<trackbot> Date: 27 October 2014

<harry> How many people want to participate remotely?

<evanpro> Hi everyone

<harry> elf, can you say something

<harry> we are testing?

<Arnaud> hi

<Arnaud> folks are still gathering here

<Arnaud> long queue at the registration

<elf-pavlik> sounds busy there :)

<evanpro> Arnaud: http://evan.prodromou.name/files/TPAC/

<evanpro> We're taking a few minutes to let the people waiting in the registration line get to the room

<evanpro> I'm going to check on the line and see if we're about ready to begin

<elf-pavlik> harry, thanks for setting up VoIP conference! looks like only i needed it :)

<Arnaud> we'll start at xx:45

<harry> I think others will dial-in

<harry> I'll send out the info to the mailing list.

<AnnBassetti> Arnaud: introduces the meeting, process, timeline

<AnnBassetti> scribenick: Annb

<AnnBassetti> Arnaud: 2 items submitted as Working Drafts (on Activity Streams) ...

<AnnBassetti> ... need decide what next steps are

<AnnBassetti> ... goal for meeting is to come out with upcoming agenda and actions

<AnnBassetti> ... tomorrow the IG will bring Use Cases, for consideration by WG

<AnnBassetti> ... WG has a specific timeline (2 years), in which to do its work

<AnnBassetti> ... IG doesn't have deadline, so has more flexibility to explore

<AnnBassetti> Evan: if WG doesn't finish some deliverables, or new ones come, when this charter runs out .. then what happens?

<AnnBassetti> Arnaud: need new charter, especially for new deliverables .. due to patent agreements

<AnnBassetti> ... today we'll work all day

<AnnBassetti> ... tomorrow we'll break between 11-2 for the ad-hoc meetings

<AnnBassetti> ... we'll adjust the agenda based on discussions this morning

<lehawes> What do IG members think about having a breakout meeting/lunch from 11-2 tomorrow? Who is available?

<AnnBassetti> I'm available

<AnnBassetti> Harry: WebApps Sec WG working on types of protection for various types of content

<AnnBassetti> ... Annotations WG has interest for discussions with Social WG .. likely Tues afternoon

<AnnBassetti> ... <discussing when it'd be good for 2 groups to meet, for cross-fertilization>

<lehawes> Thanks AnnBasetti! Anyone else from the IG interested in working together in a breakout tomorrow?

<AnnBassetti> Lloyd: +1 to Harry

<EdK> I can split out into IG session

<harry> scribenick: AnnBassetti

<elf-pavlik> Arnaud, yes I can hear!

<lehawes> OK EdK. Thanks!

<some attendees having trouble getting into IRC .. trying to figure it out>

Evan: presents "Social WG Status"

Social WG

<elf-pavlik> http://evan.prodromou.name/files/TPAC/

<evanpro> Thanks Elf!

Evan: reviews history of various groups on "social" since 2010 at W3C ..
... immediate causes:
... convergence of open standards work
... increased collaboration between OpenSocial and W3C
... Activity Streams 2.0

<AndyF> ;join #social-chat

<harry> AndyF, everything is just in #social

<Loqi> AnnBassetti meant to say: ...Activity Streams 2.

Evan: increased IndieWeb development

<elf-pavlik> also some activity on github: https://github.com/jasnell/w3c-socialwg-activitystreams/issues

Evan: <Evan covered much more; see his slides>

<Shane> Hi all, I'm around on IRC but afraid I forgot my headphones so can't currently listen in at the moment

<AnnBassetti_> scribenick: AnnBassetti_

Evan: <continuing his introductory presentation>
... very diverse set of team members; many different points of focus
... also a challenge is how to interact with the IG
... tight schedule; fairly high stakes
... big opportunity
... Next steps:
... 1. Commit to strategies for API and protocol
... 2. Push Social API First Public Working Draft for Q4 2014
... 3. Keep federation protocol effort spun up for Q1 2015
... Opportunities in 2015 and 2016, to build upon the work of this WG
... questions? discussion?

Arnaud: in addition to adding specific items on REC track, we can also have a "wish list"
... the further we get into the Rec process, the more rigorous it is; hard to add new stuff; good to put those things on Wish List
... then work on those things when re-chartering subsequent work
... really helpful to have a wish list for future opportunities

Ed Krebs (Ford): reviewed notes from earlier work ...

scribe: we might want to add PubSubHub onto Wish List .. or IG to discuss

Tantek: gives kudos to Sandro Hawke as Indie developer

<harry> +1 sandro

<tantek> greetings #social!

<harry> note that folks can dial -in - also, is KevinMarks stopping by?

Larry Hawes: could you give us guidance on Use Cases needed for social data syntax?

scribe: since a lot of that work has been done

Evan: yes; a lot of preconception that we'd end up with Activity Streams
... not sure what value Use Cases would add at this point ... maybe as validation

<EdK> sidenote - most of my Use Case work tries to focus more on federation related issues, moreso than syntax/activity streams

James Snell: we don't have very clear use cases for <I missed which>

scribe: Need: what is a social profile in activity stream

Larry: yes, that's what we've focused on

James: is there a minimal set of social objects we need to provide

<jasnell> Ann... was saying that we need use cases and requirements for the additional vocabularies beyond basic activities

<jasnell> for instance, what is the vocabulary for a "social profile"

Arnaud: yes, we inherited this work, so are sort of coming up with solution before defining the problem ..

<jasnell> how do we describe people, and the objects acted upon, etc

Arnaud: but still useful to define use cases as documentation, validation,

Larry: and also 'gap analysis'

Lloyd; use case a few years ago re: describing pictures

scribe: what happened?

Evan: SWAT0

<jasnell> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SWAT0

<elf-pavlik> Wed session Social WG + Schema.org may have some relevance to broader vocabulary(ies) https://www.w3.org/wiki/TPAC2014/SessionIdeas#Schema.org_and_Social_WG

Evan: Social Web Acid Test

<EdK> SWAT Zero - first use case from federated summit

James: describes basic use case

<EdK> multiple actors who have accounts on multiple services

James: 6 basic steps

<jasnell> the link is : http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SWAT0

Lloyd: sounds like something the IG should include in our list

<jasnell> :-)

<tantek> Note that SWAT0 is a FOUR YEAR OLD use-case

Evan: there are a lot of other use cases that could be detailed out of that one

<tantek> see also: http://indiewebcamp.com/SWAT0

<dromasca> Evan: initiate and end subscriptions, approve, etc. - complex space

<elf-pavlik> who asked question?

Lloyd: do we have preference for business-oriented use cases vs more purely social use cases

<Shane> I like the way SWAT0 works, covers a lot of problems all at once without having a massive list

Arnaud: 2 diff constituencies here
... we should be equal; a lot of shared requirements

<dromasca> Arnaud: SWAT - the only uc that was voted about

<Loqi> dromasca meant to say: Arnaud:SWA - the only uc that was voted about

Evan: there are some refinements for the business use cases

James: would be interesting to describe what addtional requirements are need for business use cases

+1

Lloyd: might be like a road map

<dromasca> Ann: add 'approval' chain as business dimension of consumer use case

Ed: recruiting is a cross-over use case ..

<Zakim> tantek, you wanted to say, plenty of space on the wiki to document new use-cases as part of the social web IG16:53:18 <jasnell> Question: How many of these business use cases/requirements actually involve interop/federation across application or organization boundaries? Are these things that the WG needs to standardize?

Tantek: I'm not convinced there are such strong distinctions

<dromasca> tantek: pls. document social business ucs on wiki

<lehawes> I strongly second tantek's plea to document any use case on the wiki. Starting point is http://www.w3.org/wiki/Socialig/Use_Case_TF

Tantek: glad we've got Activity Streams out there as public draft .. and call for exclusions
... we need better documentation on use cases for Activity Streams .. both to add and to subtract components from AS
... esp important to document use cases for components of Activity Streams that are important for you
... to ensure it gets kept in

<elf-pavlik> +1 for backing all features with a use case requring it

Tantek: AS defined as nouns / verbs / objecgts
... in IndieWeb camp we discovered we only needed nouns

Ann: how did you get rid of verbs?

<tantek> AnnBassetti: you asked for documentation see: http://indiewebcamp.com/ActivityStreams#Verbs_vs_just_posts for example

Dan Romascanu: <sorry I can't hear>

scribe: asking about how WG approves or selects use cases

Arnaud: at the end, every feature in the spec must have minimum 2 implementations
... also, at the end, if there is a feature that no one wants to implement, that can be an important cue

<dromasca> question about how use cases are filtered and consolidated - Arnaud: WG will select a list of pointers to use cases from the IG

Mark Crawford: IG has in charter, reqt to deliver Use Case report ,,, NOT to support WG, but rather to describe social business .. which is broader than what WG focus is

scribe: may use diff formats

Larry: initial focus for IG use cases has been profiles, due to WG focus, but then we'll broaden into others

<break>

<elf-pavlik> thank you AnnBassetti_ for scribing!!!

<jasnell> For those not here in person, I'm going to be showing http://tpac.mybluemix.net/index.html on the screen here once we're back from the break

<elf-pavlik> thx jasnell !

<jasnell> (warning.. only tested in chrome and firefox ;-) ... sorry ie)

<harry> note we're on break now

<harry> will be restarting in 5 min.

<evanpro> scribenick: evanpro

<tantek> talky.io/socialwg is live - someone want to try it out?

jasnell will present overview of status of AS 2.0

Will talk about where we were with 1.0, what's changed with 2.0, what 2.0 looks like, what we need to do

http://tpac.mybluemix.net/index.html

Simple activity: actor, verb, object

<elf-pavlik> tantek, I get "We could not get access to your microphone or camera." and only 'Take me back' button

No i18n, no language context

But can get basic info across

flat namespace

2.0 activity

additional context

@id for id, some other info

In 1.0 MediaLink objects

url used inconsistently in 1.0

"url"

Never really clearly defined what it points to

<tantek> elf-pavlik: talky is two way so we can see who is watching

Limited set of properties for MediaLink: height, width, mediaType

AS 1.0 base schema with common verb and object type definitions

e.g. person, video

<tantek> greetings folks on Talky - could you introduce yourselves in IRC?

embedCode for HTML version of the item

<tantek> Talky users - how's the audio quality?

Use cases in base schema were either concrete or not very concrete

<elf-pavlik> tantek, ok-ish

Not everything took off as implementable

<elf-pavlik> i think better than SIP

<tantek> great!

<tantek> better than SIP is an improvement :)

geolocation defined in base schema 1.0 became core in 2.0

<elf-pavlik> thx tantek! :)

Priority e.g.

<tantek> np elf-pavlik

Most of base schema is not yet in AS 2.0 vocabulary

Many verbs from basic schema 1.0 are in use by processors today

Do we need standard identifiers for verbs?

<elf-pavlik> can i get link to your AS1.0 feeds? :)

4 people implement activity streams

2 consuming, 4 publishing

<tantek> AS 1.0 ^^^^

<tantek> 1 consuming implementation is only behind a firewall

List of implementers are on the wiki

Changes from 1.0

<tantek> for the record - at the meeting for consuming, only jasnell and evanpro raised their hands.

http://tpac.mybluemix.net/whatsnew.html

<Loqi> tantek meant to say: for the record - at the meeting forconsumin, only jasnell and evanpro raised their hands.

<harry> elf-pavlik, you want to dial back in?

<harry> anyone else?

<elf-pavlik> harry, i stick to talky for now :)

<harry> ok.

evan asks: some mechanisms for 1.0 to 2.0 compatibility around media type

tantek: asks no backward compatibility for XML version?

jasnell: no

XML version never made it to 1.0

Added elements: inReplyTo, replies

audience targetting: to, bto, cc, bcc

priority

location

tantek: to be fair, I'm consuming from many clients, see https://github.com/e14n/pump.io/wiki/Clients

<tantek> I'm only supporting publishing AS 1.0 / XML FWIW

<tantek> because it was easy to add it to my existing Atom feed

changed options, include "duration"

<tantek> have not done the work to try to publish any JSON AS

tags and attachments changed to singular

<wilkie> rstatus only supports AS 1.0 through XML too :/

2.0 supports language map

wilkie: I believe it's the same for GNU Social and Diaspora

<tantek> wilkie AS/XML solidarity!

<wilkie> yay!

<tantek> :)

height and width are pixels

upstream/downstreamDuplicates are deprecated

<wilkie> I am writing the JSON stuff though to catch up. :) the new backend code produces AS 1.0 json but doesn't consume/discover it just yet.

<elf-pavlik> I didn't hear question

<AnnBassetti_> Sandro asked how language here relates to language in json-ld

Sandro asks what the relationship is between language in AS 2.0 and JSON-LD language

<tantek> evanpro - could you minute question?

<elf-pavlik> https://github.com/jasnell/w3c-socialwg-activitystreams/issues/21

jasnell replies that there is additional language support in AS 2.0 since not all processors will support JSON-LD language

Example: unmapped verbs

With unmapped verbs like "post", these become skolems in JSON-LD

can add some verb namespace to avoid this problem

By adding context

Another example shows expanded actors or object

multiple actors and object

actors and objects referred to just by ID

object using the link class

"Link" class means that object is dereferenceable

and has a representation by the given media type

"multiple types"

Link (= dereferenceable) and "Blog Post" (= role)

Can use multiple vocabularies using JSON-LD

e.g. vCard

Actually call out "vcard:" prefix for non-JSON-LD processors

AS relies on other vocabularies

tantek asks where namespace expansion happens

jasnell says, it's part of the JSON-LD process, using the "@context" property

how do we use microformats ontologies in the JSON-LD

<AnnBassetti_> Lloyd asked if skills are represented .. .

<AnnBassetti_> answer: no

<tantek> note that the w3c vcard namespace is old vcard3. there is vcard4 now. and h-card

<tantek> which is based on vcard4

simple collections: using AS 1.0 similar links

<elf-pavlik> Q: do you use different URI for collecition and its first page?

geolocation using W3C geo basic

geolocation using GeoSPARQL is also possible

example of attachment, similar to email

Simple link: example 14

"rel" is issue for JSON-LD

<elf-pavlik> thx jasnell

jasnell responds to elf's question: yes, we do have different identifiers per page and for the collection

language example:

"language" defines default language context for all properties

<AnnBassetti> <Tantek gives brief overview to talky.io .. noting that we're seeing elf-pavlik, and he sees at least part of the room>

This doesn't come through to JSON-LD

Since "@language" only attaches to a property

Can also add "@context" to "@language"

<Loqi> AnnBassetti meant to say: <Tantek gives brief overview to talky.io .. noting that we're seeing elf-pavlik, and he sees at least part of the room>

Question: how easy do we want this to be for non-JSON-LD consumers?

