Social Web Working Group Teleconference

08 Nov 2016

Attendees

Present: annbass, rhiaro, aaronpk, tsyesika, eprodrom, sandro, ben_thatmustbeme, cwebber2, tantek
Regrets: julien, csarven
Chair: Evan
Scribe: sandro

Minutes

<eprodrom> Can someone scribe?

<Loqi> rhiaro: tantek left you a message 6 days, 18 hours ago: do you know how we (chairs / staff) can make blog posts here: https://www.w3.org/blog/ (as other WG chairs (including non-W3C-team people) seem to be able to) ?

<Loqi> sandro: tantek left you a message 6 days, 18 hours ago: do you know how we (chairs / staff) can make blog posts here: https://www.w3.org/blog/ (as other WG chairs (including non-W3C-team people) seem to be able to) ?

<tsyesika> hey rhiaro

<eprodrom> Scribe?

<scribe> scribe: sandro

<eprodrom> PROPOSED: accept https://www.w3.org/wiki/Socialwg/2016-11-01-minutes as minutes for Nov 1 2016 telecon

<rhiaro> +1

<annbass> +1

<eprodrom> +1

+1

<aaronpk> +1

RESOLUTION: accept https://www.w3.org/wiki/Socialwg/2016-11-01-minutes as minutes for Nov 1 2016 telecon

<eprodrom> https://www.w3.org/wiki/Socialwg/2016-11-17

<Loqi> Social Web WG Face to Face Meeting at MIT (F2F8)

evan: Reminder: Face-to-Face meeting next week, in the Boston area

https://www.w3.org/wiki/Socialwg/2016-11-17

<Loqi> Social Web WG Face to Face Meeting at MIT (F2F8)

eprodrom: Media wiki tables are a challenge, but we can do it!

<annbass> very bummed I cannot be there

rhiaro: telecon tuesday or cancel it?

eprodrom: I'll be in transit
... Anyone who thinks we should have a telecon Tuesday?

<eprodrom> PROPOSED: cancel Tuesday Nov 15 2016 telecon

<eprodrom> +1

+1

<rhiaro> +0

<aaronpk> +0

<annbass> +1

<ben_thatmustbeme> +1

RESOLUTION: cancel Tuesday Nov 15 2016 telecon

eprodrom: dinner?

<rhiaro> I will lead everyone to Veggie Galaxy unless overridden by someone else

<aaronpk> rhiaro++

eprodrom: it'd be good, but let's figure it out later

<Loqi> rhiaro has 136 karma in this channel (247 overall)

VeggieGalaxy++

<Loqi> veggiegalaxy has 1 karma

AS2 issue-312

eprodrom: Must every AS2 object have a name? Complicated

<cwebber2> hiya

eprodrom: I think re resolved to have a fallback name property

<cwebber2> dialing in

w/ re / we /

eprodrom: right now we have a SHOULD requirement for .name of every Activty, Collection, Image, Person, etc

<ben_thatmustbeme> https://github.com/w3c/activitystreams/issues/312

eprodrom: question was raised -- for some objects, like a Collection, the name fields forced
... Sometimes it doesn't seem like there should be a name
... a Page of a Collection
... the counter argument: there will be many consumers that will want to present activities that they might not understand, or be able to render in any specific way, so it'd be nice to have something like name to fallback on
... esp with extension types, or obscure types
... Problem with doing that: some consumers may want to have customized presentation, and they wouldn't be able to tell which names were really specific to that object (eg my name) vs a generated name like (eg Unnamed Person)
... so the resolution is to have a fallback name required, and an option name would be allowed.
... fbname might be "a person", name "evan prodromou"
... Why is this on the agenda?

rhiaro: I put it on the agenda, because I didn't think we'd decided what the names were going to be.
... If we're making a normative change, we need to extend CR
... and I'm implementing and need to know what to do

<ben_thatmustbeme> would we need a fallbackNameMap?

eprodrom: Yeah, I don't think we'd picked the final properties

rhiaro: biggest thing is if we're going to make a normative change -- which seems unavoidable -- if we make .name the required property, or if a we add another required property that's also normative ...

eprodrom: If we use name as the mostly-required (SHOULD) and add an optional "title"
... that minimizes the normative impact

rhiaro: It'd need to be MUST to be a reliable fallback

<cwebber2> :|

<cwebber2> no MUST

<tantek> BTW such a "fallbackname" is kind of a brand new feature and having never been incubated is not really appropriate to add during CR

