Social Web Working Group Teleconference -- 18 Nov 2016

<Loqi> yeah who invited you anyway Zakim

<csarven> ---

<csarven> I'd like to join a hangout (audio only is fine) live from my sofa.

<aaronpk> we'll probably use the conference number again

<csarven> is that a diff number than the weekly w3c telco number?

<aaronpk> same

<csarven> ok, thanks. Let me know when I can join.

<aaronpk> will do. i think we said we're doing agenda scheduling at 9:30 (in 37 minutes)

<csarven> I'll join in then.. with preference to take up LDN first :)

<rhiaro> In searching for instructions on how to operate this conference room, I found http://archive.boston.com/news/education/higher/articles/2007/11/06/mit_sues_gehry_citing_leaks_in_300m_complex/

<rhiaro> Woah! ben_thatmustbeme found the secret to operating the shades and projector screen, after we spent 10 minutes pushing every button in the room

<rhiaro> hey eprodrom tantek you need to show up, we're talking about politics

<ben_thatmustbeme> hello

<rhiaro> trackbot, please draft minutes

<trackbot> Sorry, rhiaro, I don't understand 'trackbot, please draft minutes'. Please refer to <http://www.w3.org/2005/06/tracker/irc> for help.

<rhiaro> trackbot, make minutes

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

<ben_thatmustbeme> trackbot, draft minutes

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

<aaronpk> trackbot, figure it out

<trackbot> Sorry, aaronpk, I don't understand 'trackbot, figure it out'. Please refer to <http://www.w3.org/2005/06/tracker/irc> for help.

<aaronpk> csarven, we're on the conference bridge now

<csarven> Are all hands on deck?

<aaronpk> almost

<tantek> csarven, all hands we need for LDN :)

<tantek> trackbot, start meeting

<csarven> digging diggg.. where is the number..

<aaronpk> csarven, https://www.w3.org/2015/10/social-wg-telecon

<aaronpk> scribenick: aaronpk

LDN

tantek: open issue for LDN?

csarven: we received one new editorial issue after we entered CR

<tantek> https://github.com/w3c/ldn/issues

<csarven> https://github.com/w3c/ldn/issues/60

rhiaro: there's one more open issue. we've updated the text already, but we're waiting for commenter. i don't think we need to do anything more on it though.

<csarven> https://github.com/w3c/ldn/issues/55 (open world assumption..)

rhiaro: and there's an ongoing philosophical discussion that doesn't affect the spec

csarven: it's quite interesting if you have nothing else to do

rhiaro: he's nitpicking about phrasing

tantek: is it normative?

rhiaro: no

tantek: amy and sarven, can you two agree on the text changes needed to resolve this and propose it to the group?

csarven: we've made a few commits to revise what henry was concerned about.
... it's about the wording about the idea that don't take the claims in the notifications for granted

tantek: so you think you've completed commits that handle the issue as reported?

eprodrom: this seems like a security consideration. if i'm tracking the state of an object, i may not want to accept notifications from an entity that doesn't control that object. i may not want to allow updates that take an object into an unexpected state that does not reflect the previous history of the object.

rhiaro: this is just in the intro to the security considerations section. we've already covered many ways for this, this is just the intro

eprodrom: if there are specific security considerations already covered, it makes no sense to argue this point and maybe take it out

tantek: can you drop a link to the proposed wording that hasn't been committed yet

<rhiaro> https://github.com/w3c/ldn/issues/55#issuecomment-257135287

csarven: we also have a sender verification section. let's say you have an access controlled way of writing notifications to an inbox. that alone doesn't address the validity of the statements.

<tantek> PROPOSED: resolve LDN issue 55 with rhiaro's suggested text in https://github.com/w3c/ldn/issues/55#issuecomment-257135287

<rhiaro> +1

<aaronpk> +1

<ben_thatmustbeme> +1

<rhiaro> csarven?

<csarven> +1

eprodrom: "taking precautions" seems open

rhiaro: that's expanded upon in the actual section

<rhiaro> Just to have on record: I don't think henry's proposed alternative is appropriate because it ties more closely to LDP that necessary, and we don't need to talk about 'reputation'

RESOLUTION: resolve LDN issue 55 with rhiaro's suggested text in https://github.com/w3c/ldn/issues/55#issuecomment-257135287

<eprodrom> +1

tantek: great, everything else is editorial, i'll let you all handle those until you say you want input from the group

<csarven> https://github.com/csarven/ldn-tests

tantek: next question, how is the test suite coming along?

<rhiaro> https://linkedresearch.org/ldn/tests/

rhiaro: we've got tests for receivers, sender and consumer tests are not done yet

<csarven> https://linkedresearch.org/ldn/tests/receiver

rhiaro: the tests need some UI love to tell when you've actually passed

tantek: do you have a rough idea of when they will be complete?

csarven: i think another week or so to polish it up
... sender and receiver should be far simpler
... we have a way of checking whether something's working or not. what remains for the receiver test is... there's a "report" section
... you can enter a URL with an inbox and do a GET/POST/HEAD/OPTIONS and it will return the headers it receives and the body
... there's a section on a report that shows the normative parts of the spec and a check or cross
... that's the part that needs to be done

<cwebber2> present

tantek: next question is how are you coming with implementation reports?

rhiaro: we have an informal list, those are the people we'll go back to after the test suite is done

tantek: based on the informal list, what's your confidence level of the amount of the spec that's implemented

rhiaro: we have implementations of all roles that do everything

<csarven> https://github.com/w3c/ldn/tree/master/implementations

rhiaro: i think we have at least two implementations but not sure how many of those are non-editor implementations
... do we need implementation reports from generic LDP implemtations, since LDN is a subset of LDP?

tantek: i think it would still be useful to get reports from those, to test your assumption that LDN is actually working the way you expect with generic LDP

sandro: keep it clear that these implementations are not necessary for the process.

tantek: i think the value is getting feedback from the implementers

sandro: from a community perspective it's great, but from a process perspective it's gravy

tantek: looking at the LDP rec it's from feb last year, so i'm guessing most of the tests in the report were run over a year ago. so having the test results for LDP will show that these implementations are still around.
... so it's 90% frosting but making sure these implementations still exist is good to know
... based on that, we should try to transition to PR in january
... we'll try to time this with Micropub in january

csarven: one question about tests. the way we're going at it now, it's a form and you select some options and submit and it comes back with results.
... the question is the difference between tests and validators
... what we have now is kind of like the w3c html validator
... kind of like how webmention has it, individual links to each test
... we're not doing it that way, is the way we're doing it okay?

tantek: do you have instructions for someone running the test to say run it with these options and then these other options

rhiaro: we could have URLs that prefill the forms with different options, or just have a set of different forms

tantek: what does the implementation report template have?

rhiaro: the report template right now has a bunch of checkboxes. we can refine that to different combinations of things.

tantek: that would probably help people submitting reports

eprodrom: of the implementations, what i don't see are any activitypub implementations which was one of the advantages of LDN we were hoping to have

rhiaro: my implementation is also activitypub

eprodrom: it may be important to raise this as we're talking about getting implementations of activitypub

rhiaro: a conformant sender would be sending activitypub if it found an activitystreams inbox