AnnBassetti: logging bot for this channel

<tantek> wseltzer arrives

@context allows us to declare separate properties which can have default language

So "displayName_en"

AnnBassetti: yes

Need a better solution for language context

<elf-pavlik> could someone pleaes type questions to IRC?

Sandro points out that we need to have a review at Last Call for i18n

<elf-pavlik> thx :)

jasnell points out that there will be a11y issues here too

AnnBassetti points out naming conventions issue

e.g. family name and individual name

<AnnBassetti> See Richard Ishida's excellent article: http://www.w3.org/International/questions/qa-personal-names

dromasca asks accessibility, via language analysis

<AnnBassetti> he is lead of i18n activity in W3C

jasnell says much of the use case is JSON -> HTML

e.g. does a link need alt text

<AnnBassetti> i18n = "internationalization" ... 1st and last letters, with 18 letters in-between

<elf-pavlik> comment: possible issues with using as:alias to reference other resources (instead their @id) https://github.com/jasnell/w3c-socialwg-activitystreams/issues/25

Note on slides: http://tpac.mybluemix.net/ are up for review

todos: activitystrea.ms possible to use as ns:

Not a lot using the 2.0 namespace

Question around namespace URI: use dates or not?

Another question: backwards compatibility?

Use a namespace for 1.0 base schema?

<tantek> welcome barnabywalters - is that you we see on talky?

tantek asks, can we kick this question of NS to someone else?

<barnabywalters> greetings tantek!

<barnabywalters> yup, that’s me

evan asks, do we have authority to mint namespaces in /ns/

<tantek> barnabywalters joins via talky

sandro answers, and Arnaud confirms: yes

<elf-pavlik> hi barnabywalters o/

<barnabywalters> hi elf-pavlik!

sandro questions whether "/" or "#" should end the namespace

<harry> The hash versus slash debate is preventing the Semantic Web from creating the singularity, BTW

<harry> for 15 years!

One document with a hash, or multiple documents?

<elf-pavlik> +1 # :D

<wilkie> +0 :)

<harry> www.example.org/ns#term

<harry> www.example.org/ns/term

<sandro> http://example.com/activitystreams/verb

this would need to go to the JSON-LD context document per jasnell

<sandro> http://example.com/activitystreams#verb

<harry> +1

<jasnell> http://www.w3.org/ns/activitystreams#

PROPOSED: adoption of "http://www.w3.org/ns/activitystreams#" as namespace for Activity Streams 2.0

<jasnell> +1

<deiu> +1

<sandro> +1

<harry> +1

<MarkCrawford> +1

<rhiaro> +1

<elf-pavlik> +1

<dret> +1

<EdK> +1

<AdamB> +1

<dromasca> +1

<sandro> jtauber: +1

+1

<harry> any objections?

<wilkie> +1

RESOLUTION: adoption of "http://www.w3.org/ns/activitystreams#" as namespace for Activity Streams 2.0

closing this issue

jasnell continues, what to do with base schema 1.0

Does need to be heavily influenced by use cases

<AnnBassetti> Evan: as an Activity Streams implementer .. I can say that most of the verbs here are core to what happens in social software (follow, join, ...)

<AnnBassetti> ... others not so much

<AnnBassetti> ... questions if it makes sense to define a core schema, with others more tangential

<AnnBassetti> Tantek: to counter Evan's point, many implementations discovered they could make those verbs into nouns

jasnell says everything becomes a noun for IBM too

<AnnBassetti> James: in IBM's experience, that's true re: nouns

Very frist version did not have verbs

<tantek> AnnBassetti: to be clear *every* implementation done by someone in the IndieWebCamp community. not just "most"

<tantek> numerous at this point - dozens

verb becomes optional

fall back on object type

<AnnBassetti> Sandro: can you explain why you like verbs?

<AnnBassetti> Evan: more comfortable with activity referring to action / actor / object

<elf-pavlik> tantek, can you try reduce mic just a bit? it sounds link you have strong winds there ;)

<tantek> elf-pavlik: that might be the computer fan :/

<elf-pavlik> i don't think so, i just get thos FX after each word said

<barnabywalters> I think it’s just artifacts from a laptop mic being used to try to record an entire room

<tantek> barnabywalters: we're using the Logitech HD camera and mic

<tantek> I can reload and see if it helps

<barnabywalters> tantek: weird — compression artefacts maybe then?

<AnnBassetti> Ann: does this have to be decided now?

<AnnBassetti> Arnaud: no, but we should open an issue

<tantek> barnabywalters, elf-pavlik just reloaded

<barnabywalters> now there are silent clones

<AnnBassetti> James Tauber asks what was the advantage the IndieWeb folks got from this other syntax

<elf-pavlik> tantek, can we just try reducing mic sensitivity a little?

<AnnBassetti> ... this is largely a syntax issue, right?

<AnnBassetti> Tantek: good question

<AnnBassetti> ... we've been proceeding cautiously

<AnnBassetti> Sandro: is it 1- or 2- degrees of freedom?

Likes, favorites, RSVPs, favorites all work as nouns

<AnnBassetti> ... we need 2, if they are truly orthogonal

question of whether the representation is the event versus the state of the world after the fact

transmission is a notification mechanism

<elf-pavlik> Q: can we consider schema.org approach with "@type": "LikeAction" similar to your noun approach? ( jasnell might have already said someting about it earlier)

"verb": "like" vs. "@type": "LikeAction"

<Loqi> AnnBassetti meant to say: ... we need 2, if they are truly orthogonal

<Arnaud> ISSUE: Should we drop verbs and only use nouns/object type instead?

<trackbot> Created ISSUE-3 - Should we drop verbs and only use nouns/object type instead?. Please complete additional details at <http://www.w3.org/Social/track/issues/3/edit>.

<elf-pavlik> https://github.com/jasnell/w3c-socialwg-activitystreams/issues/23

<elf-pavlik> as:Verb & as:verb possibly not needed #23

<AdamB> CRUD

<sandro> CRUD = Create, Read, Update, Delete

evan asks if verb -> type has JSON-LD advantages

type and kind are different axes

<elf-pavlik> you can define properties with rdfs:domain and rdfs:range for subtypes

<elf-pavlik> see: http://schema.org/participant

<sandro> why not try to resolve the issue, if we have consensus

Implementation experience with indieweb has been that types can be implied

type can be implicit

objectType is not required for 1.0 as

some implementations use it, some don't

<harry> so as long as its not required and optional, we are OK if people end up using it implicitly

<elf-pavlik> having types might make it esasier to filter them

<tantek> sandro asked for a mapping from properties to implied types - here's what we have so far in IndieWebCamp: http://indiewebcamp.com/posts#Inferring_post_kinds_from_properties

sandro asks if we can skip the typing

or mapping properties of types

<sandro> sandro: can we have an official machine-readable mapping from properties to the types they imply?

generic property name for indieweb s "in reply to"

RSVP has its own properties

<AnnBassetti> compared to re-posts

<tantek> RSVP has its own property with its own values

<dret> maybe that's a good case for postel's law? be strict in what you produce and liberal in what you accept? implied info makes systems robust but also a bit vulnerable to misinterpretations, right?

<sandro> evanpro: I prefer explicit types to inferred types

Question for social API endpoint receiving an activity as a command

<AnnBassetti> Tantek: could we capture this as an issue?

<AnnBassetti> Arnaud: what, exactly, is the issue?

<AnnBassetti> <discussion>

<elf-pavlik> sandro, would you see IndieWeb implicit types similar to rdfs:domain inference?

Is the typing implicit in the presence of a property or explicit in a type or "verb" property

<sandro> elf-pavlik, yes, but I don't think it's helpful to get into that in this group

<elf-pavlik> roger!

<jtauber_> jtauber: jtauber: explicit types allow more generic properties (if properties can be used to infer types AND we have explicit types then the property names have redundancy)

<elf-pavlik> hey ahdinosaur :D

<ahdinosaur> hey elf :)

<barnabywalters> one interesting side effect I’ve observed and discussed of having implicit types via properties is that a single post can have multiple “types” e.g. be a like and a comment at the same time

<barnabywalters> there are several people in the IWC community who regularly publish these types of posts

<Arnaud> ISSUE: Do we rely on explicit typing or support implicit typing based on explicit property names?

<trackbot> Created ISSUE-4 - Do we rely on explicit typing or support implicit typing based on explicit property names?. Please complete additional details at <http://www.w3.org/Social/track/issues/4/edit>.

<elf-pavlik> "@type": ["LikeAction", "CommentAction"]

<elf-pavlik> barnabywalters, ^

<barnabywalters> elf-pavlik: that would probably be the AS equivalent — do any AS implementations actively publish or consume (i.e. display intelligently) activities like that?

<elf-pavlik> barnabywalters, AS1 didn't allow it

<elf-pavlik> http://tpac.mybluemix.net/whatsnew.html

harry points out that we can leave it and see what happens next

<barnabywalters> elf-pavlik: I don’t really care if it was *allowed*, I care who uses it ;)

Arnaud notes that it doesn't have to be one or the other

<Shane> barnabywalters++ Rules don't matter if people have a reason to ignore them

<Loqi> barnabywalters has 75 karma

Can use MUST, SHOULD, or MAY to switch between different options

<wilkie> people may want to use them because human beings are more creative than machines, and extensibility still matters

jasnell notes that we don't currently have a vocabulary of activities except 1.0 base schema

<barnabywalters> more specifically, the UIs which were used to create and display those activites

<wilkie> if people want to reduce social data to sludge because sludge is all protocol designers believe people need, more power to you :)

<tantek> wilkie - so far everything published by indiewebcamp implementations are *exactly* what *people* want to publish, not protocol designers

<tantek> exclusively so

<wilkie> less "here's what people do" and more "what can we do to handle people doing more without extensive changes to implementations?"

<wilkie> yeah, well, if your only benchmark is replicating a stream of tweets, then you'll get what you have, no kidding.

<tantek> wilkie - so far every different meaningful "type" has needed changes to implementations because its the UX of those implementations that matter in terms of different types

<tantek> wilkie - stream of tweets was IndieWeb 2010, we're well past that

<wilkie> no, if you don't understand a type, you can read the display name or content common to all types

<tantek> right, h-entry gives you that

<wilkie> then you filter by verb so that you only really try your best at reading 'post' activities

<tantek> "content common to all types" are the properties defined in h-entry

<wilkie> yes, there is a mapping

<tantek> so far implied-typing appears to the be things are evolving: http://indiewebcamp.com/posts#Inferring_post_kinds_from_properties

<wilkie> I really wish indie web cared about extensibility and interop instead of trying to tell everybody that they solved every problem because there is only one problem worth solving :P

<aaronpk> lol

sandro points out that subclassing

<wilkie> if only indie web were a community that wasn't so passive aggressive about how to do things right too

<wilkie> oh well. we'll get there in spite of it all, I'm fairly optimistic

<elf-pavlik> evanpro, can you please try to minute little more in IRC, i get pretty bad sound now :(

<tantek> wilkie - microformats2 is extensible - which is why the indiewebcamp community has been able to so easily add support for and implement new (implied) types

elf-pavlik: working on it

<tantek> wilkie - not sure what you mean about passive-aggressive - indiewebcamp is openly demanding of documentation of use-cases and selfdogfooding. even #social WG demands documentation of use-cases.

sandro points out that subclassing relationship ("love" is a subclass of "like") may benefit from using @type version propertyname ("likeOf", "loveOf") or verb "like" "love"

<elf-pavlik> tantek, wilkie: can we plesae try to stick to the converation in a room? i really have hard time to follow now :'(

<Shane> Also it isn't about doing things right, it is about learning from experience. Just look at how webmentions are changing with vouch etc

tantek says that users don't differentiate between types of content, e.g. image and text

so text with attachment versus image with caption

<sandro> tantek: Sometimes the users don't know what the "type" is of what they're doing, so maybe explicit typing would be hard

<tantek> remote folks, we are breaking for lunch but will leave the camera on

<tantek> on talky

<elf-pavlik> tantek, can we try adjusting mic gain a bit

<lehawes> quit

<oshepherd_> WebRTC: Brand new and more reliable than SIP. The world we live in.

<rektide> i wish talky would let me in in spectator mode. i have no mic nor video on this laptop. "You are about to join a video chat" modal dialog on screen with "We could not get access to your microphone or camera"

<wilkie> same

<gRegor> I could hear you, barely, elf-pavlik

<gRegor> That's a bit better

<gRegor> very good, elf-pavlik

<gRegor> elf-pavlik: mic quality is better

<elf-pavlik> jasnell, could you please ask tantek when he comes back to reduce mic gain? we get a lot of artefacts in the sound :(

<jasnell> elf: will do

<elf-pavlik> thx :)

<Arnaud> still waiting for people to be back in the room

<EdK> some folks still returning from lunch, should be able to get started in about 5 minutes ((1;10pm Pacific)

Arnaud: http://evan.prodromou.name/files/TPAC-API/

<lehawes> Back from lunch :>)

<jasnell> Elf, tantek is not back yet. once he is I'll ask him about the mic gain

<Arnaud> about to start

<jasnell> hey Elf.. what time was that breakout session for wednesday supposed to be? do you remember?

<elf-pavlik> 9:15AM

<rektide> there's no alternative to the talky (which those of us without webconf broadcast gear can't use) ?

<elf-pavlik> rektide, we used Zakim before but talky had better sound

<elf-pavlik> which now doesn't work that well any more...

<lehawes> Arnaud: Do we know what time the IG presentation will be tomorrow morning?

<gRegor> good, tantek

<gRegor> audio died

<harry> scribenick: harry

<gRegor> audio's back

Social API

evanpro: Presenting slides and then open discussion

<gRegor> I'm the other person, lurking :)

evanpro: this is next in deliverable list

<AnnBassetti> Evan presents slides re: Social API

<AnnBassetti> http://evan.prodromou.name/files/TPAC-API/

<scribe> scribenick: Ann Bassetti

<harry> scribenick: AnnBassetti

<Lloyd_Fassett> I can take it

<Lloyd_Fassett> Issues" Cpmf;atopm pf AS amd E,bedded experiences /

<elf-pavlik> thx Lloyd_Fassett ! sometimes we get poor audio so minutes *help us a lot* to follow

<Lloyd_Fassett> Examples of Social API: Twitter / Facebook / OpenSocial

<Lloyd_Fassett> elf - can you see the slides?

