Socialwg/2016-12-06-minutes

06 Dec 2016

Attendees

Present: aaronpk, rhiaro, cwebber, sandro, julien, annbass, ben_thatmustbeme, cwebber2
Regrets Chair: tantek
Scribe: rhiaro

<scribe> scribenick: rhiaro

Last week's minutes

<cwebber2> https://www.w3.org/wiki/Socialwg/2016-11-29-minutes

<cwebber2> +1

<julien> +1

<sandro> PROPOSED: approve last week's minutes https://www.w3.org/wiki/Socialwg/2016-11-29-minutes

<sandro> +1

<rhiaro> +1

<annbass> +1

<aaronpk> +1

RESOLUTION: approve last week's minutes https://www.w3.org/wiki/Socialwg/2016-11-29-minutes

<julien> +1

<cwebber2> +1

AS2

<cwebber2> jasnell, are you around?

<sandro> jasnell, any chance you're around?

<sandro> :-)

<jasnell> yes and no. I cannot join a call but I can answer questions here

tantek: DO we have a CR draft ready for publication

sandro: I think Evan's email said he had made all the changes
... I have not reviewed them yet

tantek: could someone drop a link to the changes?
... so we can at least review it

<jasnell> Evan updated the editor's draft this morning

<sandro> jasnell, have you seen Evan's recent changes?

http://w3c.github.io/activitystreams/core/#changelog

<sandro> http://w3c.github.io/activitystreams/core/index.html#changelog

<Loqi> [James M Snell] Activity Streams 2.0

<jasnell> I very briefly glanced at it this morning

tantek: I can't review, so I'm going to leave it to other folks
... The key thing is to make sure the stuff that we discussed at the f2f and subsequent telecons
... the whole name, summary, should

sandro: Am I right that there's no changes to vocab, just core?
... vocab does not have a changelog

<jasnell> that's how it appears

<sandro> http://w3c.github.io/activitystreams/vocabulary/index.html#changelog shows nothing since CR

<Loqi> [James M Snell] Activity Vocabulary

tantek: if there are other changes, could you take a look at those

<sandro> jasnell, are you okay with this being a new Candidate Recommendation? Do you want more time to review it -- like another day or two to yell if there's a problem?

"The summary property SHOULD be used as a fallback text representation, possibly automatically generated by the publisher. If there is no name property, the summary property SHOULD NOT include markup, and SHOULD be as short enough to be used as a text representation in the language context. "

<jasnell> sandro: let's do that, I've had a ton of context switching the past couple of days so I wouldn't feel comfortable signing off without a better readthrough

I don't think I understand the last part of that sentence about language context?

<jasnell> give me until tomorrow and if I don't raise any objections by then go for it

sandro: We can publish so long as jasnell doesn't raise objections by tomorrow?

tantek: I would be fine with that

PROPOSED: Publish new CR of AS2, giving people another 24 hours to review, pending no objections

<jasnell> +1

<rhiaro> +1

<sandro> +1

<annbass> +1

<cwebber2> +1

<cwebber2> sorry :)

<julien> +1

<sandro> tantek: +1

RESOLUTION: Publish new CR of AS2, giving people another 24 hours to review, pending no objections

tantek: Next AS2 discussion item is a dependency the Annotation group would like to see resolved
... They depend on a bunch of vocabulary terms and want us to go to PR to make sure the terms are stable
... and there's a proposal from sandro

sandro: They were hoping we would go to PR and we talked a bunch in the staff IRC and realised we don't actually need to go to PR for them to be able to reference us, just that we intend to proceed to PR

tantek: you noted that there's only one term we dont' have implementations for
... startIndex

sandro: They have at least one implementation, and may have two others

tantek: If we don't get implementation reports from them for that, we'll drop it, and they should too
... I feel we should have this on the record for a resolution
... so that it's clear

<sandro> PROPOSED: We consider the 12 AS2 terms used by WebAnnotations to be stable and will not substantively change their definitions from those in our 06 September 2016 CR. We expect to go to PR in Jan 2017 and see no likely impediments, given our plan to drop any vocabulary terms lacking 2 impls. We currently lack 2 impls of as:startIndex, but assume Anno can provide them.

<julien> +1

<sandro> +1

<cwebber2> +1

<aaronpk> 👍

+1

<annbass> +1

RESOLUTION: We consider the 12 AS2 terms used by WebAnnotations to be stable and will not substantively change their definitions from those in our 06 September 2016 CR. We expect to go to PR in Jan 2017 and see no likely impediments, given our plan to drop any vocabulary terms lacking 2 impls. We currently lack 2 impls of as:startIndex, but assume Anno can provide them.

<sandro> jasnell, are you cool with that?