<cwebber2> I'm here

tantek: i think we know what to do next for LDN then. when you feel like the test suite is complete, put it on the agenda for the next telcon
... we'll use that as a trigger to start requesting implementation reports

10 minute break

<csarven> (not about the break)

<tantek> break for 10 min

8 minutes until resume meeting

<Loqi> I added a countdown scheduled for 2016-11-18 4:00pm GMT+0000 (#5938)

<tantek> oops was scrolled and didn't see aaronpk had done it already

<Loqi> resume meeting

<cwebber2> I emailed Jason Robinson about implementations

<cwebber2> he indicated on this thread that Diaspora was unlikely to implement AP but Friendica may: https://joindiaspora.com/posts/8251228

AS2

tantek: are there any open issues since last time we talked?

eprodrom: we call out the JSONLD context as non normative, but we would still like to make it useful
... amy uses it to generate the document that lives at the namespace URL

<ben_thatmustbeme> cwebber2, i seem to get that from diaspora on most things, seems like they creaare mostly of the opinion "we aren't going to add it, but you are welcome to add a PR"

rhiaro: having extra things in it will probably not break things but having missing things might

<csarven> Note to self: look at AS2 examples (the one that extends with OA, PROV that I wrote last year) and see if they are still okay (especially with the changes in OA)

eprodrom: we have two outstanding editorial issues, actually no we dont

<tantek> https://github.com/w3c/activitystreams/issues

eprodrom: we have one outstanding editorial...
... about the relationship between link types and object types
... there is a subtle difference between referencing an object and referencing a link to the object
... james gave a good answer to it and asked him to include it in the document itself
... it hasn't been updated so i think i'm going to do it myself
... i think i understand the difference, it feels very subtle but i think i can use james' wording from the document and the examples he gave
... we made a decision this summer about using markup in the name
... as a group we made this decision that we weren't going to use markup. we asked for comments from the orgiinal commenter. there hasn't been discussion since august 4th.
... i think it's reasonable to close it at this point

https://github.com/w3c/activitystreams/issues/338

tantek: we had a joint socialwg i18n meeting at TPAC
... we asked it to get minuted in our channel
... did this come up?

<rhiaro> https://github.com/w3c/i18n-activity/issues/178

aaronpk: here's the link to the IRC chat from the meeting https://chat.indieweb.org/social/2016-09-22#t1474555160487000

eprodrom: from my POV we had a discussion and made a resolution to discuss it

tantek: okay, close it with a link to their issue

<tantek> 354 was discussed here: https://chat.indieweb.org/social/2016-09-22#t1474557612462000

<tantek> https://github.com/w3c/activitystreams/issues/354

<csarven> eprodrom jasnell: Feel free to assign me to https://github.com/w3c/activitystreams/issues/376

<csarven> I don't have the q key on my kb

https://github.com/w3c/activitystreams/issues/354

eprodrom: in JSON-LD, you can specify the context property as either an object or as a URL
... it is nice to be able to just have that single URL, and one of the ways we've sold the AS2 to the AS community is "don't worry about JSON-LD just drop in this property and you have JSON-LD"
... the problem is the default language in JSON-LD is defined within this context object

<eprodrom> "@context": "https://www.w3.org/ns/activitystreams"

<Loqi> [Amy Guy] ActivityStreams 2.0 Terms

<eprodrom> "@context": {"@id": "https://www.w3.org/ns/activitystreams", "@language": "en"}

<Loqi> [Amy Guy] ActivityStreams 2.0 Terms

eprodrom: it would require some additional complexity to change the language
... but one of the selling points is we're going to make it easy to use it and now we've made it slightly harder
... the change proposes making it a SHOULD instead of MAY

tantek: the assumption that if you give authors a SHOULD that they're going to pay any attention to is false
... so richard is wrong about how authors behave when given directions like that

sandro: i think the examples are the important part

tantek: examples are far more effective than "should" language

sandro: i'm very concerned about the increased complexity. we're between a rock and a hard place.

tantek: it's something that noone currently implements

rhiaro: no other implementations support it, but we also only have english speaking implementations

tantek: this goes back to the larger point that the i18n group has yet to incubate or implement any localization in JSON

rhiaro: this is about how JSON-LD specifies how to do it, which is already implemented

eprodrom: so right now what we have is if you would like to specify the default language, you change the context from a single string to an object with a couple properties

sandro: can activitystreams supply multiple context URLs for a bunch of languages?

<cwebber2> https://www.w3.org/TR/activitypub/#object-without-create

<Loqi> [Christopher Allan Webber] ActivityPub

cwebber2: we added language to everything that has a name
... bengo pointed out he got very confused

<cwebber2> {"@context": ["https://www.w3.org/ns/activitystreams", {"@language": "en"}]}

<Loqi> [Amy Guy] ActivityStreams 2.0 Terms

<cwebber2> {"@type": "Create",

<cwebber2> "@context": "https://www.w3.org/ns/activitystreams",

<cwebber2> "object": {"@context": {"@language": "en"}}}

<Loqi> [Amy Guy] ActivityStreams 2.0 Terms

cwebber2: so we suddenly transformed the way the inner context worked
... the object that used to be inside got shifted around. it no longer had a list within a map, it just became a map
... and bengo said it was confusing and was some magic JSON-LD-foo
... and I had only put it there because someone told me to, so I removed it
... so having the language inside there can create some gnarly scenarios to ask from users

tantek: was this ever tested in JSON-LD itself?

cwebber2: i've never done anything like this in an actual implementation

tantek: is there any evidence that there are any actual implementations of JSON-LD that are doing anything with language in the way we are being asked?

cwebber2: i don't think we're going to be able to change the way AS2 does this by renegotiating with JSON-LD

tantek: i'm not suggesting we create our own solution. i'm asking this because putting in something that we think is solved because it's in a REC which if it turns out it's not solved is worse than not having it at all... because it's providing a false sense of having internationalization
... i don't know how to resolve this except for asking i18n to point us to some actual JSON-LD producers and consumers

<cwebber2> q/

eprodrom: i think there is a fundamental assumption here by the 118n group which is that each natural language proprety in a document should be explicitly marked up as the language it is, either by the default or specificlaly on the property
... there is a contradictory principle that it could be left out and understood from human world context (not @context)
... the mechanisms we hve for doing explicit language markup seem to complicate our JSON documents

tantek: which alone isn't enough reason to not do them

eprodrom: one thing we can do is bite the bullet and say we need to add this as a SHOULD and update our examples
... one problem with the -fr @context URL is naive consumers would fail if they are doing exact string matching on the context

<cwebber2> {"@type": "Note",

<cwebber2> "@language": "en"}

<sandro> {

<sandro> "@context": "https://www.w3.org/People/Sandro/activitystreams-fr",

<sandro> "name": "Martin created an image",

<sandro> "type": "Create",

<sandro> "actor": "http://www.test.example/martin",

<cwebber2> {"@type": "Create",

<sandro> "object": "http://example.org/foo.jpg"