<elf-pavlik> yes, i have them on second screen

<Lloyd_Fassett> If we can't make something that has a passing resemblance to these API's, we're on the wrong track.

<jasnell> If it's helpful to folks, I can give a quick overview of how the OpenSocial spec is organized

<Lloyd_Fassett> We need reusable libraries, you should be able to point a social client at a new service and it should just work

<Lloyd_Fassett> It should be possible be to implement an API facade

<Lloyd_Fassett> Social API's manage social connections, publish content, manage responses..a light application platform

<elf-pavlik> tantek, if you have mic settings to reduce gain / boos just 10% or so it could reduce artifacs in sound ...

<Lloyd_Fassett> There's not just a one way follow, but a two way...join a group (manage circles), manage list of friends...manage a group of other accounts can be very useful

<Lloyd_Fassett> Larry Dawes: Does it need to be two way?

<Lloyd_Fassett> Evan: Twitter is one way, Facebook is two way, G+ is either

<Lloyd_Fassett> Social content publication management CRUD...should be able to handle basic types...not all services would handle all

<Lloyd_Fassett> Responses: typically there is a way to manage responses: Like, vote, comment, reply, share..redistributing content to your followers

<Lloyd_Fassett> Social Applications: event invitations, polls, Q&A, Games, etc.

<elf-pavlik> AFAIK both IBM and Boing have Q&A systems

<Lloyd_Fassett> Models: 1. Command, 2. Resource.

<AdamB> Boeing does have an Q&A system

<Lloyd_Fassett> Resource model is more common

<Lloyd_Fassett> represent URI's, use http to modify.

<Lloyd_Fassett> Example: post 'tantek' to a stream of information

<oshepherd> Important thing with resource model, syndicating activities is harder

<Lloyd_Fassett> You can have a uniform interface with this...get an endpoint.

<Lloyd_Fassett> downside is 'unfollowing' and other actions are harder to do

<Lloyd_Fassett> other models will take us off the beaten path of social software

<Lloyd_Fassett> non-JSON representation and non-RESTful are off the beaten path

<Lloyd_Fassett> Proposals under discussion are Open Social - very command oreinted, http verbs against an AS

<Lloyd_Fassett> use Put to update elements in that stream

<Lloyd_Fassett> very nice straight forward API

<Lloyd_Fassett> We need to make some changes from current state of it.

<Lloyd_Fassett> Pump.io (Evan works on this) is a hybrid of command and resource.

<Lloyd_Fassett> I think we need more proposals, or start working on them internally

<tantek> bunch more on https://www.w3.org/wiki/Socialwg/Social_API_candidates

<Lloyd_Fassett> we need endpoints

<Lloyd_Fassett> Tantek suggested micropub as an option to consider

<Lloyd_Fassett> We need to make a number of decions, we need to get on the path to resolving them.

<tantek> with existing multiple client and server micropub implementations

<Lloyd_Fassett> first is understanding embedded experience

<oshepherd> tantek: micropub is publish-only, no?

<Lloyd_Fassett> Embedded experiences is a nice system that works with Open Social, but we'll need a lot of work to get it going

<tantek> oshepherd: not sure what you mean? people are using it for CRUD for various post types - which is the core of social API as described

<oshepherd> tantek: Micropub is a publishing API, doesn't let you fetch /your/ incoming stream

<jasnell> Our actions proposal addressees the embedded experiences stuff, separates it from the api

<Lloyd_Fassett> How deep does social API go in a application?

<tantek> oshepherd: you mean HTTP GET ?

<Lloyd_Fassett> What verbs and domain concepts should be supported, how extensibility is needed?

<Lloyd_Fassett> Other important issues that we may want to punt on are authentication, discovery (will you be able to take an identified user and discover endpoints / AS etc.)

<oshepherd> tantek: What I get with GET http://mypumpserver.com/api/username/inbox, i.e. the activities addressed at me and of everyone I'm following

<Lloyd_Fassett> Arnaud: Thank you Evan

<Zakim> elf-pavlik, you wanted to discuss Action Handlers / Hypermedia Controls / Hypermedia API

<elf-pavlik> eg. http://www.hydra-cg.com/spec/latest/triple-pattern-fragments/#controls

<elf-pavlik> "This hypermedia control must be present because there purposely does not exist a fixed URL format that servers of triple pattern fragments need to follow. This means that clients of triple pattern fragments must not need prior knowledge of a server, i.e., they must not assume a certain URL pattern. Instead, clients must interpret the hypermedia control in each triple pattern fragment in order to retrieve another fragment."

<elf-pavlik> http://lists.w3.org/Archives/Public/public-socialweb/2014Sep/0044.html

<elf-pavlik> [Lecture] REST, Hypermedia, and the Semantic Gap: Why "RMM Level-3 REST" is not enough

<tantek> no existing hydra implementations for the record

<Lloyd_Fassett> Tantek: do you know implementation of that? Elf: no

<Lloyd_Fassett> (I can't tell what Elf is saying)

<Lloyd_Fassett> Evan: that's wildly unlike any Social API's I talked about at the beginning

<Lloyd_Fassett> those typically have well defined URL's. I don't know what the starting point would be if we don't give structure to URLs.

<Lloyd_Fassett> Elf: you can sense it in the header.

<Lloyd_Fassett> Evan: how?

<Lloyd_Fassett> (I didn't understand answer)

<Lloyd_Fassett> James: when he said page collection, you're going to have a number of links...you need a template...if the API is a follow your nose

<Lloyd_Fassett> every response includes information about the object

<Lloyd_Fassett> Evan: If I were creating a client, I'd ask for an entry point and then there would be a discovery process. Is that right

<Lloyd_Fassett> Elf: yes

<Lloyd_Fassett> James: This is the usecase for Webfinger

<Lloyd_Fassett> Evan: link relations, on page or in header

<Lloyd_Fassett> Eric: I am from (didn't hear)

Siemens

(I think)

<Lloyd_Fassett> Evan when you say API is it http?

<Lloyd_Fassett> Evan: yes

<tantek> re: Webfinger - it conflates user URLs vs API URLs which I think is a problem. we've found it better to use h-card for user URL discovery, and <link> etc. HTTP LINK header for API URL discovery.

<Lloyd_Fassett> Evan: we have a lose definition in our charter. We have a lot of pressure to do http restful api, if we don't, we have to explain why

<elf-pavlik> Hypermedia API = full REST API

<Lloyd_Fassett> Eric: since AS are http based, how is Social API different? HOw do they relate?

<tantek> FWIW - we have found that the "static site generation" use-case (which numerous people do), makes REST impossible, so we have dumped strict REST

<Lloyd_Fassett> AS is a data format

<oshepherd> tantek: I think I've mentioned this already, but in the draft I'm going to bring shortly, WebFinger is used *only* for mapping acct:username@domain URIs into dereferencable objects and nothing else :-)

<Lloyd_Fassett> Eric: with links in it

<Lloyd_Fassett> Eric: I think a RESTful api with AS where things are posted/put is close to what is in our charter. Does that make sense?

<tantek> oshepherd: in that case, we have in the simple case that you can skip webfinger and just use the user's URL to get their representative h-card and go from there

<tantek> there's no actual need for webfinger

<elf-pavlik> http://martinfowler.com/articles/richardsonMaturityModel.html

<Lloyd_Fassett> Eric: You need a relative mapping to explain how other social sites show up

<oshepherd> tantek: Backwards compatibility with user@domain URIs?

<tantek> such folks have profile URLs - done

<Lloyd_Fassett> Evan: That's not where I was going with that. There are a number of API's...http based mostly using JSON that have patterns so you can see your followers, your activity

<Lloyd_Fassett> Eric: the goal is not to have a Social API....

<Lloyd_Fassett> Evan: I think there is a great possibility of doing that....the API could talk to Facebook etc on the back end.

<Lloyd_Fassett> Evan: we'd have a social network server to handle that.

<wilkie> I've found that people will debate whether or not an '@' or a '/' is better until the sun cools

<Lloyd_Fassett> Tanktek: A social proxy server

<tantek> wilkie - turns out people that actually post on the web have a URL for their profile

<tantek> so if they're debating it on the web, they're already using a '/'

<tantek> we can judge by actions rather than words

<wilkie> they're just URIs that map identities. webfinger's a little more indirect, so it might be better for service migration, but meh. whooooo caaaaarrrreeees. it's not like webfinger is HARD or anything.

<Lloyd_Fassett> Arnaud: adding use cases for this would be useful. We all come with different backgrounds. IT's important to explain where you are coming from so people know where you are coming from.

<tantek> hard or not, webfinger is unnecessary. YAGNI = people don't bother implementing at scale.

<Lloyd_Fassett> Arnaud: Even terminology is an issue

<wilkie> tantek: we can say 'turns out' this and that and point to our userbases allllll day. we'll both be correct. we have seen all of these things. that's why they exist.

<oshepherd> tantek: Its' 1 trivial web request if the user who you're attempting to follow has a user@domain URI. Boohoo

<tantek> a lot of what's been done in indiewebcamp protocols / formats is drop a lot of "1 trivial X"

<elf-pavlik> folks, let's leave webfinger in peace for today, it makes it really hard to follow if we have 2 parallel conversations :(

<Lloyd_Fassett> Harry: I think we have time for James Snell to present to talk about embedded experience

<Loqi> aww, cheer up

<tantek> dropping a lot of 1 trivial x type stuff = more publishers and more implementations

<jasnell> Can definitely give an embedded experience overview

<tantek> if you like webfinger, great, implement it on your own site

<Lloyd_Fassett> Harry: Authentication is out of scope. It is being recharged in another WG. We expect new charters by early 2015.

<tantek> I'm just pointing out trends towards minimalism

<Lloyd_Fassett> Harry: What we can use today, longer term thinking, assign JSON payloads, if we need to do digital signatures...do it the ITF way, The CRG...if you're into digital signatures...don't worry there will be ways to do that are supported by browsers.

<Lloyd_Fassett> Harry: public test suites by the end of Dec.

<Lloyd_Fassett> Harry: Tanket IETF: SSL issue brings process into question

<elf-pavlik> Lloyd_Fassett, thanks for scribing!!! with current audio i can only read...

<elf-pavlik> so all in Lloyd_Fassett hands (literaly) :)

<Lloyd_Fassett> Harry: rolling your own crypto and signatures is a bad idea. For discovery it's the same thing. Things that have been tried include Host meta, webfinger, Mike Jone's version. Look at IETF for the work they've done

<dret> speak up, everybody!

<dromasca> https://datatracker.ietf.org/doc/charter-ietf-anima/

<Lloyd_Fassett> James: when we were talking about verbs and syntatics...events vs states...is that the same as commands vs resources?

<dromasca> charter on the iesg table for approval as we speal

<dromasca> speak

<elf-pavlik> dromasca, link?

<Lloyd_Fassett> Evan: answer: no. Who has looked at Open Social API? (3 hands). It has a single place where you expect Activities, like ATOM. The command model is to fire new activities at that end point.

<Lloyd_Fassett> James: they could represent either?

<Lloyd_Fassett> Evan: yes

<Lloyd_Fassett> Evan: the idea of a resources model where you get a collection of objects, if I was going to get that same information from AS I'd go through the stream

<Lloyd_Fassett> James: The handling of deletion...it's a version of a state...you'd need to know all the events.

<Lloyd_Fassett> Evan: Yes, it's a hard way to manage that mechanism

<Lloyd_Fassett> Evan: One nice thing about that structure is that it's simple and we can publish quickly, but we'd be kicking the deeper issue into the future

<Lloyd_Fassett> James: It's basically the ATOM model. While we can get so far with that approach, this is signicantly better

<Lloyd_Fassett> James: you could provide both current state and audit streams

<Lloyd_Fassett> EdKrebs: in business lense you need some persistence. In current public state you don't.

<Zakim> tantek, you wanted to go through the slides and make comments, slide by slide

<Lloyd_Fassett> James: this is good for use case work

<jasnell> Lloyd... There are two james' :) James s. And James t.

<Lloyd_Fassett> thank you.

<jtauber> jasnell: we can't even say the left-handed one ;-)

<elf-pavlik> which page?

<Lloyd_Fassett> Tantek: embedding is a different beast than CRUD.

<oshepherd> tantek+1 on this

<dret> for an interesting approach how to persist events such as deletion, https://tools.ietf.org/html/rfc6721 (authored by james, btw) might be interesting to look at.

<Lloyd_Fassett> Evan: one thing about embedded experience with embedded streams is perscibed by the activity generator.

<Lloyd_Fassett> Tantek: I thought managing connections vs CRUD interesting...push semantics down into the format.

<Lloyd_Fassett> Jasnell: that gets back to the Resource model

<elf-pavlik> evanpro, could you type into IRC nr or title of discussed page?

<evanpro> http://evan.prodromou.name/files/TPAC-API/

<evanpro> But I'm not scribing right now

<oshepherd> tantek: Audio/video posting from browser... easy if you're just uploading premade files...

<elf-pavlik> evanpro, but you control slides in a room? we just need page nr to follow

<Lloyd_Fassett> Tantek: Video and Audio are hard. Mobile browser support is not working. If we can do Text and Image, Vid and Aud are versions of the first two.

<oshepherd> tantek: But with AS1 audio/video is no more complicated than images in th API

<Lloyd_Fassett> Evan: In a business context, Word Processing, Spreadsheets, would be common

<Lloyd_Fassett> Tanktek: pointing to vid conference call "this is real different"

<oshepherd> Punt realtime off into future/IG?

<oshepherd> Realtime is kind of an "event"

elf demonstrates the upside-down use case

<dret> RTC could be mapped into a different URI-based interaction space, in the same way as tel: URIs allow to provide links into the telephone interaction space.

<Lloyd_Fassett> Tantek: Other models are difficult non-JSON....the indie camp..using micropub, you can post type of object and number of properties and that's it. If you want to post something with structure the Micropub models requires you post them one at a time.