<cwebber2> tantek: it's not a new feature

eprodrom: I see that point, but we've had it as a should for so long, and it's serving that same role

<cwebber2> tantek, this used to be two separate fields

<cwebber2> tantek, and then they were combined

rhiaro: Okay, so we're relying on developers taking the SHOULD very seriously

<cwebber2> tantek, and now we've realized that's a mistake

eprodrom: Well, that's how it's been used

<tantek> cwebber2, ok, in that case ok with "at risk"

<cwebber2> tantek, granted, "fallbackName" is a new *term*

<cwebber2> since it used to be displayName iirc

eprodrom: Making it a MUST would be hardening that requirement, which seems unneeded

<Zakim> tantek, you wanted to note that IIRC at the f2f we decided to drop the SHOULD for name, which should not affect any conformant implementations. the fallbackname did not have

tantek: What I recall from F2F, with one objection, was the drop the SHOULD from .name
... having a fallback name, people were open to that, editor was supposed to provide proposal
... regarding fallback name, Chris and I spoke, and I'd be okay adding something at-risk

cwebber2: So, I'm looking at the notes

<cwebber2> https://www.w3.org/wiki/Socialwg/2016-09-23-minutes#AS2

<rhiaro> That way of thinking about it makes it sound almost like it's a non-normative change, that's what I was looking for :)

RESOLUTION: Return distinction between "user entered or otherwise significant name" and "text fallback" and shift SHOULD from meaningful name to text fallback.

<tantek> yay citations

<Loqi> :D

cwebber2: I think that was a good resolution
... so what to call them?

<tantek> ugh, the whole reason we renamed to "name" was because others using "meaningful name" were using that

cwebber2: main motive is avoid a new CR --- so make name be the fallback

<rhiaro> I agree with everything cwebber2 is saying

cwebber2: But I like the name as the meaningful name

sandro: if it's an extension, then it doesn't need a new CR

cwebber2: I think this is core

<rhiaro> It's only like 4 weeks, right

tantek: I'd like another CR to do this

<rhiaro> It's not so bad :p

<rhiaro> AS2 is for life, not just for the charter period.

eprodrom: Yeah. My feeling here is ... the spec is forever, so let's try to do the right thing for the long term.

<annbass> +1 on 'doing right thing'!

eprodrom: let's not have our short-term scheduling issue be the problem.
... so let's have .name be the meaningful name, and something else .depiction? as the fallback

cwebber2: how about .displayName ?
... or .fallbackName

<tantek> I can live with fallbackname

+1 fallbackName

<tantek> however I have a counterproposal for avoiding a 2nd CR (I think)

eprodrom: I like fallbackName
... sounds a bit like tempInteger but I see the point

<tantek> (that I made at the f2f)

<Zakim> tantek, you wanted to note that I did propose a way to potentially avoid a second CR at the f2f

tantek: another idea, without CR
... drop SHOULD from name

<ben_thatmustbeme> sandro, i was going to cover that too

tantek: say consumers that need some kind of fallback name
... should use the .summary property
... and say publishers, if you don't provide a name, you SHOULD provide a plaintext .summary
... In my experience, the situations where you don't provide a name
... and consumers still want a something, my exoperience and indiewebcamp experience, is summary works well for that
... eg a LIKE doesn't have a meaningful name, but it does have a summary which can be used in contexts that don't understand Like
... that works well
... it works in clients that don't understand this kind of reaction, but understand reactions in general

<tantek> <tantek> +1 with use as "summary" for fallback text since that seems to work, and does not require adding a new term (thus does not require a new CR)

tantek: I'm advocating for this, that this doesn't need a new CR
... would this break any existing implementations?
... I'm trying to find a way to keep the funcationlity people want, in a way that folks have seen working

eprodrom: sounds like a solid proposal, I like it
... only objection I might have is
... the name/title slot typically has max six words, but in some cases, the summary might be several sentences long. An Abstract or TL;DR type thing.
... that's my only concern here

tantek: I agree that's a real case

<rhiaro> I can hear tantek and eprodrom just fine

<eprodrom> Did I just fall off the call, or did tantek?