<sandro> }

<cwebber2> "@context": "https://www.w3.org/ns/activitystreams",

<Loqi> [Amy Guy] ActivityStreams 2.0 Terms

<cwebber2> "object": {"@type": "Note",

cwebber2: we do have an alternate way of doing @language in AS

<cwebber2> "@language": "en"}}

<cwebber2> {"@type": "Note",

<cwebber2> "@language": "en"}

<cwebber2> {"@type": "Create",

<cwebber2> "@context": "https://www.w3.org/ns/activitystreams",

<Loqi> [Amy Guy] ActivityStreams 2.0 Terms

<cwebber2> "object": {"@type": "Note",

<cwebber2> "@language": "en"}}

<cwebber2> https://github.com/w3c/activitystreams/issues/354

rhiaro: so here's the thing
... the way we support extensions or using any of the vocabulary is you have to use the context. so you can have the AS2 context in a string and then an object with a bunch of prefixes and more contexts
... so if i'm publishing a bunch of data that uses AS and a bunch of other vocabularies, which I am, I have to list the context that way
... a plain AS consumer i'm happy to ignore the vocabularies it doesn't understand
... but if it doesn't understand context as an object then it's not going to recognize it in the first place
... so in order to make AS and JSONLD play together at all, all consumers need to be able to understand context as an object, not only as a string

sandro: anyone using extensions it woudl not fall back

tantek: that was a decision last time this came up

eprodrom: are we okay with bad consumers that will error out? we're not okay with that

rhiaro: so if we have to support @context as an object then we can put @language there as well
... so we don't need to put @language in all the examples because we can say that once
... but they have to be able to understand the @context as an object

eprodrom: i'm going make a suggestion that might be controversial for our friends in i18n
... we do have a specific language tag (und) undetermined. what we could use is sandro's cool hack for multiple contexts that define language
... so that we can have single string contexts for each natural language
... the base namespace could be "und"
... that would be roughly the same as what we have now
... it would be obeying the letter of the suggestion

rhiaro: but then consumers would have to string match against every language

sandro: we could say that the context is this string followed by a hyphen and a language tag
... if every example in the spec there's this extra curly braces and @language then it makes all the examples look a lot more complicated

rhiaro: we want the examples to be simple, but we want to tell consumers to be prepared to consume an object as context

<eprodrom> aaronpk++

<Loqi> aaronpk has 70 karma in this channel (1142 overall)

<eprodrom> You're a hero for scribing this

rhiaro: we need the majority of the examples to be simple, and then one section with the context being an object

<cwebber2> I agree that seeing the @context with an array does make it seem more complicated

<cwebber2> but

sandro: people making consumers are more likely to check against the test suite and read the spec

<cwebber2> I also think sandro's suggestion makes things complicated, based on my own implemntations

eprodrom: more likely simple consumers would do a string match

rhiaro: i think we should require consumers understand objects

<cwebber2> :<

eprodrom: consumers would do a string match and would notice that some percentage of what they were getting didn't need work with the string match and then they would do a regex match. then in a small number of cases they'd have an array thing and may or may not support that.

<wilkie> you're making consumers more complex to make producers 'feel' less complex?

<cwebber2> wilkie, yeah I think this makes things a lot more complex

rhiaro: they can not support extensions and drop all the terms and that's fine. but dropping any AS that includes extensions dropping the whole thing is a problem.

<wilkie> does the HTML lang attribute make HTML feel too complex?

sandro: that means people to me people won't use extensions at all because a chunk of consumers will fail if the AS includes extensions

eprodrom: what you're suggesting is probably to use context as an object in all our examples so consumers get used to seeing that

<sandro> { @context: { @id: "{

<sandro> "@context": "https://www.w3.org/People/Sandro/activitystreams-fr"} ... }

rhiaro: we're not telling publishers to publish an object, but we're telling consumers to expect an object

<sandro> oops

eprodrom: we're in a weird space where "have to" is a difficult term to define

rhiaro: are consumers going to look at the examples in the spec or use the test suite?

eprodrom: they'll probably run it against the test suite and then see things in production

<wilkie> the context should be the same regardless of what language you use so that JSON-LD generically understands two documents are comparable

<sandro> { @context: { @id: "https://www.w3.org/People/Sandro/activitystreams-fr" }

<sandro> name: "..." }

<tantek> sandro - has anyone suggested or tried this in the broader JSONLD / i18n communtiies?

<tantek> it sounds like experimentation in the context of a CR which I feel uncomfortable about

eprodrom: we have a base version that does not define language. we make N many language specific contexts you can pull in to use a single string context. and finally we have a MUST on consumers that they MUST handle a context as an object.

<cwebber2> tantek, I agree this feels like experimentation

tantek: i don't think we should be experimenting in CR

<eprodrom> "@context": "https://www.w3.org/People/Sandro/activitystreams-fr"

rhiaro: if a publisher wants to use language, they'll be okay with using the extra step of the object with language

sandro: about 1/4 of the examples we put the @language in

<cwebber2> yes, I agree having about 1/4 of the examples with @langauge

<wilkie> HTML lang isn't required. it defaults to 'unknown'. from a cursory glance, the HTML spec doesn't recommend it be there, instead it warns that absent 'lang' can produce unexpected results.

eprodrom: if we make language a SHOULD then we should have it in most of the examples

<cwebber2> I agree with SHOULD

<wilkie> I agree with SHOULD also

tantek: the problem is history does not support "SHOULD"
... you end up with lang="en" as the default

sandro: i thought the phrasing was "if you know the language you SHOULD put it there"

tantek: that is very different

eprodrom: that is very different

ben_thatmustbeme: <general agreement>

<cwebber2> tantek, is that what you wanted to raise?

eprodrom: another possibility is to make a simpler way to define language

<eprodrom> "language" : "en"

sandro: that's what chris was getting at

<eprodrom> "@context": "...", "language": "en"

sandro: this is the tradeoff for people using JSON-LD processors or not

<tantek> from r12a: "Could we either change MAY to SHOULD, or at least add a note to say that actually language information is quite important (since very few developers seem to recognise that)?"

<cwebber2> :\

rhiaro: we can tell consumers there are three ways to identify an AS document
... as a publisher, i want to know i can publish it in this object format and not have consumers fail

<Zakim> cwebber, you wanted to explain why sandro's suggestion is also complex, but suggest we just use @language in *some* examples.

<cwebber2> manu.sporny.org/2016/json-ld-context-caching/

cwebber2: i think sandro's path, while i like it in an entertaining way, but would be a huge mess
... i don't like resorting to string parsing as a shortcut either

<rhiaro> +1 not string parsing

cwebber2: there were a couple suggestions that were good. if you wrote in a note that it's really important to do language things that's really easy
... also 1/4 of the examples could include language and that'd be reasonable
... to set the expectation that sometimes this happens
... we're also not making it so that every example has this overhead of the language

<sandro> +1 say "it's really important" and 1/4 of examples use it.