<Lloyd_Fassett> Lloyd: power went out on projector

<Lloyd_Fassett> Evan: Let's take a break

<Arnaud> http://evan.prodromou.name/files/TPAC-API/?full#19

<Lloyd_Fassett> Tanket: Non-json point is that we're not even using html

<Lloyd_Fassett> webmention is a versino of ping back

<Lloyd_Fassett> we used that with Micropub as well

<Lloyd_Fassett> Tantek: In a LInked Data way we post them one at a time we post each item to individual URLs and reference those

<Lloyd_Fassett> we haven't run into limitations

<Lloyd_Fassett> Arnaud: that's expensive for systems.

<Lloyd_Fassett> Tantek: not really

<Lloyd_Fassett> Tantek: Non-REST. Everyone in INdie world is from REST

<Lloyd_Fassett> STatic site generators are a show stopper for REST

<Lloyd_Fassett> You have to use a common end point

<elf-pavlik> also CDN ...

<Lloyd_Fassett> Tantek: This is a solid use case. Even the New York Times generates a static site that creates this problem. They are big enough to be considered Enterprise class. THis is an important market signal we need to look at

<Lloyd_Fassett> Tantek: we need to support that

<Lloyd_Fassett> Tantek: we make micropub post calls to a published endpoint.

<Lloyd_Fassett> Jasnell: The dataformaty for the client going back to the server is not an important as a shared vocabulary. Making the terms understandable is important.

<elf-pavlik> +1 shared vocab

<Lloyd_Fassett> Jasnell: we have lots of evidence for both, we need to match up vocabulary

<Lloyd_Fassett> Evan: We need to specify something.

<tantek> FYI micropub: https://indiewebcamp.com/Micropub

<Lloyd_Fassett> Jasnell: JSON-LD vs form POST are the two options

<Arnaud> http://evan.prodromou.name/files/TPAC-API/?full#20

<ahdinosaur> how are SSGs a show stopper for REST?

<Lloyd_Fassett> Tantek: there are mulitple micopub implementations now.

<tantek> http://aaronparecki.com/metrics

<Lloyd_Fassett> Tantek: people are creating multiple clients...a person created a Pebble app to publish what he's eating as an example.

<Lloyd_Fassett> Tantek: reviewed what the person posted to the group. There's a lot of implementation work there.

<Arnaud> http://evan.prodromou.name/files/TPAC-API/?full#21

<Lloyd_Fassett> Evan: comments on decision slide?

<Lloyd_Fassett> Tantek: Strongly wants to separate embedded experience. It's important, but a different problem.

<Lloyd_Fassett> Evan: Is there a procedure we can take that embedded experience as a next step after a client API? It is on our charter.

<Lloyd_Fassett> Harry: The charter is vague on purpose. It would be sad if we didn't deliver an API

<tantek> welcome aaronpk to the talky!

<Lloyd_Fassett> Arnaued: like AS, we are free to split the purpose

<Lloyd_Fassett> Tantek: how many people are interested in Embedded experience? (ans: about 5)

<Lloyd_Fassett> Tantek: Parralell to CRUD API

<oshepherd> aaronpk: Elf is upside down. I'm the one with the beard. Hopefully the one with the people is obvious :-)

<aaronpk> what is a summary of "embedded experience"? (sorry I joined late)

<Lloyd_Fassett> Jasnelss: I see actions as separate from the API work.

<tantek> and we lost power in the room again

<Lloyd_Fassett> Evan to James: Are you embedding actions into a UI that would respond?

<Lloyd_Fassett> Jasnell: yes

<Lloyd_Fassett> Evan: That's different than Embedded Experience?

<Lloyd_Fassett> Jasnell: yes

<Lloyd_Fassett> Evan: I would be more supportive of that than something that got to DOM level.

<Lloyd_Fassett> Evan: The Open Social API, you can fire anything at that end point, it doesn't tell you anything about followers of where you post or anything else. It's an activity lifecycle

<Lloyd_Fassett> Evan: We can do that, or we can support verbs (following, voting, etc.)

<Lloyd_Fassett> Tantek: I"m leaning toward including all the verbs in the format and not the API

<oshepherd> tantek++ on in the format. Want to syndicate "follow" to followers, for example :-)

<Loqi> tantek has 112 karma

<Lloyd_Fassett> TanteK what is room for extensibility?

<Lloyd_Fassett> Evan: Can we support new verbs in the AS sense. How should a server handle activities it doesn't understand?

<Lloyd_Fassett> Jasnell: We're using AS for in devops..checkin, checkout for code, it's not something a typical social system would see. How would a system react if it saw that and didn't understand it or be gerenic on CRUD? That's what we need to figure out

<Lloyd_Fassett> Evan: If the extensibiliy is in the payload, we can say go ahead and put in whatever you want.

<Lloyd_Fassett> Evan: Tweets deliver exactly just tweets.

<Lloyd_Fassett> Tantek: Twitter changed the payload for pictures

<oshepherd> tantek: Question of does server /act/ on your activities

<oshepherd> e.g. does posting a "follow" activity make the server do something?

<oshepherd> If not, need separate "follow" endpoint

<Lloyd_Fassett> Jasnell: Twitter defines the payload and I can only supply that. The alternative is that the endpoint is generic and I decide what is in the payload.

<tantek> oshepherd: I suppose, it depends?

<oshepherd> tantek: Probably the pragmatic option

<Lloyd_Fassett> jasnell: IBM Connections you can submit many things to the API, but it might not surface in the UI

<tantek> oshepherd: typical micropub scenario: publish a post of some time, your server sends webmentions to everything the post links to, and then *those* servers react in some way.

<Lloyd_Fassett> Jasnell: does an endpoint need to accept anything?

<oshepherd> jasnell: IMO should. Extensibility dictates so.

<Lloyd_Fassett> AnneB: What's an endpoint? Tantek: An endpoint accepts http post requests. It can be the same server can publish information. STatic sites use proxies to handle requests.

<Zakim> sandro, you wanted to point out that API in the charter is not this

<Lloyd_Fassett> Sandro: http api's are not what the charter says. The charter language use "client side"

<Lloyd_Fassett> Harry: You can read "client side" in both ways.

<elf-pavlik> client (app) - server (web service) ?

<Lloyd_Fassett> Harry: it was written to mention html 5, many companies requested embedded experience, that's why it's listed in the deiverables. We are not bound to deliver an API thought for Open Social.

<Lloyd_Fassett> Harry: recharting at this point is a waste of time.

<Lloyd_Fassett> Harry: embedded can be just retrieving AS and it can be thought of as embedded.

<Lloyd_Fassett> Tantek: I think we're streching the charter.

<Lloyd_Fassett> Sandro: I'm concerned we don't get patent coverage

<Lloyd_Fassett> Harry: we'll get it because the licensing comes in at the end. The initial comits to the drafts. At the workshop everyone was asking for embedded experiences, not CRUD, but they are not here.

<tantek> aaronpk, you may want to read scrollback about micropub

<Loqi> tantek meant to say: aaronpk, you may want to read scrollback aboutmicropu

<Lloyd_Fassett> jasnell: there are lots of API's, Open Social implementations focuses on their gadgets, not the API. Let's get a good model for embedded experiences and let people create it on their side.

<Lloyd_Fassett> Harry: OpenSocial has both

<harry> Note that the OpenSocial Submission mixes CRUD and non-CRUD: http://www.w3.org/Submission/2014/SUBM-osapi-20140314/#rfc.section.2.3

<Lloyd_Fassett> Tantek: Is anything in OpenSocial in the charter?

<Lloyd_Fassett> Evan: no. It was contribution to the process

<harry> The term 'client-side' was used on purpose

<Lloyd_Fassett> Arnaud: I agree with Harry that we're making a mountain out of a mole hill.

<Lloyd_Fassett> Arnaud: Can we get that in writing?

<harry> We can interpret the term 'client-side' API to include both a HTTP CRUD API and Embedded Experiences.

<harry> In particular, both are in the member submission.

<Lloyd_Fassett> Anne: I think Sandro is correct. I don't follow the nuance of the discussion. But I think it's important to document the consensus of the group.

<Lloyd_Fassett> Eric; I would read the charter as Client Side

<Lloyd_Fassett> Evan: only?

<Lloyd_Fassett> Eric: yes

<Lloyd_Fassett> Eric: Sandro has a good point

<sandro> We seem to agree to that "Social API" includes HTTP APIs, not just embedded experiences APs

<Lloyd_Fassett> Harry: We don't want to put embedded experiences out of scope yet. We need to give members a chance to respond. The WG can deliver them in separate documents.

<Lloyd_Fassett> Evan: It sounds like there would be discomfort on an http API

<Lloyd_Fassett> Tantek: no

<sandro> PROPOSED: This Working Group interprets "Social API" in include HTTP API (and maybe other things)

<sandro> +1

<evanpro> +1

<aaronpk> "to include"?

<Lloyd_Fassett> +1

<oshepherd> +1

<jasnell> +1

<MarkCrawford> +1

<dromasca> +1

<rhiaro> +1

<elf-pavlik> +1

<jtauber> +1

<lehawes> +1

<harry> +1

<tantek> +1

<Loqi> AnnBassetti meant to say: AnnBassetti joined #social

<AdamB> +1

<sandro> RESOLVED: This Working Group interprets "Social API" to include HTTP API (and maybe other things)

<claudio> +1

<Loqi> harry meant to say: +1

<Lloyd_Fassett> jasnell: the charter says embedded format but nothing about CRUD. Should we include that?

<Lloyd_Fassett> Sandro: it's too early for that.

<dret> +1

<Zakim> sandro2, you wanted to ask whether anyone has a solid proposal for a Resource-style API (because I'm somewhat skeptical that all activities map to this style)

<Lloyd_Fassett> Sandro: resource style...I like the feel of the Resource better, but can you do that for every activity type? I'm dubious

<Lloyd_Fassett> Evan: I agree, it gets complicated. We may be able to do something that handles mainstream actions. We could extend it with command structure

<Lloyd_Fassett> AdamB: I agree with that.

<Lloyd_Fassett> Sandro: ok

<Lloyd_Fassett> Evan: question about authentication. I agree with Harry. Authentication is already well covered. Most are using OAuth 1.0 or 2.0. There are methods people expect

<Lloyd_Fassett> Jasnell: Oath 2.0 has a notion of scopes. If there are specific scopes for Social, we may want to define what those are.

<Lloyd_Fassett> Harry: W3C does lias with IETF. I prefer not to reinvent OAuth.

<Lloyd_Fassett> AnnB: We use SAML

<aaronpk> building on top of OAuth 2.0 is a good idea

<harry> http://oauth.net/2/

<aaronpk> defining scopes would be useful

<oshepherd> oauth2++

<Loqi> oauth2 has 1 karma

<Lloyd_Fassett> EdKrebs: We use SAML internally, but with collaboration across partners OATH 2.0 is becoming more important

<oshepherd> need to think about extending it to cover the federated use case

<aaronpk> also note that obtaining authorization is separate from using authorization

<oshepherd> My phone client wants to be able to identify itself to yoursite.com as mysite.com/me

<tantek> aaronpk - I wonder what it would take to plug SAML into IndieAuth

<aaronpk> OAuth 2.0 is a way to get authorization, which is represented by an access token string

<harry> We regularly co-ordinate with both IETF re OAuth and OASIS re SAML

<harry> so asking for stuff and having questions should be straightforward.

<aaronpk> then clients present the access token to the server

<Lloyd_Fassett> Harry: We can reconnect with SAML if needed. We can use both with W3C. W3C monitors developments.

<aaronpk> IndieAuth is built on top of OAuth 2.0 where identifiers are URLs, but the end result is also an access token

<oshepherd> SAML is XML based; OAuth2 is JSON based and more widely deployed on the open web. Related tech, e.g. JSON Web (encryption/signing/token) also aligns well with our use case

<Lloyd_Fassett> EdKrebs: Social Applications, idea management. 3 years ago it was horizontal, 2 years ago they crossed processes (prod dev). Now I see Idea Management. It's largely around enterprise activities. It might be somehting the WG would want to prioritize back to the IG.

<aaronpk> this means it's possible to have multiple (incompatible) ways of obtaining access tokens, and as long as the "social API" can accept a bearer token we can deal with authorization as a separate issue

<harry> yes, you will need "an access token"

<Lloyd_Fassett> AnnB: I don't know if it's that different. IBM Jam did that.

<harry> In consumer-facing space, likely OAuth

<Lloyd_Fassett> EdKrebs: it's more than that.

<harry> In enterprise, likely SAML

<Lloyd_Fassett> AdamB: lifecycle of ideas?

<harry> Regardless, we don't reinvent either.

<oshepherd> Note OAuth is picking up adoption in enterprise space

<aaronpk> tantek: I think it wouldn't be a huge stretch to use SAML in an IndieAuth context

<harry> We like both communities and work with both regularly, in particular with help of ISOC

<Lloyd_Fassett> EdKrebs: there are interesting nuances that are new. We've had a lot of vendor demo's in the last few months that sparking some ideas.

<elf-pavlik> aaronpk, any special conditions you can think of with CORS enabled?

<Lloyd_Fassett> Jasnell: Provinance is the piece that is being added. Manage it and track it

<aaronpk> elf-pavlik: I'd have to go brush up on SAML before I can answer that

<Lloyd_Fassett> Tantek: Provenance is important in indie web. We need it to establish trust. Everything is so distributed.

<elf-pavlik> i meant fo oauth, i know that persona has some issues there since server needs to limit Audience for security reasons

<Lloyd_Fassett> Harry: We seen failed experiments with digital signatures but I expect it to come back.

<Loqi> elf-pavlik meant to say: i meant fo oauth, i know thatperson has some issues there since server needs to limit Audience for security reasons

<Lloyd_Fassett> AnneB: Boeing does not allow anonymous posting. Identity is controlled.

<oshepherd> tantek: Suggest something JSON Web Token based. Use PKI to establish chain of identity from "authority" (Server controling URI)

<Lloyd_Fassett> Claudio: responding to TAntek from an hour ago. I think video conferencing is important.

<Lloyd_Fassett> Tantek: we're using peer to peer in this meeting

<Lloyd_Fassett> Claudio: webrtc?