<eprodrom> I think it was me

<eprodrom> I'm going to call back in either way, since I can't hear.

tantek: but in all the cases where there's a longer article, there's a name provided in practice.
... I think

<eprodrom> tantek: would you mind chairing for 5 minutes while I reconnect?

tantek: When summary is longer, there tends to be a meaningful name
... Also, it's possible for people to provide really long names anyway

<rhiaro> If summary is long, the consumer knows they're allowed to truncate it, since it's sthe fallback

tantek: Maybe provide guidances, saying .name and .summary MIGHT be long

<tantek> chair: tantek

<Zakim> ben_thatmustbeme, you wanted to ask about SHOULD / MUST of fallbackName

ben_thatmustbeme: in either case, whether we use fallbackName or .summary, is it a MUST or a SHOULD?

<cwebber2> -1 on MUST

tantek: Yes, if there's no name provided, summary becomes a SHOULD. SHOULD is strong enough, doesn't need to be MUST, because we might get empty summary values.

<cwebber2> it's just fallback text, if someone doesn't have it, it's hot a *huge* deal

<eprodrom> back

tantek: this keeps a smaller change.

<eprodrom> name -> summary -> "an object"

ben_thatmustbeme: If .name isn't present, the publisher SHOULD provide a .summary

tantek: Right, but we should also give guidance in case both are missing.
... that's a possible thing in the real world.

sandro: The test suite should have an entry missing both\

<tantek> eprod I would be ok with that

eprodrom: it does now

<tantek> eprodrom, I would be ok with something like that up to the consumer (localization etc.)

<Zakim> cwebber, you wanted to say it would be fine, but explain I think it'll still result in a new CR though

<tantek> chair: eprodrom

cwebber2: I'm fine with .summary, and I'm against MUST, as we discussed in F2F

<tantek> +1 cwebber2

cwebber2: I want to raise a Devil's Advocate argument
... if we move the SHOULD, that still might need a new CR
... if BigBlueHat were here, he'd argue that SHOULD's aren't normative

tantek: It's not that SHOULD's arent normative, it's about new features, and breaking implementations.
... so if implementations do something different, then you'd need a normative change in response

<rhiaro> It means that most of my activities don't have summary..

tantek: case by case

rhiaro -- or a name?

<rhiaro> sandro: aye

tantek: If we think this is compatible with implemnentations, it should be okay

cwebber2: Let's check with James

<rhiaro> ..maybe it's fine

eprodrom: No, we don't need to wait for James. He's aware of our schedule.

<tantek> jasnell can raise an issue based on this conclusion if he wants and we can consider that if/when that happens

<tantek> +1 eprodrom

eprodrom: This is in the spirit of what we agreed at the F2F

<cwebber2> +1 sounds good

<Zakim> rhiaro, you wanted to ask if we'd have to kick html out of summary for this to work?

<eprodrom> !!

rhiaro: This all sounds good. I'm fine using .summary this way. Great to have it not considered a normative change.
... what about Markup, though

sandro: name does not have markup, but summary does ?

eprodrom: leave it as is

sandro: advice to implementors?

<eprodrom> PROPOSED: name -> MAY, summary -> SHOULD, add section on string representation of object

<tantek> not sure if that made it through

<tantek> summary is a SHOULD *only* in the absence of name

<tantek> don't want a general SHOULD on summary

<eprodrom> PROPOSED: name -> MAY, summary -> SHOULD if no name, add section on string representation of object

<rhiaro> Also there was that thing about keeping it required for Article ..?

<tantek> and informative consumer guidance on what to do with too long of a name/summary or if both are empty

sandro: yes, but that doesn't address markup

<tantek> summary -> SHOULD be there and be plain text if no name

sandro: would be say something like: when you're using summary as a backup name, it's okay to strip out the markup.

eprodrom: Yes, non-normative advice.

<tantek> yes

sandro: I'm concerned stripping markup might change semantics

eprodrom: okay to leave the markup, too

<tantek> oops disconnected, reconnecting

<cwebber2> oof

eprodrom: I'd note that summary has markup

<cwebber2> welllll

<tantek> sandro no

sandro: I think we said name needs to work without markup, so fallback name does too

<tantek> implementations of Atom etc. already drop markup from 'name' or 'summary' when displaying

<cwebber2> that doesn't mean that summary renders in the same way as name...