<tantek> Aside: looks like JSON-LD test suite and reports did a bunch of explicit @language processing testing (syntactically) and that implementations in general "passed" in terms of transforming from input to output. http://json-ld.org/test-suite/reports/

<tantek> ht: ben_thatmustbeme for finding that link

<eprodrom> PROPOSED: Resolve issue #354 of Activity Stream with a Note that points out the importance of explicitly marking up language of natural language properties

<Loqi> [Gregg Kellogg] JSON-LD Implementation Report

<eprodrom> PROPOSED: Resolve issue #354 of Activity Stream with a Note that points out the importance of explicitly marking up language of natural language properties and use such markup for some examples

<sandro> +1

<cwebber2> tantek, yeah expansion and compaction will do smart things that "understand" the @context

<rhiaro> +1

<cwebber2> +1

<ben_thatmustbeme> +1

tantek: you said something earlier i thought was useful. "if the producer knows the language"

<cwebber2> tantek, but I've still found combining two documents together to be a bit tricky, unless you do an expansion then compaction... I guess expansion then compaction is the "cleanest way"

sandro: we can add text that explicitly says we omitted language from the examples for brevity and to avoid people blindly copy/pasting the language in the example without changing it

rhiaro: we're concerned about people blindly following the examples

eprodrom: they primarily use the language mechanism built in to AS (map) which is easier and more obvious

rhiaro: it doesn't need to be all the examples, just some of them
... what we're saying is you should do language somehow, but we can't include all possible ways so we include language in the simplest way to illustrate it

ben_thatmustbeme: that makes sense, we don't include every tag in every example

tantek: evan it sounds like you have some proposed text

<ben_thatmustbeme> +1

<eprodrom> +1

<cwebber2> +1

<sandro> +1 with explanation of WHY it's not in every example

<sandro> (to avoid hypocracy)

<rhiaro> +1 with explanation of WHY it's not in every example

<wilkie> +1 also with explanation so people don't accidentally think English is default

<rhiaro> PROPOSED: Resolve issue #354 of Activity Stream with a Note that points out the importance of explicitly marking up language of natural language properties and use such markup for some examples, and points out that it's not in every example because we want to avoid the copypaste EN everywhere thing

<cwebber2> +1

<sandro> +1

<aaronpk> +1

<rhiaro> +1

<wilkie> +1

<ben_thatmustbeme> +1

<ben_thatmustbeme> +0 i don't think its really necessary

sandro: our examples leave out language because we have a reason to leave them out

eprodrom: so maybe we just make it a should anyway

tantek: if you know the language

<tantek> note as an example: Twitter acknowledges it doesn't "know " the language but offers "language detection" in its developer API: http://www.digitaltrends.com/social-media/twitter-adds-language-detection-and-tweet-filtering-to-api/

eprodrom: we're going to say you SHOULD explicitly mark it up, but we're not going to recommend a particular way of doing so

rhiaro: right, based on the most appropriate way of doing it

<eprodrom> PROPOSED: Resolve issue #354 of Activity Streams with a SHOULD that points out the importance of explicitly marking up language of natural language properties if you know the language, and use such markup for some examples, and points out that it's not in every example because we want to avoid the copypaste EN everywhere thing

<cwebber2> +1

<rhiaro> +1

<aaronpk> +1

<sandro> +1

<ben_thatmustbeme> +1

<eprodrom> +1

RESOLUTION: Resolve issue #354 of Activity Streams with a SHOULD that points out the importance of explicitly marking up language of natural language properties if you know the language, and use such markup for some examples, and points out that it's not in every example because we want to avoid the copypaste EN everywhere thing

tantek: i think this will communicate a positive response to richard which is great
... i also want to acknowledge what evan said about people ignoring it, like on twitter, where they don't know

<cwebber2> btw

<cwebber2> I asked the i18n team to give guidance on this

<cwebber2> https://www.w3.org/International/wiki/LanguageDetection

cwebber2: just pointing out this link. when the i18n team asked us to add SHOULD to activitypub, i said this sounds great but realistically i don't know the language. i asked them to produce a document so that we can link to it, so they started this
... here's some guidance on how to figure out what language to apply

tantek: if they publish this as a note i'd be okay, but i don't feel comfortable referencing a random page on their wiki
... chris if you want to reference this page, i encourage you to reach out to them and request they publish it as a NOTE

cwebber2: i can do that

<rhiaro> I would like to point out that I have a multilingual blog post now http://rhiaro.co.uk/2016/11/decentralising-social-tokyo (half English half Japanese)

eprodrom: this does mean our examples will have a warning on the validator

rhiaro: that's fine, because we have a good reason

sandro: maybe we add a checkbox to the validator asking whether you know the language

rhiaro: an extra checkbox on the validator saying explicitly to validate language stuff would be good

<tantek> FYI: the related i18n issue was closed 23 hours ago: https://github.com/w3c/i18n-activity/issues/205

LUNCH

<cwebber2> I'm on the phone

<cwebber2> what's going on

<tantek> we are adjourned until 14:00 EST for lunch

<cwebber2> aaronpk, ok what issues are we working on?

<cwebber2> aaronpk, making a proposal?

<cwebber2> aaronpk, wasn't totally clear to me

<cwebber2> aaronpk, and do you want to do so now, or in an hour?

sorry hi

so we need to write up a description of the community group in order for the group to propose creating it

starting this here: https://www.w3.org/wiki/Socialwg/swicg

<cwebber2> aaronpk: could we take 20 minutes to eat, then regroup?

sure

<cwebber2> aaronpk: cool.

say something on the phone when you're back, i'm stepping away

<cwebber2> kk

<tantek> julien departed ~noon (for the minutes)

<cwebber2> hi

<cwebber2> aaronpk: I'm back. Dialing back in now

<cwebber2> aaronpk: would you prefer to handle this over voice or irc?

there's a whole discussion happening here so maybe irc?

<cwebber2> sounds good to me

actually i'll put on headphones

oh i got disconnected too, one sec

<cwebber2> ok!

so i started this https://www.w3.org/wiki/Socialwg/swicg

<cwebber2> good start :)

we really just need a little description for the form to submit here https://www.w3.org/community/groups/propose_cg/

<Loqi> You must be logged in to propose a group or ( Request a W3C account ).

<Loqi> Once you are logged in, you will provide the name of the group, a short description, and a short name. The shortname is used in various places (such as automatically generated U...

but also need to agree on the "ground rules" for the group especially around communication methods

<cwebber2> so I think it would be nice to use gitlab as communication, because it supports afaict pretty much every github feature and *might* end up federated

<cwebber2> we can just use the ironic mega-silo for now

yeah i agree gitlab is probably the best bet, but i'm concerned about it for two reasons

1) it doesn't actually federate so it's kind of an aspirational use of gitlab until it actually does

<Loqi> that doesn't make sense

<cwebber2> shh Loqi

2) it means we're on our own little island whereas on github at least everyone already has an account and there's good cross-linking of issues and stuff with other repos we'll be referencing

<Loqi> [@Zaninel] Millions of dollars in engineering has led us to this point... https://pbs.twimg.com/ext_tw_video_thumb/799080862113480704/pu/img/FwQMKwmV4RQPhHkf.jpg

in an ideal world, this issue tracker would just be an aggregator and i would be posting comments from my own software (e.g. my own gitlab instance) and they'd just be aggregated at this group's

but in the mean time .... :)

<cwebber2> re: 1) https://gitlab.com/gitlab-org/gitlab-ce/issues/4013

yeah, it doesn't even have a proposed solution yet right? also that's just about pull requests, not about issues

<cwebber2> right

<cwebber2> aaronpk: the other side of things is that I know some people who won't use github because it requires proprietary javascript, and that's a no-go for them

i'd love to be able to demonstrate federated issue tracking

<cwebber2> so gitlab is not ideal

<cwebber2> but I think it might be a bit beter

<cwebber2> better

<cwebber2> there's tradeoffs for sure; I hear you on it's more of an island

<cwebber2> though, in theory github users can log in with their github accounts

<cwebber2> so it's not so bad

<cwebber2> but they won't get notifications from one notification UI

gitlab uses omniauth and i'd love to be able to pull in the indieauth gem for it

that way i don't even need a separate gitlab identity

<cwebber2> that'd be cool

<cwebber2> so, aside from github vs gitlab

<tantek> can you file an issue for them to pull it in?

actually i want to update that gem to do the auth server discovery itself rather than delegate that to indieauth.com first

<cwebber2> would also be good to discuss scope of what we intend to accomplish. the current description on that page is pretty good, though

<cwebber2> aaronpk, something that doesn't seem to be described, but I want to continue

<cwebber2> is related things to make the federated web more usable, like spam filtering / anti-abuse tools

there's not much in that description, just half a sentence pulled from our current charter :)

<cwebber2> aaronpk, I assume the CG will be a good fit for that?

