starting Sept 6, with the option of filling in the missing weeks if we have lots to talk about
  3. rephrase description of endpoints to clarify that its scope is for endpoints that tend to be shared on a domain/server
  4. Accept Accept/Reject on Follow language from Editor's Draft with adjustments to example discussed on this call to resolve #244.
  5. Issue new ActivityPub CR with changes as noted in changelog at https://w3c.github.io/activitypub/#changes-22-august-to-present

[End of minutes]