<Lloyd_Fassett> Tantek: yes

<aaronpk> i'm using Chrome btw

<Lloyd_Fassett> Tantek Chrome Firefox

<aaronpk> for talky

<aaronpk> &yet wrote an iOS app so I can join this talky room from my phone

<Lloyd_Fassett> Tantek: we'd love to see rtc in IOS

<aaronpk> would be nice to see webrtc support in iOS safari though

<elf-pavlik> aaronpk, Mozilla Persona limitation which limits its usage with CORS (Explicitly specify the audience parameter) https://developer.mozilla.org/en-US/Persona/Security_Considerations

<Lloyd_Fassett> Larry: an extension to Sandro's point. There's a problem now for lack of activity definitions. Liking isn't the same everywhere.

<tantek> Larry, see http://indiewebcamp.com/like and http://indiewebcamp.com/favorite and in particular http://indiewebcamp.com/like#Variants

<Lloyd_Fassett> Evan: That's reasonsonable. In the AS world we have a schema that is super fuzzy. Especially as words in English like Attach, Consume (eat or watch?)

<Lloyd_Fassett> Activy Schema has to do this

<Lloyd_Fassett> jasnell: throw it up at the wall and document. We need to narrow it down to standardize it.

<Lloyd_Fassett> Eric: If you need something we precise semantics, do your own vocabulary and publish it. We won't see that in our base Schema

<Lloyd_Fassett> Evan: our core set should cover many things

<Lloyd_Fassett> Tantek: 500px has both a heart button and a like button. That pattern is not going to end. Some places have a lot of smily's.

<MarkCrawford> Since I have heard the dreaded IP words raised, I think it might be worthwhile to remind folks in the meeting that only the WG members have made the IP commitment because if I remember correctly, the IG members have no licensing commitment. Also, since we are taking resolutions of the WG, perhaps we should limit the actual polls to WG members only.

<Lloyd_Fassett> jasnell: 500px 'heart' adds to your list, like ads a count of likes

<lehawes> tantek: Thanks! I'll check these out.

<oshepherd> AS1 conflates "favorite" and "like". Perhaps we should disentagle them?

<oshepherd> "favorite" would be a more long term thing, probably have a list of your favorites

<oshepherd> "heart button" would map to favorite

<evanpro> oshepherd: yep, that's pretty complicated

<Lloyd_Fassett> MarkCrawford: IP...we have IG members that have not made WG comittments. We need to be care about taking their contributions. We are taking +1 in the minutes from non WG members.

<Lloyd_Fassett> Arnaud: Fair point, but the outcome would not be different.

<jtauber> FWIW: here's my collection of things that a "favorite" in Twitter can mean: https://thoughtstreams.io/jtauber/taxonomy-of-twitter-favorites/

<Zakim> tantek, you wanted to say Micropub uses IndieAuth which is pluggable, uses OAuth, Persona, SMS, TOTP etc.

<Lloyd_Fassett> MarkC: we need to be careful because IG members are not making a commitment

<dret> i'd be in favor of not boiling the ocean of formalizing/cataloging/mapping all possible interaction models on social sites.

<Lloyd_Fassett> Harry: Good point

<Lloyd_Fassett> Tantek: Any API that has write to a server will need Authentication. I recommend Indieauth (?).

<Loqi> evanpro meant to say: oshepherd: yep, that's pretty complicated

<harry> IndieAuth via SAML may be a world of pain :)

<Lloyd_Fassett> Tankek: I asked someone to look at SAML through Indieauth

<aaronpk> sounds like fun :)

<harry> It's got a very nice metadata framework, but just warning you guys

<tantek> thanks aaronpk!

<jasnell> Anything with SAML is a world of pain

<Arnaud> PROPOSED: Separate Embedded Experiences from "basic" client API

<oshepherd> +1

<tantek> +like

<tantek> ^^^like

<Lloyd_Fassett> Arnauld: Everyone is assuming we'll separate Embedded Experiences from the API. I'd like to make a proposal.

<elf-pavlik> +elf:like

<Lloyd_Fassett> Jasnell: The way OS did it conflated the issue. They did a gadget that did many different things.

<tantek> +the_coffee_break

<MarkCrawford> LOL

<Lloyd_Fassett> jasnell: OS embedded experience is too complicated.

<Arnaud> PROPOSED: Separate Embedded Experiences from HTTP client API

<Lloyd_Fassett> jasnell: separating it will make it much easier to get things done

<jasnell> +1

<oshepherd> +http://www.w3.org/ns/activitystreams#like :-)

<Arnaud> PROPOSED: Separate embedded experiences from HTTP client API

<evanpro> +1

<oshepherd> +1

<dret> +1

<MarkCrawford> +1

<jtauber> +1

<elf-pavlik> +0

<jasnell> +1

<tantek> like

<sandro> +1

<AdamB> +1

<Arnaud> RESOLVED: Separate embedded experiences from HTTP client API

<jasnell> "@type":"as:like"

<tantek> 💛

<sandro> ♥

<sandro> <3

<elf-pavlik> "@type": ["elf:like", "as:like"]

<oshepherd> proposal: rename "as:like" to "as:♥" ;-)

<oshepherd> Audio has gone to hell :-(

<elf-pavlik> oshepherd, nice IRI :)

<elf-pavlik> as:♥

<elf-pavlik> http://www.w3.org/ns/activitystreams#♥

<oshepherd> Anonymous person on talky is me, crawled into bed :-)

<harry> We are re-convening

Federation

<dromasca> federation protocol

<dromasca> going through the charter

<dromasca> evan: distributing status updates between different servers

<dromasca> two use cases

<dromasca> distributed consumers

<dromasca> 2. business colaboration

<dromasca> taking data between two security domains probably implemented as servers

<dromasca> feed identified by address or id and subscribe model

<dromasca> other mechanisms that may work - peer2peer ; but we assume it's rather a server to server mechanism

<dromasca> feed and sindication - depends in the way social api ends looking like

<elf-pavlik> harry, i have PoC decentralized system implementation which uses The Bayeux for pub/sub http://svn.cometd.org/trunk/bayeux/bayeux.html

<Loqi> elf-pavlik meant to say: harry, i have PoC decentralizedsystem implementatio which uses The Bayeux for pub/sub http://svn.cometd.org/trunk/bayeux/bayeux.html

<harry> Note that we did ask for Pubsubhubbub to be listed as an input to the charter which would require W3C RF licensing, and BradFitz didn't see any reason to move the spec from its current location or change anything.

<harry> So I think Pubsubhubbub is off limits for Rec-track.

<oshepherd> +1 for integrating both into one common protocol

<elf-pavlik> http://www.w3.org/community/pubsub/

<harry> ACTION: See if BradFitz did sign on off W3C OWF licensing. [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action01]

<trackbot> Error finding 'See'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<harry> ACTION: hhalpin to see if BradFitz did sign on off W3C OWF licensing. [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action02]

<trackbot> Created ACTION-5 - See if bradfitz did sign on off w3c owf licensing. [on Harry Halpin - due 2014-11-03].

<dromasca> Webmention

<elf-pavlik> http://indiewebcamp.com/webmention

<dromasca> previous implementations were not federating with other

<dromasca> exception - diaspora one way federation

<harry> http://www.w3.org/community/pubsub/participants

<harry> BTW, the answer seems to be neither Brett nor Brad signed the W3C Community License :(

<dromasca> every implementation gor its federation protocol

<dromasca> looking at pingback

<dromasca> 1-1 use case - comment on my side, see it on your side

<dromasca> source marked up - this is an entry, got author, content,

<dromasca> federating comments - more implementations than any other procotol

<dromasca> now events, notifications - use webmention

<dromasca> thin protocol based on existing publishing resources

<dromasca> basic flaw of pingback - automated spam problem

<dromasca> invite open review - is the automated spam problem solved?

<dromasca> Arnaud: with pingback one can accept or not . tue here?

<dromasca> yes, true

<dromasca> James: what if firewall in between?

<dromasca> uses httppost

<dromasca> harry: solved the spam problem - what about industry implementation?

<Lloyd_Fassett> dromasca: should be either jasnell or jtauber

<tantek> http://indiewebcamp.com/vouch

<dromasca> jasnell

<oshepherd> tantek: I suggested hashcash in the past. What happened to that line of thought?

<dromasca> analog art flowchart on screen

<dromasca> vouch flowchart

<dromasca> vouch url - you believe the recipient will approve, and links to sender

<elf-pavlik> eg. list of participants in some event which you participated

<dromasca> a trust model that does not require any terms of trust!

<elf-pavlik> you could use https://www.w3.org/2002/09/wbs/35125/TPAC2014/registrants#SW as vouch url

<dret> webmention looks interesting as a solution to the problem it is solving (better pingback), but it's bit hard to see how it actually applies to the federation API mentioned in the charter.

<dromasca> bar raised enough to solve the automated spam problem?

<oshepherd> If you have WebMention follow actions and disconnected PubSubHubub subscriptions, you've reinvented the worst aspect of OStatus!

<dromasca> Sandro - assumption that everybody on its own domain? what if not - vouch would not reflect

<dromasca> Sandro - not vouching from twitter but vouching to twitter

<dromasca> why is vouch based on domain?

<tantek> https://indiewebcamp.com/lj2006

<dromasca> history - explained on subdomain page

<dromasca> webmention - when sending again the semantics is 'i updated' or 'i deleted my comment', etc.

<tantek> https://indiewebcamp.com/reply-context

<dromasca> replay-context - tw came with it, came on url,

<tantek> https://indiewebcamp.com/recursive_reply-contexts

<dromasca> recursive reply-context ; chained reply-context or 'transitive'

<tantek> jasnell: agrees with the domain-level vouch

<oshepherd> jasnell: If you can't make that callback, the link is useless anyway!

<dromasca> jasnell: needs to cross firewall, cannot be just dropped

<dromasca> put content in initial post, not to lose the content

<elf-pavlik> jasnell, with application/x-www-form-urlencoded nesting may not work very well...

<dromasca> jasnell: how you mention profile information? use webfinger to discover?

<dromasca> jasnell - lot of objects that do not have http id - need to be discovered

<sandro> jasnell: In a lot of our use cases, we have a lot of objects which do not have HTTP URIs

<dromasca> jtauber - how generic can this go? any resource to any resource

<dromasca> yes - can go general

<jasnell> sandro: it can be possible indirectly... webfinger is one way of accomplishing that

<sandro> thanks

<dromasca> AdamB - how it would work in enterprise environment where subdomains are not wide spread?

<dromasca> use direct addressing, no need for vouch - do they have spam problems internally

<dromasca> jasnell - may have croo-org

<dromasca> cross-organizations

<dromasca> Evan - Linked Data Platform next?

<dromasca> sandro - does not understand why it's there

<dromasca> harry - it was reuested

<dromasca> sandro - take the http idea of posting to create resource - formalize create a container which has a link to each content list

<dromasca> Arnaud - LDP at lower level, not app specific

<tantek> cleaned up "webmention with vouch" flow diagram: https://indiewebcamp.com/File:2014-285-webmention-vouch-drawing.png

<dromasca> not specific for feferation

<dromasca> Arnaud - process point of view - LDP goes to PR, if it fits could be used

<Zakim> tantek, you wanted to ask is anyone federating comments (or other interactions) using LDP?

<dromasca> sandro - only if it fits

<dromasca> tantek - anyone attempting to federate?

<dromasca> sandro - me and Andrei

<Loqi> AnnBassetti meant to say: s/follow the discussion./follow the nuance of the discussion. But I think it's important to document the consensus of the group./

<elf-pavlik> tantek, check out http://cimba.co - https://www.youtube.com/watch?v=z0_XaJ97rF0

<dromasca> sandro - emulate twitter experience, but does not scale (60k users, 20k servers), using simba, LDP server + access control

<dromasca> sandro - one client implementation, one server

<elf-pavlik> sandro, cimba == kima ? https://github.com/rww-apps/kima

<dromasca> evanpro - would it be reasonable to sketch this process on wiki as proposal> looks close with things done on JSON-LD. Does it make sense? can we sketch on wikki

<dromasca> sandro - hesitates, not sure about use case

<dret> aha. thanks, elf

<evanpro> http://www.openwebfoundation.org/announcements/introducingtheopenwebfoundationagreement

<harry> I'm happy to talk through OpenSocial HTTP API with Sandro

<dromasca> owf - dates 2009 - still valid?

<dromasca> harry - somebody familiar explain pluses and minuses of pubsubhubb -

<tantek> dret: PuSH is not a federation protocol, it's a pushing protocol

<dromasca> dret- pubsubhubbub - push protocol -

<tantek> dret: the latest version hollowed it completely

<tantek> dret: going to be hard to get interoperability

<dromasca> sandro - 0.4?

<tantek> evanpro: 0.4

<tantek> evanpro: idea was to decouple from Atom (XML)

I like @dret's terse statement: "pubsubhubbub is not the federation protocol; it's the push protocol" ... would be helpful to have such an annotation on all these

<tantek> evanpro: and make it more applicable for other payloads such as images or JSON

<tantek> evanpro: it doesn't specify a ping paylod, so you can't do fat pings

<dromasca> evanpro - other payloads

(although all you geeks may be clear on all of that)

<Zakim> elf-pavlik, you wanted to discuss friendship claims in federated network

<tantek> hence you need provenance

<dromasca> elf-pavlik: silos, symmetric relation, i can publish any claims - how people can verify claims?

<tantek> this is one of the reasons we found we had to nounify everything in indieweb

<Loqi> AnnBassetti meant to say: (although all you geeks may be clear on all of that)

<tantek> so you could cite the posts

<tantek> that made the claims

<jasnell> we need symmetric webmention... initial webmention can include an identifier, that the receiver must include on their side, mention back to the original, that is then verified

<dromasca> evanpro - rough idea - verify information stored on different servers

<aaronpk> jasnell: interesting... see current work on webmention status using identifiers for individual webmentions: http://indiewebcamp.com/webmention#Asynchronous_status_polling

<dromasca> tantek - pubsuhubbub - working since 2010

<tantek> tantek looks up a citation for that ;)

<jtauber> sounds like a slogan

<dromasca> evanpro - the primary use activity strings, allows for activities info to flow accross the connection

<dromasca> sandro - xml only?

<dromasca> evanpro - not any longer

<dromasca> feature to broadcast to large number of subscribers

<tantek> I've published notes with PubSubHubbub on tantek.com since 2010-02-01: http://tantek.com/2010/032/t3/bits-tweeting-from-my-site-favicon-twitter-pubsubhubbub

<dromasca> well supported for produces of atom feeds - bloggers, etc.