<cwebber2> right CG

<cwebber2> oops

I *think* we can do whatever we want in the CG?

<cwebber2> aaronpk: just making sure you and I agree on what's good :)

yeah i'd like to be able to work on things like that too :)

<cwebber2> so we're basically thinking vocabularies and protocols to support the distributed / federated social web, and surrounding technologies that "support" them (such as spam filtering and anti-abuse tooling)

that works

go ahead and add that to the page?

<cwebber2> aaronpk: cool, doing so

<cwebber2> aaronpk: updated description, refresh

cool

so the group will get a wiki on w3.org by default, do we want to use that?

<cwebber2> aaronpk: I think we might as well

<cwebber2> aaronpk: the socialwg has used a wiki to good effect mostly

yeah seems like it

<cwebber2> aaronpk: so I suggest we use gitlab, but yes, since it's not federated, use gitlab.com. It'll have some tradeoffs, but given the work we're doing I think it's worth it, and it might allow some users in related spaces to participate who wouldn't on github

oh gosh gtilab.com??

<cwebber2> gitlab.com

<cwebber2> not gtilab.com, I don't know what that is

lol oops typo

i think i am actually more opposed to using gitlab.com compared to github.com

<cwebber2> aaronpk: ok, do you have an alternative?

<cwebber2> http://www.gtilab.com/ appears to have something to do with Mattel and Hasbro

self-hosted gitlab?

<cwebber2> aaronpk: I'm ok with that, though I don't volunteer to host it :)

let me check what the recommended server stats are for it

<cwebber2> ooh I like that it uses capital-S Scheme

<cwebber2> scheme based expert system ;)

<cwebber2> 1980s MIT classic ;)

wow they want 4gb ram free

know anyone who could sponsor hosting this server? ;-)

<cwebber2> aaronpk: ping mattl :)

ha oh yeah

<cwebber2> aaronpk, I believe there's https://git.gnu.io/explore/projects which I think mattl got some gitlab people to sponsor/maintain

<cwebber2> but I don't know much about it or how that maintenance is done.

well i'm happy to run the server cause i already run the indieweb server and more for my own sites

<cwebber2> aaronpk: ok!

<cwebber2> aaronpk: so I say we still resolve at this time to do something like

<mattl> aaronpk: i can get you a server

<cwebber2> oh or there we go

\o/

i don't even need a big one, just a 4gb VPS is enough

<mattl> yup. we can figure it out next week

that'd be awesome

how would you describe our group's use of the wiki?

<tantek> conclusions

<tantek> everything besides discussions

cwebber2, can you hear us? i switched back to the speakerphone

<cwebber2> hi

<cwebber2> I need to dial back in, one se

<cwebber2> c

<cwebber2> whoa

<cwebber2> sounds like a room full of robots

i'm gonna call back in

<cwebber2> kk

i may have confused it by switching from headphoens to speakerphone

<cwebber2> sounds better now

<wilkie> I hear it

<rhiaro> The other thing is the bus factor of aaronpk being in charge of everything

<rhiaro> I think we just need another route clearly documented for them to participate

<rhiaro> so they're not dismissed altogether

<tantek> is IRC still not good enough as a start?

<rhiaro> hey cwebber2 what license should my software be

<cwebber2> rhiaro, that depends on what you're writing; I think https://www.gnu.org/licenses/license-recommendations.html is a pretty nice guide

<rhiaro> thanks!

<cwebber2> rhiaro: for clarity, I'm using Apache v2 for ActiviPy (which is a library only), MediaGoblin is AGPLv3, but for newer server applications, especially ones that are both server + client, I tend to just use GPLv3+ (though AGPL is still a good choice)

<cwebber2> rhiaro: but that page probably says better than I will here

<tantek> https://wicg.io/

ah right we already have this github org: https://github.com/w3c-social

we'll make a new github org for this

cwebber2, check out the updated page https://www.w3.org/wiki/Socialwg/swicg

<cwebber2> aaronpk: looks good

<cwebber2> aaronpk: also can we agree that we don't try to pronounce "swicg" as one word

swickguh

<cwebber2> it sounds a bit like retching

i think the "g" is silent

<tantek> resume meeting postponed til 14:15

<cwebber2> swick-gee is not bad

<tantek> swi-kig

is that a hard or soft G

<tantek> because WICG is pronounced wi-kig

<tantek> hard

<tantek> like keg

<cwebber2> swi-kig is ok

<cwebber2> (factoid: did you know that python's WSGI protocol is supposed to be prounounded "whiskey"??)

(added a sentence about not using github.com)

brb

<tantek> not using or not having to use?

<wilkie> that's unacceptable. soft-g, hard-g and now g-that-sounds-like-k

(if you don't want to use)

<cwebber2> I think we're good

<ben_thatmustbeme> scribenick: ben_thatmustbeme

SWICG

<csarven> swicg comes a bit close to SWIG. Semantic Web Interest Group.

<csarven> And that had been around forever

<eprodrom> SOC-CG

<csarven> Has

<eprodrom> What about WebSub?

aaronpk: i don't want to derail this with another long round of choosing a name for something

<rhiaro> SWUB

<csarven> Do we need a Google docs for this? Like for push

aaronpk: AnnB proposed Social Web Incubator Community Group.

<csarven> (not a real suggestion)

tantek: which paralells Web Incubator Community Group wicg.io

sandro: the "Incubator" seems bizarre to me

<aaronpk> https://www.w3.org/wiki/Socialwg/swicg

aaronpk: anyway... i did not plan on talking about the name, i planned to share the description that cwebber2 and I came up with, its not a charter length description, its what shows up on the group page and the list of all groups

<aaronpk> "The Social Web Incubator Community Group is to continue and extend the work of development of vocabularies and protocols to support the distributed / federated social web, as well as "supporting" technologies (such as anti-abuse and anti-spam tooling suitable for a federated network). This group continues the work of the W3C Social Web Working Group."

aaronpk: i would like feedback on if this is clear
... most of that is from our charter

eprodrom: "the work of development" can just be "development", can get rid of the scare quotes

sandro: related technologies

eprodrom: "tooling" is an interesting word there

<cwebber2> +1 techniques

tantek: techniques

eprodrom: the rest sounds great

aaronpk: the last sentence about 'continues the work of' is that appropriate?

sandro: seems fine to me

<cwebber2> agreed

eprodrom: can i throw a couple of other things in here, to the extent that we have registries, like the namespaces for as2, can we add that in there

sandro: i read that as being in there, but we can add that in there

tantek: do we need that as part of the summary?