<jasnell> yep

<julien> I think? Can't hear hm anymore :/

sandro: While we're waiting for tantek, I think that's it for what's explicitly on the agenda

Document status

sandro: Raise your hand if you have a document status update to report

<aaronpk> 👋

<julien> he waives

<julien> raises his hand I believe ;)

various: discussion of aaronpk's problemantic use of emoji

sandro: giving tantek another minute..
... Did everyone see the list of AS2 implementations?

http://as2.amy.gy/reports.html

scribe: It does make me wonder how many things we're going to have to drop from as2

annbass: it's a great way to present that information

sandro: Shall we go ahead and proceed without tantek?
... aaronpk, go ahead?

tantek: *returns*

<julien> I do no have a status report :/ sorry for the confusion

aaronpk: this is for WebSub, from julien as well
... We've been working on getting it ready for CR

<aaronpk> https://w3c.github.io/websub/

aaronpk: I have the editor's draft up to date with the conformance classes and the last trailing issues that julien's been merging in
... I have one pendin gPR to add the exit criteria to it

<aaronpk> https://github.com/w3c/websub/pull/77

aaronpk: the same stuff that was written in the issue
... that should take us to everything we were planning on doing for taking it to CR

tantek: does that mean you have a CR draft ready?

<julien> done and done!

aaronpk: once that PR is merged it should be

<julien> PR merged and CR ready

tantek: changelog?

<aaronpk> https://w3c.github.io/websub/#changes-from-20-october-fpwd-to-this-version

aaronpk: Oh I didn't update that yet
... I believe the list is very small though

tantek: any outstanding non-editorial issues open?

aaronpk: A couple waiting for commenter
... julien, do you know offhand?

julien: One waiting for commenter feedback, but I don't think there's anything blocking

tantek: if it's still open it's blocking
... If we previously resolved on then that should be clear in the issue. If youthink you have a resolution but the commenter hasn't responded yet then walk us through that issue
... basically by this discussion we should see if we have zero non-editorial issues that need input from the group

<julien> https://github.com/w3c/websub/issues/68

julien: number 68
... The proposal is to replace rel=self with rel=canonical
... I don't think there's any benefit to it, but it would break a lot of implementations

tantek: do we have a proposal to resovle that issue without changes?
... can you propose that both in the issue and IRC

sandro: hold on..
... it's currently tagged as waiting for commenter, but the commenter has since replied, an dnobody has answered

julien: he's asking for the same thing over

sandro: he phrased as a 'new suggestion', so to close without answering him.. how do you reply?

julien: I think we should not ?? the rel=canonical urls

sandro: his point is he wants people to be able to use rel=canonical in their html

julien: I don't think there is any good reason for it

aaronpk: everything is the same, except in html people should look for rel=canonical and THEN rel=self
... since canonical is already existing, people might already have it, and then wouldn't have to add a new one

sandro: what does websub use this for? what happens if you don't have either?

aaronpk: then you don't know what topic url to send

sandro: can't be the URL of the page?

aaronpk: theoretically the same but some cases where it's not
... the spec makes it explicit that it should always use the advertised the value

tantek: one of my requests from a while ago was that if there's no self then just use the URL of the page

julien: one of the problems that we've found over and over with feeds is query strings
... having urls with query strings, people subscribe to that, but that is not what gets sent to the hub, so the subscriber never gets data

sandro: the arguements for requiring rel=self, already implemented, but aside, we want to avoid the error cases where the URL is wrong, and rel=canonical might have the wrong URL as well?

julien: it's different things
... a need for a way to identify the topic, should not be implicit, it will fail silently
... if it's explicit, I don't see the point in having two possible values

aaronpk: is there something about rel=canonical that has a different meaning?
... the example I could think of is an atom feed sometimes points to a rel=canonical of the html page

<sandro> ah... so the rel=canonical would be wrong

aaronpk: so if you want to subscribe to the atom feed, you can't use the rel=canonical that would be wrong

tantek: the confusion here is from the fact that the term self as an English term si very overloaded
... and what it really means is like rel=update. Yes your'e looking at this page, and if you want to get updates of the kind of stuff you see on this page, here's the URL that you subscribe to. That's my understanding of the meaning of rel=self in WebSub
... the unfortunate thing is that people see rel=self and think of the english term
... it does not mean all the things people think when they see self
... and it's unfortunate, but there's so much interop on this I would not propose changing

<sandro> maybe rel='websubChannel" would be the clearest English :-)

tantek: maybe an informative note about what does rel=self mean

<julien> yiiiikes!

tantek: I do think .. wasn't there an atom rel=self alsothat meant something else?

julien: ours is the meaning of atom rel=self that we're using here

sandro: it sounds like if we modify the draft to include a note including what we mean by self

<aaronpk> rel=topic would have been a better choice but that ship has sailed

sandro: and tell the commeter we thought about it and are too concerned about the fact that canonical might be different, and we want to keep it a separate relation

tantek: mention the experience of silent failure

julien: silent failure, if you do not subscribe to a URL that's the one the publisher advertises you will not get data and will not understand why

aaronpk: most obvious example is query string parameters

<sandro> +1 silent failure is a bigger issue than convemience of re-using rel=canonical

tantek: our irc logs, with /today
... can we move this along?