<jasnell> http://blog.superfeedr.com/state-of-pubsubhubbub/

<dromasca> downside - no access control mechanisn,

<dromasca> no way to limit broadcasting to groups, circles, etc.

<dromasca> fewer and fewer feed providers are supporting it now

<dromasca> Lloyd_Fassett - open source?

<dromasca> evanpro - not sure, gogle published only the spec

<oshepherd> I think once you reinvent it for ACL'd posts you might as well drop it and lose the complexit yand bagage

<MarkCrawford> With respect to the Federation Protocol, I would hope that we can define a protocol that can support multiple solutions and not limit to LDP or webmention or any other. There are those that are focused on OData driven solutions for outbound feeds and experimentation around federations.

<tantek> MarkCrawford: there are those using OData for their outbound feeds

<tantek> MarkCrawford: including SAP and Microsoft

<lehawes> Re SubPubHubbub: Lack of access control is a showstopper for social business

<dromasca> MarkCrawford - must it be based on LDP, or webmention - newer cases based on social data

<lehawes> To clarify on MarkCrawford's comment, the social data is published using oData

<harry> There was a conversation around mapping oData and RDF in W3C ages ago

<Zakim> tantek, you wanted to reply to elf re: verifying relationships in a distributed social network, why we nounify things even like following

<harry> not sure where that went.

<dromasca> tantek - delayed response how to verify relations - every verb becomes a post - list of followers can be linked

<dromasca> evanpro - digital signatures can be another idea

<elf-pavlik> reminds me of hosted and signed assertions in https://wiki.mozilla.org/Badges

<dromasca> jasnell - many of these techniques based on pings / fat pings (content included)

<dromasca> evanpro - pompeio - inbox endpoint that you ping

<evanpro> pump.io

<dromasca> pumpio

<elf-pavlik> http://pump.io/

<dromasca> we're not ready, open to proposal, may need to wait til next f2f

<harry> ACTION: hhalpin to cycle with Wendy and Google on PUSH licensing [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action03]

<trackbot> Created ACTION-6 - Cycle with wendy and google on push licensing [on Harry Halpin - due 2014-11-04].

<dromasca> Arnaud - what is the next step? not clear yet

<tantek> for federation, perhaps we can start a wiki page like we did for social api?

<evanpro> wow!

<evanpro> oshepherd: wow

<dromasca> harry - few suggestions - one draft on http-based api and a strawman proposal for federation

<wilkie> oshepherd: very nice!

<dromasca> ldp work , action handlers work already have owners

<elf-pavlik> jasnell, we never got to http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams2.html#actions

<evanpro> https://www.w3.org/wiki/Socialwg/Social_API_candidates

<jasnell> elf: not yet. I believe we'll be talking actions in detail tomorrow

<evanpro> Micropub is on there

<elf-pavlik> jasnell, cool! :)

<dromasca> tantek - wiki page for api proposals - add more content?

<tantek> we have this now: https://www.w3.org/wiki/Socialwg/Social_API_candidates

<harry> that wiki page looks like its covering everything

<dromasca> sandro - ldp only a place holder

<tantek> shall we start: https://www.w3.org/wiki/Socialwg/Federation_candidates ?

<oshepherd> pffft. EVERY TIME I need to log into w3.org on a new device I have to do the forgot my password thing

<dromasca> Arnaud - try to wrap-up

<dromasca> Arnaud - showing agenda for tomorrow

<dromasca> annotation wg - move to 3:30 (to avoid conflict w/ AC meeting)

<elf-pavlik> I wish you all nice dinner! :)

<KevinMarks> is agenda on wiki right for tomorrow?

<tantek> KevinMarks: yes

<oshepherd> I'll be back on tomorrow evening (afternoon your time, I guess). See you all tomorrow

<KevinMarks> so 9am start?

<evanpro> KevinMarks: what are you doing right now?

<evanpro> Come have dinner with us!

<harry> KevinMarks, do you have any recs for food in the area?

<harry> Faultline Brewery?

<evanpro> We need to find a place to eat nearby?

trying to figure out a nearby place to eat

<elf-pavlik> tantek, thanks once again for setting up talky!

<KevinMarks> I'm having dinner at home shortly, sorry

<KevinMarks> hm, not many good palces over that side

<tantek> elf-pavlik: you're welcome! glad it worked out.

<oshepherd> Yeah, its 00:30AM here. I'll be back tomorrow after work :)

<KevinMarks> to eat there you're either in the hotels nearby or a drive into town

<tantek> KevinMarks: there's all the places in the Mercado Center too

<tantek> I think we're considering the Italian place

<tantek> forgot the name

<Lloyd_Fassett> Informal dinner right now at Tomatino 3127 Mission College Blvd, Santa Clara, CA 95054

<rektide> are the 0.8.2 consumer changes different than 0.9.0 consumer rewrite?

<Arnaud> trackbot, start meeting

<trackbot> Meeting: Social Web Working Group Teleconference

<trackbot> Date: 28 October 2014

<MarkCrawford> Arnaud: we did pretty good yesterday. We started out with a loose agenda but ended up filling the whole day. It is always hard at the beginning when we have no preset issues, so we did pretty well.

<MarkCrawford> Arnaud: It is hard to solve a problem that is not well defined. When we talked about Social API, people don't really know what we mean - client or server. If we come to consensus, we can do a better job of solving defined problems. The IG has a broader scope for use cases, and can come up with use cases over time whereas the WG has to clearly identify what we want to focus on in a shorter period of time.

<MarkCrawford> Arnaud: There is a feeling that the WG doesn't need to redefine use cases, but we should have a document that we can point to that identifies what we want to focus on. If there is a document that we can leverage, then fine, but we have to agree to do this.

<MarkCrawford> Arnaud: The IG has agreed to initially focus on those that are most important to the WG. the IG report is not important, but we need something for the WG that we can point to that clearly identifies what we are working on and what problems we are trying to solve.

<MarkCrawford> elf: - Tanteck is not here yet

<MarkCrawford> Arnaud: identified use cases prove extremely useful as we move forward to develop our deliverables. This issue came up several times yesterday which just reinforce my thinking.

<MarkCrawford> Sandro: do we need only 3 or so valid use cases to do our work?

<Loqi> wseltzer meant to say: rrsagent, this meeting spans midnight

<AdamB> can somebody throw in here the link to the swat0 use cases sandro mentioned?

<MarkCrawford> Sandro: a paragraph or so should be sufficient for each use case

<elf-pavlik> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SWAT0

<rhiaro> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SWAT0

<AdamB> thanks elf!

<AdamB> and others!

<rhiaro> elf-pavlik: What's the talky address?

elf -- we're going to try to get you going with Talky

I have a web cam

<Arnaud> ok, we're trying to get talky set up

<elf-pavlik> Thank You :)

<MarkCrawford> elf: the presentation we are about to see is in the IG wiki main page

<elf-pavlik> http://talky.io (room: socialwg)

<lehawes> Link to IG presentation is http://www.w3.org/wiki/File:Social_IG_TPAC_Update.pdf

<elf-pavlik> MarkCrawford, got it!

<elf-pavlik> thx lehawes!

<rhiaro> elf-pavlik talky should be good to go

<rhiaro> \o/

<elf-pavlik> i'll just watch and listen

<elf-pavlik> pleaes go on!

<elf-pavlik> no worries

<elf-pavlik> we can sort it out during coffee break

<elf-pavlik> AnnBassetti, I see slides directly from pdf

<MarkCrawford> Larry: Social IG Update. Agenda is to talk about work that has preceeded us, review our charter/scope/deliverables, give status update, look at example use cases and next steps.

<MarkCrawford> Larry: slide 2 - review Pre-IG use case work. Social Web Incubator Group has 5 years of effort and at least one solid use case in their final report.

<MarkCrawford> Larry: the other very influential report is the Cloud standards customer Council which has 4-5 solid use cases.

elf, OK .. trying to show you the person presenting ... in this case Larry Hawes .. guy in blue shirt, dark hair

<MarkCrawford> Larry: we have also brought in those already identified by the WG as well as some of our own.

<elf-pavlik> perfect! thanks AnnBassetti :)

<MarkCrawford> Larry: slide 3: charter and scope. Chartered to deliver use case report by 12/31/2016. Intend to harvest existing and produce new use cases that illustrate non-technical requirements.

elf, any idea why my video is split?

<elf-pavlik> no, but let's just roll with it for now

<MarkCrawford> Larry: Slide 4 Deliverables. Use Case and requirements report; Social architecture report (Leveraging Headlights report); social vocabularies inventory.

ok, sorry

<MarkCrawford> Larry: Slide 5 - Membership. Enough to get good work done

<MarkCrawford> Larry: Slide 6 - Process.

<MarkCrawford> Sandro: When is the IG meeting this week?

<MarkCrawford> Larry: during the ad hoc time today

<elf-pavlik> it wasn't me!

<elf-pavlik> ok, good luck with overcoming those issues!

<MarkCrawford> Larry: Slide 7 - Status. # meetings, task forces. Scenarios are just a narrative for use cases. Use Case is more detailed per our template which will give us a formal way to document and publish them.

<elf-pavlik> I'll keep an eye on the room! you folks can got out get some fresh air :)

can you hear, elf?

<MarkCrawford> resuming

<MarkCrawford> still on slide 7 status

<elf-pavlik> yes i can hear

we have the doors open, behind me ... is there too much ambient noise

<elf-pavlik> i can hear good (even better than yesterday!) video got frozen though

<MarkCrawford> Larry: working to categorize use cases and we have a first stab at that on the wiki

hmm .. I see people moving on video

<MarkCrawford> Larry: Slide 8: just read

elf, do you need me to restart video?

<elf-pavlik> AnnBassetti, yes please

<MarkCrawford> Larry: Slide 9 - Next steps: discussed

hmm .. maybe I need to drop off and come back in

<harry> LarryHawes: Going over a slide-set describing use-cases

<harry> ... see wiki for updates

<MarkCrawford> Larry: switched to wiki and discussed how you can find the use case TF from the social ig wiki.

<elf-pavlik> sandro, can you unmute yourself so i can hear?

<harry> ... most of focus on profile federation and social data syntax use-cases

<elf-pavlik> please :)

<elf-pavlik> thx!

<harry> ... we want to work on API and Human interaction, but that's next

<harry> MarkCrawford: We've created a few links to non-existent wiki pages

<harry> ... we will create web-pages in the wiki

<harry> ... and use blanks to refine categorization

<harry> ... currently red ones are from the W3C Headlights report

<harry> ... there's a bit of duplication, will try to refine it.

<harry> LarryHawes: Jump to the use-case themselves

<elf-pavlik> sandro, please mute again :)

<harry> ... mapping to the CRUD framework

<elf-pavlik> AnnBassetti, can you unmute please? :)

<harry> ... social profile creation

<harry> ... its in the template

<scribe> done

<elf-pavlik> great! thank you :)

can you hear?

<harry> ... narrative, actors, goals of actors, success scenario, success criteria, failure criteria

<MarkCrawford> Larry: Social Profile Creation Use Case. We want your feedback on both the template as well as the content.

<harry> ... feel free to give us input on the template itlsef

now there's a big fan behind me .. is that too noisy?

<elf-pavlik> a bit but not a big deal

<MarkCrawford> Larry: Social Profile Creation wiki page

<sandro> https://www.w3.org/wiki/Socialig/Use_Case_TF/Profile_Use_Cases

<MarkCrawford> Larry: Working with an openID model around social profile creation such as google that could then be federated to other social profile maintainers.

<MarkCrawford> Larry: Noted Success Scenario and Success and Failure Criteria.

<MarkCrawford> Not crazy about verb: federate.

<MarkCrawford> Sandro: copied or mirrored

<elf-pavlik> AnnBassetti, can you ask tantek to mute?

<MarkCrawford> Harry: do you want to copy, open, or share?

<harry> copy on request to another system or automatically update other systems?

<elf-pavlik> AnnBassetti, your laptop has better sound then tantek's webcam

<MarkCrawford> James: should at least spell out the different options

ok

better?

<MarkCrawford> Larry: some could have graphic representation. We also need to add id for original submitter.

<elf-pavlik> AnnBassetti, yes and with 2 cameras i can now see everything :D

yes, cool

<MarkCrawford> ?? How do we get comments?

I wish I could zoom my view

<elf-pavlik> + remote control ;)

<jtauber> :-)

<MarkCrawford> use cases need to be on individual wiki pages, have comment section, and tolerate inline changes to leverage wiki concept

<cwebber2> oh, are we still doing our meeting today?

<cwebber2> I guess everyone else is at TPAC

<elf-pavlik> https://talky.io/socialwg

<elf-pavlik> cwebber2, join us :)

<MarkCrawford> Larry: federation of Profile Information to a nNew Network or System.

<jasnell> Mark: there are two James' in the room :-) ... please show my comments as jasnell to differentiate

<MarkCrawford> Larry: read Social Profile Updates UC

<MarkCrawford> Larry: read social Profile Deletion UC

<dromasca> +q

<harry> elf-pavlik, feel free to type question as well

<Zakim> elf-pavlik, you wanted to discuss any implementations we derrive usecases from? eg: PLP has one https://github.com/hackers4peace/plp-docs

<elf-pavlik> https://www.w3.org/wiki/Socialig/Use_Case_TF/Profile_Scenarios#Portable_Linked_Profiles

<elf-pavlik> http://www.open-steps.org/introducing-the-new-open-knowledge-directory-with-plp-profiles/

<MarkCrawford> elf: we have a prototype for the use cases here, wondering if we have others to compare?

<elf-pavlik> we have implementation and i wonder about other implementations

<MarkCrawford> Larry: elf is pointing to a scenario, not something in our UC format.

<elf-pavlik> so we can compare and test interoperability

<MarkCrawford> Arnaud: I don't understand elf's point.

<MarkCrawford> Larry: this is detail oriented, but not UC

<elf-pavlik> do we track somehow implementations?

<MarkCrawford> Arnaud: I am interested in - we just saw an overview of what we saw as profiles. Is that capturing what we are thinking, or is there others that we also need to capture so that my problem can be solved.

<MarkCrawford> ?? Does profile include privacy options?