aaronpk: the first post will be a longer summary of the group

eprodrom: who is going to update erata for the specs? thats something the CG should handle

sandro: the group can't publish anything normative, but they can update the list of erata

eprodrom: i don't know if we need to call them out, but those are definitely two important things for the CG to manage

<Zakim> tantek, you wanted to say s/for a federated network/for the open web and to also make it clear about incubating

tantek: the phrase "suitable for ..." i was saying "suitable for the open web"
... to keep other types of federated network work out of the group
... i specifically say open web to set the bar high here
... the summary should clearly call out the method of incubation
... we will only have during the period of the extension that we presumably get, to have a WG, whereas the webWG has one

<tantek> e.g. https://wicg.github.io/admin/charter.html

tantek: so i want to explicitly call that out
... i think there is a lot from this page that i find is very useful
... do we need a charter? I don't think so

aaronpk: i don't think so

tantek: i'm not saying you should copy that charter, but its a good place to mine from
... i think that middle paragraph that you cited is a pretty good one
... and the key thing is that its not saying what group its going to, but if they have enough prototypes they can use that to propose a new working group

s/sandro: I'm a little teapot, short and stout//

aaronpk: does the description look good?

tantek: would you consider putting formats in place of vocabularies?

aaronpk: i can add formats

<aaronpk> added formats

tantek: for participation i would say the same requirements from the wicg charter
... where you have to sign this license agreement

sandro: the bit about maintaining the namespace, thats beside the point, its co-ordinating extensions
... traditionally extensions use different namespaces

eprodrom: i think that from my perspective, if i were trying to track down whose job it is to update that namespace

sandro: hopefully that would be in the namespace, its a rule that every page on w3.org is signed
... it says editor, amy guy at the top

cwebber2: there was a review process for joining this group, is there one for this group

tantek: there is no such requirements for community groups

aaronpk: as i remember joining the group requires you to sign this stuff automatically

sandro: the one surprising thing there is that if you work for a w3c member, your AC rep has to review it, but the system does it
... public can easily join, but member organizations without going through their company first
... its just surprising

tantek: also in participation, we haven't had a history of posting things on a blog for this group, so i'm not sure that a claim that we would do that in the future is something we should promise. if you want to say that we MAY..

aaronpk: i wanted to make sure to make it clear that you would not expect to see discussion on the blog

tantek: so maybe put that as a catch all absolutely last
... and then after the 'the groups wiki is used for documenting conclusions' i have found those wikis incredibly dead
... i would suggest we instead use the w3c's wiki

sandro: why is that bad?
... because they have to be in the CG?

tantek: you get less people randomly joining in and making editing your wiki

<aaronpk> wait i didn't say that

sandro: if your TRs are on your wiki, as you used to do...

<tantek> e.g. https://www.w3.org/wiki/Swicg

sandro: i think they raised the bar on the wiki, where you have to be in a group to edit the wiki
... woah, w3c says it requires member access, but you (aaronpk) aren't a member

tantek: did you register as a swicg org on github?

aaronpk: in order for me to submit this, we do need to have a name. Are there any serious objections to the current name?

sandro: so swikig is how its pronounced? i don't see a lot of incubation in that description

tantek: thats why i asked for it up front in the description

sandro: i'm okay with it

<tantek> PROPOSED: create the Social Web Incubator Community Group (SWICG) with description and participation as noted https://www.w3.org/wiki/Socialwg/swicg with aaronpk and cwebber as co-chairs

<rhiaro> +1

<eprodrom> +1

<cwebber2> +1

<ben_thatmustbeme> +1

<aaronpk> +1

<sandro> +1

RESOLUTION: create the Social Web Incubator Community Group (SWICG) with description and participation as noted https://www.w3.org/wiki/Socialwg/swicg with aaronpk and cwebber as co-chairs

<KevinMarks> Bait and Swicg

<KevinMarks> +1

sandro: he's just being goofy

eprodrom: he's being Kevin

<wilkie> +1

<eprodrom> ha ha

<aaronpk> https://www.w3.org/community/blog/2016/11/18/proposed-group-social-web-incubator-community-group/

<aaronpk> https://www.w3.org/community/groups/proposed/#swicg

<Loqi> This is the list of proposed Community and Business Groups. To express support for a group, you must have a W3C account. Once a group has sufficient support (five supporters), W3C announces its creation, lists it on the current groups page, and peopl...

<aaronpk> https://www.w3.org/community/swicg/

<aaronpk> we're live!

<cwebber2> should I hit join on that or am I added as a chair and that happens automatically?

<cwebber2> I'm trying to figure this out, hold on

cwebber2, you need to join

<cwebber2> "I confirm that I have a significant employment relationship with GNU MediaGoblin (If incorrect update affiliation first) "

<cwebber2> not super sure if GNU MediaGoblin really counts as significant employment right now

<cwebber2> but I guess

<csarven> wait.. you read agreements/TOS/license/patents warnings? Who are you?

<csarven> I see checkboxes, I'm like.. I got this.

<cwebber2> csarven: haha

<Loqi> rofl

<aaronpk> cwebber2, once you join i will make you a chair

<cwebber2> I did join!

<cwebber2> "Your request to join the group with only individual commitments has been sent for review to the W3C Staff. You will be notified by email when your request is approved or denied."

<cwebber2> should I have said as my "employer"? :P

<cwebber2> with GNU MediaGoblin only remotely sort of being my employer

TOPC: AS2 non-spec issues

AS2 non-spec issues

eprodrom: before we go to test suite, we have closed 10 issues, they are updated in the ED, the most important is the issue around name and summary from last week, so we push them to the ... whats the next step?

sandro: are there normative changes?

eprodrom: this was a change to a SHOULD, (reference previous resolution)

<tantek> stubbed: https://www.w3.org/wiki/Swicg

sandro: do we think the language thing effects implementations?

tantek: thats a publisher requirement, isn't it?

eprodrom: the syntax is the same

sandro: wasn't it true that you could have a conformant reader, that just looks for the string

eprodrom: can we capture that as an issue? thats orthoganal to what we just resolved

sandro: you could have made a conformant reader before but doesn't now right?

eprodrom: we did have examples before that had the object format for context
... we didn't specify that you don't blow up

rhiaro: so is this about just making it explicit? or do we not need to do anything? how about I open an issue and we can discuss it there.

eprodrom: even to call out all the places where its like that in examples

sandro: given that we have an example in the spec, its not a breaking change, it was clearly our intent
... we don't need a new transition call, but we need a new CR

rhiaro: i'm going to confirm that its 4-weeks

eprodrom: we are adding 2 SHOULDs, they probably will not break any implementations

sandro: if those issues get in to ED, and those get circulated to all implementations, that may be good enough to not need a new CR

tantek: and explicitly call them out as normative changes

sandro: they aren't
... they match examples in the spec now

tantek: okay, they are clarifications that effect normative text, we should call that out as seperate from editorial changes, we should be up front about that

sandro: i agree we should be up front about it, but there are many editorial changes in normative text

tantek: i don't want it buried in editorial changes, put them as seperate 'important clarifications'

sandro: can we fix change log to link to previous version's change log

eprodrom: so it sounds like its good to not publish a new version yet

rhiaro: i think it would be good to get these changes formally in the CR version if we are going to use those 4 weeks anyway, and its the right thing to do

tantek: evan, what would you like to do next?

eprodrom: lets discuss test suite
... a couple of things happened, we have updated changes that match changes we made. such as using https context URL, and the name issues

<cwebber2> huh?