julien: I will add a note to the spec about why it is like this
... as well as why it's important to have a value and not nothing

<sandro> PROPOSED: Add note to spec about the naming of rel='self', and do not change to use rel=canonical at all

tantek: if you're willing to consider it, julien, the example of IRC logs, really helped convince me when aaronpk told me about that of the value of self

<rharo> +1s? ^^^^

<sandro> +1

tantek: vote on proposal

<rhiaro> +1

<annbass> +1

<aaronpk> +1

<julien> +1

<julien> (back)

RESOLUTION: Add note to spec about the naming of rel='self', and do not change to use rel=canonical at all rehttps://github.com/w3c/websub/issues/68

<julien> https://github.com/w3c/websub/issues/13

tantek: Next issue?

sandro: URL vs URI inconsistency?

tantek: I think we had something similar with webmention?
... we expect browser like behaviour with discovery
... I think I commented on the issue

<julien> but then what's the resoltuion tantek?

sandro: in webmention we reference the URL spec

tantek: and that's based on what other specs at w3c are doing, and it better reflects how URLs are treated in html, and that's essentially what we're doing, makes more sense than the IETF rfcs

<julien> Basically change RFC 3986 to https://url.spec.whatwg.org/ ?

sandro: one of the counter arguements would be that it is a political statement

tantek: if anything that's an arguement *for*

<rhiaro> Is this the one that was updated in webmention with a note to go with it?

tantek: we should do what web platform wg is doing

sandro: you mean what html has been doing?

tantek: if there's a question of politics, politial inertia, that's on the side of using the urls pec

sandro: could be either way
... there will be people who will be upset with either choice
... hopefully not terribly

tantek: not clear that it will make any difference. Worse for us to be inconsistent

sandro: consistency is better

tantek: we should record if people are upset, and if they implement, and move on

<sandro> PROPOSED: resolve https://github.com/w3c/websub/issues/13 saying we're going to consistently us the https://url.spec.whatwg.org/

<julien> +!

<julien> +1

<ben_thatmustbeme> +1

<sandro> +0 not sure about the politics, but I guess, and consistency is important, and we went this way with WebMention

<aaronpk> +1

tantek: unless there are necessary technical reasons for referencing one over another, we should have the TAG say something

RESOLUTION: resolve https://github.com/w3c/websub/issues/13 saying we're going to consistently use the https://url.spec.whatwg.org/

<sandro> <julien> +!

<annbass> +1

<rhiaro> Do we need that note that was added to webmention too though?

sandro: I think it's better to ask the commenter to close it

tantek: as long as we have a group consensus recorded, for the purposes of going to CR, it's better to leave it open as sandro suggests

sandro: somebody needs to go back and tag the 39 closed issues
... so we can have an analysis in the CR transition meeting
... most of the ones that don't have tags are by the editors, but still
... We do have to check over the closed issues and make sure they are legitimately closed

tantek: for the open ones, if it doesnt' say editorial then we need a resolution. Any left/

sandro: I see three

tantek: Unless those are mis-tagged I don't think we're going to resolve 3 issues in 2 minutes

sandro: two are raised by aaron

aaronpk: 27 is just waiting on me to write up, already resolved
... I lost track of 47
... I dont' remember why it's important any more
... and there's a new one after last eweks' call, I haven't analysed yet, 73
... 73 is the only one to talk about now

tantek: sounds like we need more time on that one
... I think we have enough edits done that you can publish a new WD with a changelog
... preferably with the edits we all just agreed on as of this call

PROPOSED: Publish a new WD of WebSub including the resolutions from today's call

<rhiaro> +1

<sandro> +1

<aaronpk> +1

<cwebber2> +1

RESOLUTION: Publish a new WD of WebSub including the resolutions from today's call

sandro: do the editors want to try to do a CR vote next week?

<ben_thatmustbeme> +1

julien: yep

aaronpk: yep

<sandro> sandro: should we try to super-rush to get it out before the moratorium

<sandro> sandro: probably not...

tantek: we can publish WDs any time

<sandro> tantek: yeah, probably not

tantek: next week I'm hoping Evan can chair

trackbot, end meeting

Summary of Action Items

Summary of Resolutions

approve last week's minutes https://www.w3.org/wiki/Socialwg/2016-11-29-minutes
Publish new CR of AS2, giving people another 24 hours to review, pending no objections
We consider the 12 AS2 terms used by WebAnnotations to be stable and will not substantively change their definitions from those in our 06 September 2016 CR. We expect to go to PR in Jan 2017 and see no likely impediments, given our plan to drop any vocabulary terms lacking 2 impls. We currently lack 2 impls of as:startIndex, but assume Anno can provide them.
Add note to spec about the naming of rel='self', and do not change to use rel=canonical at all rehttps://github.com/w3c/websub/issues/68
resolve https://github.com/w3c/websub/issues/13 saying we're going to consistently use the https://url.spec.whatwg.org/
Publish a new WD of WebSub including the resolutions from today's call