<tantek> so no this is not an issue in practice

<cwebber2> taking markup out of summary seems bad

<rhiaro> I think it's fine saying if publisher is using summary as a fallback, SHOULD NOT have markup

sandro: Tantek, you're saying it's well know, even if not in spec, that markup might be stripped from sumary

<cwebber2> -1 on taking markup out of summary, +0 on permitting stripping markup from summary (though depending on markup, that might not always be easy)

tantek --- no sound

<tantek> implementations already do it (consuming code removes markup from name / summary in Atom)

<tantek> sandro I am still disconnected, attempting dialing

<wilkie> the biggest dependency I've had in implementations is libxml which is a troublesome dependency to support on many machines. I only need it to strip HTML out, which I do 100% of the time.

<tantek> right

eprodrom: consumer can do anything, so no point in saying they can strip markup
... maybe advice against markup in summary when used as fallback

<rhiaro> That sounds good to me

-1 eprodrom consumers can do whatever they want

<cwebber2> I want waffles

<eprodrom> PROPOSED: name -> MAY, plaintext summary -> SHOULD if no name, add section on string representation of object

tantek: Also informative guidance on both being too big or empty

<wilkie> HTML in fields are good if I can just ignore them completely if I'm likely going to strip the HTML and can't rely on preserving the intended representation

<eprodrom> PROPOSED: name -> MAY, summary -> SHOULD if no name, add section on string representation of object including informative guidance on name or summary being very long

tantek: because we have thoguht about

<ben_thatmustbeme> PROPOSED: change name to may, if no name, SHOULD provide a plaintext summary, add a sectino on string representation, add guidance to the fact that summary or name may be too long

<ben_thatmustbeme> hows that

<ben_thatmustbeme> ?

<ben_thatmustbeme> GAH

<aaronpk> haha I type "functino" all the time

<Loqi> aaronpk: lol

<eprodrom> PROPOSED: change name to may, if no name, SHOULD provide a plaintext summary, add a section on string representation, add guidance to the fact that summary or name may be too long

<cwebber2> +1

+1

<eprodrom> +1

(okay with "very long)

<rhiaro> +1

<wilkie> +1

<aaronpk> +1

<ben_thatmustbeme> very long is fine too

<tsyesika> +1

<ben_thatmustbeme> didn't tantek want "or both may be empty"

<tantek> +1 ok with s/too/very, and ok adding guidance on if both name & summary are empty

<ben_thatmustbeme> +1

<rhiaro> This is just the text of the proposal, not what will go in the spec, it's okay ;P

RESOLUTION: change name to may, if no name, SHOULD provide a plaintext summary, add a section on string representation, add guidance to the fact that summary or name may be too long

<annbass> +1

AS2 Validator

eprodrom: plan to have new version for F2F
... patches welcome

<rhiaro> Thanks!

pubsub.rocks

aaronpk: pubsub.rocks now has test tool for publishers and subscribers

aaronpk++

<Loqi> aaronpk has 67 karma in this channel (1137 overall)

aaronpk: next step is if you're building a hub
... that'll build implementation report

<ben_thatmustbeme> just to note, there are a bunch of new implementation reports for webmention https://webmention.net/implementation-reports/summary/

<tantek> pubsubrocks++ !

<Loqi> pubsubrocks has 1 karma

<ben_thatmustbeme> aaronpk++

<Loqi> aaronpk has 68 karma in this channel (1138 overall)

aaronpk: likely hub test tool will be done next week, but maybe not submitting results

eprodrom, Any other business?

<ben_thatmustbeme> just to note, there are a bunch of new implementation reports for webmention https://webmention.net/implementation-reports/summary/

<Zakim> rhiaro, you wanted to note pubsub name change on wiki

<aaronpk> https://www.w3.org/wiki/Socialwg/push-name

<tantek> any leaders?

<tantek> going to be a close call, vote for your pubsub bikeshedding today!!!

<annbass> thanks Sandro and Evan!

<rhiaro> o/

<eprodrom> trackbot, stop meeting

<trackbot> Sorry, eprodrom, I don't understand 'trackbot, stop meeting'. Please refer to <http://www.w3.org/2005/06/tracker/irc> for help.

eprodrom: Thanks everyone, see everyone at F2F Thursday

<eprodrom> trackbot, end meeting