<eprodrom> https://github.com/w3c/activitystreams/issues/351

sandro: what? https context URLs?? there was a whole blog post on why we don't do that. HSTS solves that anyway
... which string are they supposed to be looking for if they are using just string matching?

eprodrom: what i'd like to do is while you read that over, we continue talking about test documents
... updated the test docs to match the current state of the spec. second the validator, there are outstanding issues on the validator, nothing that can't be fixed but they haven't been done yet

sandro: #351 seems fine, the namespace didn't change, the context changed

tantek: but its inconsistant with annotations

eprodrom: can we deal with this next?
... we have outstanding issues on the validator, i'm not sure what our process is for that. Do we need to have no issues on the validator? Most of them are around loosing the validator.

rhiaro: i have issues of validating because of ...

eprodrom: yes, but thats just a matter of how to get the AS2 on to the stream, if you paste it in, it still validates
... there are certainly things to improve, but does this stop us from getting to PR?

rhiaro: it seems harmful to me asking people to use the test suite and it doesn't work correctly

tantek: is it blocking people from implementation reports?

eprodrom: what i'd like to do is go through these and mark them as blocking or not blocking. my goal is to have no issues, but get blocking ones out of the way first, then non-blocking ones
... i'm going to give that to myself as a task
... as far as implementation reports go, right now we have 4 implementation reports for AS2, 1 implementation report that we are waiting for

rhiaro: i expect csarven will submit one as well

eprodrom: i think we'll land somewhere between 10 and 12 implementation reports before we are done

sandro: and how many are updates from AS1?

eprodrom: i know of 2 that i think are

tantek: how many are besides the editors?

eprodrom: right now none are from me or james

tantek: how many outside the group?

eprodrom: only one is from outside the group
... is it fair to make the validator to be an implementation

aaronpk: i don't feel like the validator is a valid implementation

sandro: i feel like AS2 its more valid to use it, but we might as well err on the side of not including it

eprodrom: i will do an implementation report for the validator, but if there is an question, we'll remove it
... we have a lot of features that are not implemented at all
... every part of the vocab is a feature thats at risk, except for a very small core

sandro: are they literally at risk?

tantek: we did, we had a long discussion about that
... are there any non-at-risk features that have only 0 or 1 implementations?

eprodrom: i think not

tantek: do we have a summary of them?

eprodrom: thats probably my next task

tantek: with the required features first, then at-risk after that

eprodrom: if i was going to do an estimate, they are mostly the core-types and their properties, as well as the activtypub parts
... that will probably put us in the area of 5 or so extension
... i think that we may get down to a small enough vocab document that we question why we have a seperate vocab document

sandro: so if i implement it a week after REC, its referenced all from the rec, and people can still use it, there is not a big loss if we cut things from the spec
... thats the impression that i get

rhiaro: it kind of matters of where you go for documentation

sandro: i really want the extensions to be highlighted
... does the spec link to an extensions page?

eprodrom: we do not have anything that says "look here" i don't know where that would be

sandro: we could just say the namespace document, the downside is 6-months from now there will be no w3c staff member paying attention to that

rhiaro: isn't that what the CG is there for?

<tantek> activitystrea.ms source: https://github.com/activitystreams/website

sandro: who has control of activitystrea.ms?

eprodrom: i might, i know that jasnell does
... going back to implementation reports. I think that we have a couple we would still like to see.
... i need to work on a summary table, but let me see if i can say this correctly

<cwebber2> they will at least *use* activitystreams implementations

eprodrom: activitypub implementations will be activitystreams implementations as well

tantek: if we wait for those implementations, would we have more properties exit CR?

eprodrom: yes

tantek: i think we should seriously consider that, its a bit of a delay

eprodrom: i have no complaint with that, i think it would be a better spec with us waiting for more activitypub implementations

tantek: that would buy you a lot of time for CR changes
... just another considerations

<cwebber2> ok everyone

<cwebber2> I need to leave

<cwebber2> is there anything you need me for last minute?

eprodrom: back to one thing, if the vocab gets short enough, is it worth us merging the two documents?
... the core is very structural, and the vocabulary is much more focused toward the social domain?

sandro: can we wait and see?

tantek: we have 1 test suite, 1 validator ..

sandro: its more been easier logically

eprodrom: would we need to do a new CR?

tantek: i have no idea

sandro: its just editorial really
... you just publish a tombstone

tantek: you don't redirect?
... its worth it for future readers of the spec, it will help implementations
... i would want the short names to redirect too

sandro: well TRs don't redirect, its a symbolic link, its sort of the normal flow so in this case you actually would want to do a redirect

tantek: there are some

sandro: on another issue, the name / summary issue

<tantek> e.g. w3.org/TR/CSS2 shows CSS 2.1

eprodrom: those have been changed in the ED (gives examples)
... i've given myself a task to fix the ones that sound like they were auto-generated to sound like they were created more by a human being
... i think thats all i have to say about AS2

tantek: we have 6 issues that were opened today
... from amy and csarven

eprodrom: and some from me

tantek: do we want to spend time disucssing the 1 normative issue, #377

<tantek> https://github.com/w3c/activitystreams/issues/377

<csarven> I'll have a PR ready in the next minutes

tantek: this is amy's issues with is seperate from the language issue

rhiaro: i think we resolved that, but we should make it explicit

sandro: and we need tests for that

<eprodrom> http://w3c.github.io/activitystreams/core/#jsonld

<Loqi> [James M Snell] Activity Streams 2.0

eprodrom: (quotes doc)
... we could have an implementer's note here that the value of the context property could be a string, object, or array

(after some discussion)

eprodrom: I'd be fine saying there are only these 3 ways to review it

rhiaro: my only concern is there are other versions that could be generated by JSON-LD libraries

sandro: before we publish that in a CR, lets try to get some review of that by a JSON-LD expert

(discussion of exact wording)

eprodrom: (updating issue with suggested wording)

<csarven> https://github.com/w3c/activitystreams/pull/381 should be o-kay

eprodrom: i'm comfortable with that

tantek: does that resolve it for you amy?

rhiaro: yes

tantek: that brings us back to just editorial
... given that we are talking about waiting more than 4 weeks, i think its better that we republish a new CR

eprodrom: do we need a new call for that?

sandro: as long as there are no normative changes

tantek: but thats manual still i believe and we need to do a group telcon to resolve that update

eprodrom: are we having a meeting on tuesday?

tantek: its scheduled right now

eprodrom: we typically don't do them right after a F2F

aaronpk: we just skipped one before this
... we were planning on reviewing pubsub edits

eprodrom: we'll have enough to work on then, its worth meeting

tantek: we're completed for AS2, i was going to have an update of PTD, but i wasn't able to get done what i wanted

eprodrom: is it reasonable to hold off
... what about name for pubsub?

<csarven> SoftHub?

PubSub renaming

<tantek> choices: WebFollow WebSub WebSubscribe

rhiaro: i'm for WebSub

<csarven> WebSub sounds good.