<MarkCrawford> Larry: It should, and i think we tried to capture.

<MarkCrawford> ?? I define privacy as between the owner and provider.

<MarkCrawford> Larry: No, we don't currently capture that?

<MarkCrawford> Larry: we have not yet gotten to requirements out of these yet. My vision is you build high level use cases and then build requirements out of that.

<MarkCrawford> ?? Is your federated system considered an n to n type of relation?

<MarkCrawford> Larry: that is an implementation choice and there should be a way to control

<dromasca> ?? = dromasca

<MarkCrawford> james: so, the actors in this case we cant create or delete and the system controls. We may be the subject, but not the owner. there is an additional role - profile owner might be the providor.

<Zakim> sandro, you wanted to ask how this connects to identity, and changing identity

<Loqi> dromasca meant to say: ?? = dromasca

<cwebber2> whoo, finally connected

<elf-pavlik> hi cwebber2 :)

<MarkCrawford> Sandro: the way I think of profiles as different than this. Profiles as the art of identity. The way I talk about people is their profile. You frmaed this as like facebook. What are the identifiers?

<rhiaro> So profiles are for personas rather than persons?

<MarkCrawford> Larry: We do not plan to touch ideentity. there is a relationship between these.

cwebber2, can you hear OK?

<cwebber2> yup!

cool

<MarkCrawford> evan: I want to express concern about identify management, it is a big topic that has been addressed by other organizations and we may want to defer.

<MarkCrawford> Ann: won't be solved here.

<sandro> +1 harry: avoid use of the word "identity", and just deal with profiles

<MarkCrawford> Harry: I have a simple rule - dont use the word Identity as there is confusion when it is used. littany of organizations who deal with this. For this particular use case the word profile helps clarify that we are talking about editing and sharing profiles.

<MarkCrawford> Harry: access control is something that is very hard.

<MarkCrawford> Arnaud: Charter did not address the access control issue because it could be a rat hole so we agreed that we would only address requirements/use cases. Next version of charter can address once we are clear on what we are concerned with.

can you guys hear Tantek? (at far end of room from me)

<elf-pavlik> good enough

good

<MarkCrawford> tantek: I share evan's concern about the use of the word federation. Need to be more clear about what you mean. A lot of us use more specific terms like syndication which is much more specific.

<MarkCrawford> tantek: pushing or pulling is something you can implement to.

<MarkCrawford> Larry: I hope it would be as we do requirements as to what choice we are making. We are purposely trying to find a vague term and once we get to requirmements we can refine.

<MarkCrawford> Larry: may be a difference between business and development.

<MarkCrawford> tantek: whatever thing you are federating, you use the same identifier for that across the network.

<MarkCrawford> tantek: phone number and email are examples of federation across networks

<MarkCrawford> tantek: example of breaking federation is twitter t.co. that is not federation as it breaks the identifier.

<sandro> tantek: It's only federation if you're using the original identifier

<MarkCrawford> tantek: both openID and indiauth use the same identifier.

<MarkCrawford> Evan: we have different ideas about this.

<MarkCrawford> sando: is there another word?

<MarkCrawford> tantek: looking it up in the context of technology passport and passport number is an example of federation across countries.

<MarkCrawford> me: tanteks vision strikes me as a federated network for a specific id rather than a federated profile.

<MarkCrawford> ann: trying to understand the distinction between identifier and federation. IDs like phone number, passport number use to identify me, but I don't see that as federated. the example I have in my brain is where we all create our user profile in boeing where the sys creates a generic profile and the subject can then amplify. We then have sharepoint which creates a separate profile. We want to create where sharepoint pulls the profile.

<sandro> http://en.wikipedia.org/wiki/Federation_%28information_technology%29

<MarkCrawford> Tantek - that is not federation, its centralization.

<MarkCrawford> Ann: I don't understand how the passport is federation and not identification

<elf-pavlik> tantek, maybe worth capturing your point somewhere in https://www.w3.org/wiki/Socialwg#Hypotheticals

<harry> I think Tantek has a very good point BTW

<elf-pavlik> or https://www.w3.org/wiki/Socialwg#Federation

<MarkCrawford> action - need to define what we mean by federation

<trackbot> Error finding '-'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<harry> i.e. decentralized creations of identifiers that are accepted across different systems

<harry> but arguing is hard

<MarkCrawford> Ann: we need to collectively define.

and have consensus on our use of these terms

<evanpro> ISSUE: need glossary for terms used in use cases e.g. "federation"

<trackbot> Created ISSUE-5 - Need glossary for terms used in use cases e.g. "federation". Please complete additional details at <http://www.w3.org/Social/track/issues/5/edit>.

<Zakim> AdamB, you wanted to share that within the enterprise its common that people have the same identifier / identity across systems

<Zakim> EdK, you wanted to say that business use cases can extend beyond the internal company, i.e. collaborations beyond (in response to AdamB

<dromasca> +q

<dromasca> http://datatracker.ietf.org/doc/draft-ietf-scim-use-cases/

MarkC: I want to know, as we share profiles across systems, that we have commonality on profile component definitions

Ann: yes, we will come to the day when we have external partners and suppliers logging in at our firewall
... and we will need to federate profiles in each direction

<tantek> when two or more orgs agree to federate, they agree to treat each others' identifiers as first-class citizens, use them directly, and expect that users on each system may use them directly, without requiring "local" wrappers/versions of each identifier.

<MarkCrawford> all - for the last ten minutes my typing was not posting in irc. If you spoke, please enter your comments in irc so we don't loose them

<MarkCrawford> Larry: now that we have a better understanding, we can refine what we are doing.

<AdamB> It’s interesting the different perspective that a corporation brings to these situations. Corporation have one unique identifier for an employee, the ‘employee identifier’. That identifier is used in all the systems I use throughout the course of the week so identity isn’t nearly as much of a problem for corporations.

<MarkCrawford> Evan: I want to get more involved

<elf-pavlik> MarkCrawford, cany you copy paste from your machine?

<dromasca> we seem to be talking about federation of identity information - not all belongs to the social space

<dromasca> i posted the url of prior art work in the scim wg in the ietf - http://datatracker.ietf.org/doc/draft-ietf-scim-use-cases/

<tantek> Examples of federated orgs and identifiers: phone companies & phone numbers; email hosts & email addresses; countries & passport numbers

<tantek> Non-example: US social security numbers: single provider, many consumers.

<AdamB> There is what we refer to as Human Resources (HR) data and then user provided data. A person cannot delete their HR data but can remove their user provided information.

<EdK> EdK discussed that primary "federation" is about profile information between dissimilar systems, primarily between business collaboration efforts e.g. suppliers, university partnerships, etc. This IMHO is beyond identity - for example SKILLS is a profile component, not an identity component.

<MarkCrawford> Arnaud: before we leave, we need to talk about the next F2F meeting. I am hoping that what we have already accomplished these two days convinces folks about the importance of F2F meetings. Perhaps a two day meeting is most reasonable.

<tantek> half-way examples: OpenID and IndieAuth - many *providers* of OpenID & IndieAuth identities, however, not all *providers* also consume all identifiers from other providers.

<MarkCrawford> Arnaud: The F2F is hosted by one of the WG members. Provide a meeting room big enough for the number of expected participants, food, etc.

<hadleybeeman> annbassetti/tantek: Passports are a particularly good example, because they're issued by multiple authorities (countries) and recognised/used by multiple entities (other countries).

<MarkCrawford> Ann: we should be fair to geographic considerations. I need to get permission before I can commit.

<MarkCrawford> Sandro can also volunteer

<elf-pavlik> fair to geograhic considerations - Europe? ;)

<MarkCrawford> Tantek: if folks don't show up they feel left out.

<rhiaro> +1 Europe :)

<MarkCrawford> tantek: Sandro suggested that we try to co-locate with e.g. an IndieWeb camp.

<cwebber2> if doingit near MIT, might be interesting to do it near libreplanet

<cwebber2> time-wise

<MarkCrawford> general back and forth on different locations

<cwebber2> some relevant people might already be in cambridge :)

<elf-pavlik> pool?

thanks Tsyesika

<cwebber2> (libreplanet is the FSF's conference)

are you, cwebber2?

<MarkCrawford> Arnaud: not trying to settle anything yet. Question is timing. List of possible dates and locations.

<cwebber2> AnnBassetti: I'm not in cambridge, but I will be for libreplanet

<rhiaro> tommorris and oshepherd are also Europe I believe

aha

<elf-pavlik> oshepherd?

<jasnell> how many days? 2 ...? with possible interop event included to test implementations?

<tantek> welcome bret on talky!

<MarkCrawford> evan: Perhaps align with due dates when we are expecting to have somehting like a federation protocol that we can discuss?

<harry> It's about 8 folks from Europe, 6 folks from Asia/Austrailia

<bret> You are not getting sound from me right?

<MarkCrawford> ?? Doodle poll?

<elf-pavlik> +1 pool

Harry, and, how many from NAmerica?

<tommorris> rhiaro: I've resigned from the WG for personal reasons, but thanks for the thought.

<Tsyesika> bret you're muted if you're the person who just joined :)

<MarkCrawford> Harry: typically WGs do 6 months.

<bret> cool

<bret> (sometimes my mic still makes noise on Zakim if I mute client side.. no idea why)

<MarkCrawford> Harry: we should rotate so we are not geographically biased.

bret, we're not using Zakim

<Tsyesika> I vote march around the same time as libre planet

<cwebber2> I'm going to be proposing a federation hackathon at libreplanet, for those who might be there

<Tsyesika> as I'm likely to be in boston then

<cwebber2> evanpro: maybe we should coordinate on doing such a thing

<evanpro> cwebber2: that'd be nice

<KevinMarks> I'm about to head over to the location.

<KevinMarks> Should I bring my external mic for later on?

<harry> Mostly

<bret> KevinMarks: yeah audio is pretty rough

<elf-pavlik> KevinMarks, do you have mic with protection from wind?

<MarkCrawford> general discussion around dates.

<evanpro> cwebber2, Tsyesika: that's the end of March, too far out

<lehawes> First week of February in Cambridge, MA works fine for me

<KevinMarks> There is wind? I normally put a towel over it for that

<cwebber2> evanpro: gotcha

<cwebber2> I can probably make something in feb in cambridge also

<tantek> First week of March is better for me

<evanpro> ISSUE: make sure we have teleconferencing equipment for next F2F

<trackbot> Created ISSUE-6 - Make sure we have teleconferencing equipment for next f2f. Please complete additional details at <http://www.w3.org/Social/track/issues/6/edit>.

<MarkCrawford> discussion around providing virtual attendance capabilities for future meetings

<elf-pavlik> let's do next one in summer europe then :D

<elf-pavlik> second next

<tantek> Cambridge MA is fine, also I can host in San Francisco at Mozilla SF for early March.

<rhiaro> elf-pavlik: +1

<rhiaro> Summer in Turin

<elf-pavlik> +1 Turin

<MarkCrawford> evan: How do we push to a decision point?

<MarkCrawford> Larry: don't we need to include those not here now?

<MarkCrawford> arnaud: we need to make sure this is minuted and broadcast so that objections can be raised.

dromasco says there was good WebRTC support for remote participation for joint IETF / W3C meeting

<MarkCrawford> tantek: wants to keep first week of march on table.

we need to find out how they did that

<MarkCrawford> Arnaud: we can have poll

<MarkCrawford> Arnaud to take action to set up doodle poll with different options

<Arnaud> ACTION: arnaud to set up a doodle poll for next F2F, with 1st week of Feb and March [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action04]

<trackbot> Created ACTION-7 - Set up a doodle poll for next f2f, with 1st week of feb and march [on Arnaud Le Hors - due 2014-11-04].

<jasnell> q

<MarkCrawford> Evan: suggest we have as a minimum the chairs and editors can make it to the meeting.

<jasnell> q

<elf-pavlik> can't hear now...

<MarkCrawford> We will resume at 3:30

<tantek> background reading before joint meeting with annotations WG at 3:30: http://indiewebcamp.com/marginalia and http://indiewebcamp.com/fragmentions

<KevinMarks> Yay for fragmentions

<KevinMarks> (biased, me?)

<bret> heh, probably a little

you folks on the camera ... I have to go to AC meeting .. break now, then others will have task force meeting.

they'll use Tantek's camera

see hyou later

<elf-pavlik> thanks AnnBassetti have a good AC meeting!

<bret> which camera is pointing at the projector? seems to be the most clear

<elf-pavlik> bret, frozen one? :D

<bret> well frozen now. Was working earlier

<elf-pavlik> really, stayed frozen for me all the time ...

<bret> weird

<bret> webRTC is funky

<tantek> cool - looks like I confirmed with elf-pavlik that we're live

<elf-pavlik> tantek, unmute?

<tantek> I'll be leaving that connection in the room for the next 4.5 hours but won't be here in person

<elf-pavlik> yes

<oshepherd> Of all days why did my Occulus Rift have to turn up today? :p

<elf-pavlik> oshepherd, you want 3D stream?

<oshepherd> elf-pavlik: No, I want to play with it :p

<elf-pavlik> well you have your chance now for a bit...

<oshepherd> Hah it'll probably take me that long to get it set up

<bret> oshepherd elf-pavlik pair it with https://www.flickr.com/photos/bretc/15466384617/

<oshepherd> bret: WTF is that?

<bret> 360 webcam!

<oshepherd> hah

<elf-pavlik> cool! put it on a drone and fly around room ;)

<evanpro> On the subject of use cases, I added a couple of links to the wiki

<evanpro> https://www.w3.org/wiki/Socialig/Use_Case_TF/Social_API

<elf-pavlik> we do it now?

<elf-pavlik> i just listen

<rhiaro> scribenick: rhiaro

lehawes: Offical meeting of use cases TF begins
... goal to talk about AS scenarios

EdK: Talking about internal and federated

lehawes: How will AS be used within Ford?

EdK: Information streams, by IBM & Ford. Unique value proposition of social platform. Connecting people who don't normally connect, and to connect people with information that they wouldn't normally have connections with
... Nodes aren't just individuals, but the information itself
... Driven from core applications within the company, where information posted from the system itself, rather than by individuals
... eg. all data centre incidents get published (server down, network congestion etc)
... this information published into a stream as events
... once issues are known, reduces calls to helpdesk

<jasnell> for those joining remotely... we're doing a quick ad hoc Social IG meeting. The full WG meeting is reconvening at 3:30 pacific

<lehawes> Thanks jassnell!