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

eprodrom: my thought is the editor who is still here picks his favorite, then we +1 / -1 on that
... and if that doesn't work, we go to the next

aaronpk: i know juliens preference on this, and i now agree with it

sandro: which is

aaronpk: WebSub

<sandro> PROPOSED: Rename pubsubhubbub to websub

<aaronpk> WebSub specifically

aaronpk: mainly WebSub and not WebSubscribe because of the similarity to PubSubHubbub

<csarven> Long name is what? WebSubscribeulator?

<sandro> aaronpk: this is my first choice, and I believe Julien's

<wilkie> that sounds good actually. easy to search.

<sandro> aaronpk: with WebFollow being available later for the user-facing feature, like Julien's Follow feature

tantek: do we have any non-native speakers that can confirm the pronouncability of it?

aaronpk: i believe phonetically its fine

sandro: julien was ok with it and he's a non-native speaker

<sandro> PROPOSED: Rename pubsubhubbub to websub

<eprodrom> +1

<rhiaro> +1

<sandro> +1

<wilkie> +1

<ben_thatmustbeme> +1

<aaronpk> +1

<aaronpk> PROPOSED: Rename PubSubHubbub to WebSub, and move the shortname pubsub to websub

<rhiaro> +1

<eprodrom> +1

<csarven> +1

<sandro> +1

<ben_thatmustbeme> +1

<wilkie> +1

<rhiaro> +9000

<csarven> There is no company with that name right?

RESOLUTION: Rename PubSubHubbub to WebSub, and move the shortname pubsub to websub

<aaronpk> there are surprisingly few search results for websub in general

<aaronpk> "About 78,700 results"

tantek: the only other thing on the agenda was SWP

<eprodrom> csarven: https://www.google.com/search?q=websub

SWP

rhiaro: i have not had a chance to work on it recently
... can bring it up next meeting

rechartering

<csarven> whois websub.org

tantek: we have already agreed to request an extension, which we believe we will get
... putting that in our extension request for 1 to 2+ months with one outlier
... there is a lot to consider for our recharter
... the recharter would start in may

sandro: i think we should ask for 6 months to keep it simple

tantek: that buys us additional runway for if something comes up in PR
... do we feel that it is good?

<tantek> PROPOSED: Request 6 month extension to take our drafts to REC, with plans already to take Micropub to PR in January, AS2 & ActivityPub to PR in February, WebSub (formerly PubSub) to PR in April; no f2f meetings expected, monthly telcons.

<csarven> 6 months sounds good/safe. We probably only need 3 (just pulling that out of thin air). If we finish early, we go for lunch early.

<sandro> csarven, yes, I told the group I've owned websub.org for ages, and I'm happy to hand it over to whoever's acting for the group on this.

<ben_thatmustbeme> +1

<rhiaro> +1

<aaronpk> +1

<sandro> +1

tantek: and you are available to chair in that time as well correct eprodrom?

eprodrom: yes

<eprodrom> +1

aaronpk: whats the CGs work in that time?

<csarven> +1

tantek: everything outside of our specs, building extensions, bringing people in

RESOLUTION: Request 6 month extension to take our drafts to REC, with plans already to take Micropub to PR in January, AS2 & ActivityPub to PR in February, WebSub (formerly PubSub) to PR in April; no f2f meetings expected, monthly telcons.

<csarven> Should LDN be in that PROPOSAL?

sandro: that comes back to your rechartering question, if we see a lot of companies that are w3c members or are potential members

<tantek> PROPOSED: equest 6 month extension to take our drafts to REC, with plans already to take Micropub to PR in January, LDN & AS2 & ActivityPub to PR in February, WebSub (formerly PubSub) to PR in April; no f2f meetings expected, monthly telcons.

<tantek> (sorry for the omission)

sandro: thats a good indicator for rechartering

<aaronpk> +1

<wilkie> +1

<rhiaro> +1

<ben_thatmustbeme> +1

RESOLUTION: request 6 month extension to take our drafts to REC, with plans already to take Micropub to PR in January, LDN & AS2 & ActivityPub to PR in February, WebSub (formerly PubSub) to PR in April; no f2f meetings expected, monthly telcons.

<rhiaro> PROPOSED: rename CG to *swish*

<rhiaro> Social Web Incubator Says Hi

sandro: the question is what happens to all our .rocks domains

<csarven> Oh..

tantek: sounds like a good thing to add to the list of things we intend to maintain after the WG finishes

sandro: its okay to say that you own that and its not the group that owns it

aaronpk: i'd be surprised to have the community group own a domain since they aren't official
... if i was w3c, i would not want be endorsing the code written by the CG

sandro: is what eprodrom getting at is can the CG find someone to take over as2.rocks?

eprodrom: i think aaronpk has more domains there

tantek: capture it as something for the CG to deal with
... going back to rechartering
... as sandro pointed out there is a need for 20 members

<jasnell> going back to earlier conversation: the activitystrea.ms website domain is owned by Chris Messina. Publishing is via github site

tantek: we don't even need to figure out what to put in a recharter now, we can say, what the CG incubates can determine what we recharter with

sandro: so the charter is open ended?

(talking about web platform CG)

<jasnell> previously, Chris indicated that he is willing to transfer ownership as long as there is agreement to transfer it back to him should there be no desire to continue maintaining it

eprodrom: i don't know why we would recharter right now, we have a CG that will be maintaining things, my guess is they will not be making many proposals in the next 6 months, there is a lot of implementation work to do
... there is a lot of advocacy work to do
... i don't think there is a lot of work that we have on the table

tantek: we have very deliberately been trying to reduce that and focus the group, thats why there isn't a lot left

<tantek> (explicitly)

eprodrom: i guess my question is, if we get to the end of our extension, and we don't have a lot of work to do, are we ok with that?

tantek: i think there is a growing set of communities because of the work we are doing, i forsee growing momentum of the next 3-4 months
... second, if i had to drop a recharter today, i could 3 things on it that have multiple implementations already
... Vouch, Salmention, and private webmentions
... if i were to say that in 6 months, we are going to draft another charter, i think we have a good case
... i'm not making a proposal

sandro: i see this as try to build up pressure in the CG, if it turns out to be WG levels, thats ok

eprodrom: if i may ask, i understand there are WG that don't have a fixed schedule as we do.

tantek: they do, every single one of them has to recharter each time

eprodrom: is it more a horizontal group

tantek: horizontal groups don't typically produce their own specs

<csarven> speaking of after 6 months

<eprodrom> Cool

<csarven> No audio for me. It cut off. I came back on .. still no audio.

<eprodrom> csarven: working on it

<csarven> thanks

<wilkie> sound good

<aaronpk> i think google voice timed me out

<aaronpk> should be back now

tantek: i think we leave the question of rechartering open
... i wanted to say that explicitly, and remain hopeful, and there is enough work to recharter if enough people decide that

aaronpk: is it good or bad to have the group expire and then recharter again having had a gap?

sandro: it doesn't really matter

FIN

<rhiaro> trackbot, end meeting

- DRAFT -

Social Web Working Group Teleconference

18 Nov 2016

Attendees

Contents

LDN

10 minute break

AS2

LUNCH

SWICG

AS2 non-spec issues

PubSub renaming

SWP

rechartering

FIN

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output