EdK: Also, if system posts there's an outage somewhere, somebody elsewhere is having the same issue, which allows issues being tracked and eg. escalated to global vs local

<lehawes> To clarify, this meeting is the Use Case Task Force subgroup of the Social IG

EdK: Need to bridge between social discussions around things, and the system itself. System doesn't have a concept of the social interactions going on
... The social thread becomes an important part of how the system functions

<Zakim> elf-pavlik, you wanted to discuss tracking implementation efforts for use cases we gather and prioritizing them accordingly

<elf-pavlik> i see it important that we track implementation efforts for use cases and prioritaze them accordingly

<lehawes> go ahead elf

<elf-pavlik> in WG we need feedback from people who implement AS etc.

<lehawes> elf: who is "we"

<elf-pavlik> if no one implements use case i would prioritize it lower

<elf-pavlik> ok

jasnell: For every use case, identify someone doing osmething with it - bit early

EdK: Proof of concept not necessarily implementation

<lehawes> elf-pavlik: Understood and agreed on prioritization. I think we're too early in process to begin doing that now, but should happen later.

<jasnell> right now we're in the process of identifying the use cases. Documenting implementers comes next. It's still too early and would be counterproductive to focus on implementations too much at this point

EdK: Four or five procurement systems in use. The person making a request sends an email to someone who puts it into the system, then a buyer gets involved, then there's a finance system, then a recieving system
... The person who needs the item isn't a part of any of those systems. So can't subscribe to notifications from those systems.

<lehawes> Reminder to self to harvest EdK's procurement use case from Social Business Community Group's work

EdK: Security/authorisation issue in large companies.
... So subscribing is one part in the social system. Further, actions that can be taken on things in the stream
... Using an embedded experience to grab data (if allowed), preserving the original security around the data object
... Then to be able to take action about something with the same system, is beyond the normal definition of a social system
... Notion of priorities in AS. If you're expected to take action, you need to be actively notified, not just see it by chance in the stream.

<lehawes> For those just joining the #social IRC chat, this is an ad hoc meeting of the Use Case Task Force of the Social IG. We're having an informal discussion of Activity Streams 2.0 related use cases until Noon'ish.

EdK: This is from the social business community
... It was a proof of concept, not finished building.

<lehawes> Sounds like use cases should be focused on core functions of AS2

Lloyd_Fassett: Sounds like AS didn't have priorities, and now it does, that was the gap.
... The standards we're discussing are generic, not specialised, but a hook for a specialised vocabulary
... Specialisations like domain expertise
... But currently no way to hook that into AS?

jasnell: That's what JSON-LD context is for
... you can have multiple contexts
... publish specialised vocabs in a public location and add it as a context. Reciever can't necessarily do anything with it, but it can be processed

KevinMarks arrives!

<elf-pavlik> R2-D2 style mic ;)

Lloyd_Fassett: So we need to delineate types of use cases
... We're not going to worry about vertical issues, specialisations

<lehawes> elf-pavlik: yes, it's quite shiny too!

<dret> big question is: will AS2 *require* JSON-LD context mechanism, or allow the AS1 way of using terms. my preference would be to allow but not require JSON-LD contexts, and have a registry for namespaces so that people can make their vocabularies know to the world. these may be described in the JSON-LD way, or like the AS1 vocabulary: as a simple list of terms annotated with human-readable descriptions.

jasnell: Two camps of consumers with JSON-LD
... some people just want JSON
... some people want semantics

<lehawes> elf-pavlik: Good. Sorry, but we tried to make it better for you. :>)

<elf-pavlik> thanks! much appreciated :)

jasnell: We can have a publisher of an AS document write a document in more than one way. comes out the same regardless of whether you use JSON-LD
... People looking at things from a pure JSON point of view can't interoperate, because every consumer has to understand JSON-LD
... there's code available, it's not too difficult
... but people who just do json might not like that
... Some clients only care about one endpoint, so they're going to write their client to interoperate with that endpoint
... and that's fine
... if they want to optimise so they're only parsing json, that's fine
... Publishers can add JSON-LD augmentations or not. It's the consumers who have the problem, not knowing what they'll be consuming

<elf-pavlik> in which cases exactly one can't treat JSON-LD as JSON? (not AS1 JSON)

KevinMarks: Adding more constraints doesn't necessarily make things work properly

jasnell: Key thing is we can't target / make it easy for both camps
... (recognising that not everyone will be happy)

<dret> all of this will need to be answered in the processing model section of the AS spec: as a generic client, when you see AS data flowing by, what can you say about it? and also: in which ways can you rewrite AS data (as an AS proxy) without changing what it means?

<KevinMarks_> adding JSON-LD extra to a JSON structure is fine, but requring them to be correct makes this much harder

<lehawes> dret: thanks for injecting that into the meeting record!

<jasnell> the key challenge is that if we allow publishers to use JSON-LD mechanisms, consumers MUST be capable of processing those, or we break interop

<dret> elf, JSON-LD always is JSON, of course, but for somebody with JSON-LD goggles, the world looks different. if you want to see the world the same way, you essentially have to implement JSON-LD.

EdK: Client-side resolution starts sounding like aggregation. Talking about resolving things into a thread so that the conversation makes sense. Keeping the social fabric seperated. It's useful to end user, but not to overall enterprise

<KevinMarks_> jasnell if you MUST the LD parts you break interop directly

EdK: So question becomes: where is the connection between the system and the social?

<elf-pavlik> dret, even if we recommend always using JSON-LD compacted with AS2.0 context?

<jasnell> publishers are not required to use the LD stuff

<jasnell> they MAY

EdK: Direction of discovery is an issue. User has to find something vs thing finding user through system

<dret> jasnell key question is if AS2 allows free-form JSON as AS1 did. this would be invisible to consumers with JSON-LD goggles.

Lloyd_Fassett: That's data-first. Kind of different from this group, which is profile-first

<KevinMarks_> I thought JSON-LD was supposed to add context, not remove it

<elf-pavlik> dret, can you post some concrete examples to mailing list?

<jasnell> the AS2 @context maps unknown properties to blank nodes, so they are at least preserved through the expansion algorithm

Lloyd_Fassett: In data first, define the data well, and add the social layer on top

lehawes: To bring this back to AS, jasnell: are there specific use cases to help move forward
... ?

jasnell: No

<dret> jasnell if you say MAY JSON-LD, then JSON-LD is out of the picture in the processing model section, and there's just an appendix that explains why the base syntax supports JSON-LD, but processors have to be prepared to deal with any JSON. that will make life hard for people working with JSON-LD toolsets, because they cannot assume to always see everything through their JSON-LD goggles.

jasnell: Where we use AS, it doesn't matter if it's coming from a system or a person. Something happens and we need to let people know what happens based on their relationship to the actor or the object
... Entry point to a workflow / starting point to an interaction

<KevinMarks_> are these old use-cases relevant? http://wiki.activitystrea.ms/w/browse/#view=ViewFolder&param=Examples

jasnell: I have to know what's going on in order to be able to respond to what's going on
... AS is a lower priority than actions. Acting on something in context is the next step
... Right now, push notifications send an action, but use their own format. Needs mapping to Android, iOS, webapps, etc. Need a consistent model for this.

EdK: Once you get to that, you want the response mechanism as well
... to also be consistent

lehawes: So you would see this task force focussing on the actions / embedded experiences use case, rather than sharing the stream?

jasnell: Sharing is pretty well defined
... AS is the latest iteration of this, been working on for years
... We don't need to worry about documenting use cases for AS itself
... But for embedded experiences, potential actions, we do
... need use cases
... This will be discussed more this afternoon.

lehawes: Any other comments?

Lloyd_Fassett: There are actions at the AS 2.0 level, then there are embedded actions?

jasnell: They're basically the same
... AS lets you attach a Potential Action to any Activity
... eg. potential http request or opening a browser view
... describes specific actions (or commands) that can occur
... eg for carrying out a like or a follow

Lloyd_Fassett: Can't that adapt to a market place scenario?

jasnell: Yes, that's one of the use cases being worked on
... where multiple third parties make offers, and how you handle it depends on who is providing it, but it's all tied to the same 'offer' action

EdK: It's the same offer, but if you already have a relationship with someone it might be different to if you don't - but implementation worries about that. 'Offer' action is the same

lehawes: Food market scenario can be fleshed out to become a use case
... by Lloyd_Fassett

jasnell: another use case is pushing to mobile devices
... pushing an activity out as a notification
... We want to be able to integrate with notifications without having to write extensions to the mobile device APIs
... Need simple declarative ways of describing what an action is supposed to do

KevinMarks: At the moment this is done by people embedding forms and iframes

jasnell: Currently you can say eg. you do a share by embedding this html
... if it's a browser view, you do a share by opening this window
... if it's a REST API, you do a share by passing these params
... etc.
... Currently no standards for deep linking
... Every platform does it differently, and 3 or 4 competing efforts
... 4 different for Android currently in use

lehawes: We need actions use cases
... Post scenarios to turn into use cases please!

<oshepherd> Not sure what you mean by "4 competing efforts" on Android for sharing things

Lloyd_Fassett: Could flesh out with jasnell if time

<oshepherd> Erm, you fire off an ACTION_SEND intent

<KevinMarks_> on android, you can claim http URLs as intent from an app; the OS manages contention - on iOS you need to define your own protocol per app and there is no contention resolution

Lloyd_Fassett: it is a real-world scenario

<oshepherd> KevinMarks_: Only if you're explicitly targetting another app. On Android sharing is a part of the platform

lehawes: We can get multiple scenarios around the same concept that can be normalised into one use case

Lloyd_Fassett: To do a marketplace you probalby need JSON-LD

jasnell: yes, for domain specific information
... There are existing vocabs out there

<KevinMarks_> on android, my app can claim http://twitter.com/*

<KevinMarks_> the OS will manage contention over them

<elf-pavlik> FYI http://blog.schema.org/2012/11/good-relations-and-schemaorg.html

<KevinMarks_> on iOS behaviour is undefined

<oshepherd> KevinMarks_: But claiming URIs isn't at all involved in the process of "sharing" things

Lloyd_Fassett: MarkCrawford says vocab task force should create a list of vocabs

<oshepherd> To share things on Android you send an ACTION_SEND intent without a URI

lehawes: Watch out for licensing issues

<KevinMarks_> see http://epeus.blogspot.com/2013/06/how-apples-ios-fragmentation-problems.html#connect+between+applications

lehawes: but they should be on the list so we can eliminate things
... so put everything on the list
... might get ruled out, but don't have that constraint building list
... Time to wrap up!

<elf-pavlik> for finding vocabs: http://lov.okfn.org/dataset/lov/

<KevinMarks_> share is an OS verb on android that anyone can hook; on iOS this was hard-coded, just changed a bit

Lunch time!

<elf-pavlik> enjoy your meal everyone!!!

<lehawes> Thanks everyone for your thoughts around scenarios and use cases related to action-oriented aspects of AS2!

<KevinMarks_> "Note: If more than one third-party app registers to handle the same URL scheme, there is currently no process for determining which app will be given that scheme." https://developer.apple.com/library/ios/documentation/iPhone/Conceptual/iPhoneOSProgrammingGuide/Inter-AppCommunication/Inter-AppCommunication.html

<oshepherd> KevinMarks_: Yeah. There seemed to be insinuation that there were ~4 ways to share things on Android. Maybe I was misreading the conversation

<elf-pavlik> cwebber2, still with us?

<elf-pavlik> Tsyesika, online?

<Tsyesika> yep

<elf-pavlik> great! :)

<elf-pavlik> i prepare comparison of actions in AS, schema.org and Hydra

<elf-pavlik> and may use example of movie review with rating

<elf-pavlik> do you have review/rate feature in MediaGoblin?

<Tsyesika> not at the moment, just commenting

<Tsyesika> would be a good feature to have

<elf-pavlik> schema.org uses such example in their docs http://schema.org/docs/actions.html 'Example: Movie review site API with -input and -output'

<elf-pavlik> cool, once i have it published maybe you could give me feedback?

<Tsyesika> okay, how long until you publish

<Tsyesika> you cought me as i was going off for the evening :P

<elf-pavlik> oh, please just go ahead :) anytime in next days will work just fine!

<Tsyesika> okay cool :) feel free to ping me on here or via email and i'll pick it up tomorrow :)

<evanpro> I added a diagram to the wiki

<evanpro> https://www.w3.org/wiki/images/5/54/Social-api-layers.png

<Tsyesika> i think chris will be around later too

<cwebber2> elf-pavlik: oh hi

<evanpro> Which I think might be useful to guide our ideas about the Social API

<cwebber2> I didn't realize things were continuing, opening back up

<elf-pavlik> ok, have a good one Tsyesika!

<cwebber2> elf-pavlik: I did get your email btw

<cwebber2> and will respond to it... I just got back from another conference

<elf-pavlik> cwebber2, no rush!

<elf-pavlik> evanpro, looks cool :)

<evanpro> elf-pavlik: it is!

<cwebber2> evanpro: elf-pavlik: btw, I know it's maybe not directly in the scope of this working group, but I wonder if maybe it would be good to get feedback from here, esp from those with experience

<cwebber2> tsyesika is currently overhauling our db schemas to handle the additions of federated activitystreams stuff

<cwebber2> it's quite a lot of stuff to figure out how to handle moving this into a relational db design, especially given that simple foreign key patterns don't work anymore once you're about to link to just about anything (though we have some ways around we're investigating)

<elf-pavlik> evanpro, do you understand my point with advantage of hypermedia controls over arbitrary paths like /evan/notes /evan/images /evan/friends /evan/meals /evan/books etc. ?

<cwebber2> I'm not sure if there's anyone with experience or interest who might be interested in reviewing... it could be helpful, also, for mediagoblin to write up how we handled the schema in a relational db so other implementers can reference it

<cwebber2> I realize that's out of scope of the wg specifically tho :)

<evanpro> elf-pavlik: yes I absolutely do

<evanpro> I think that's an API design issue that we'll need to address

<elf-pavlik> cwebber2, please send it directly to WG (implementers) list, i don't see it out of scope at all!

<cwebber2> elf-pavlik: okay, great!

<cwebber2> Tsyesika: ^^^^

<evanpro> The big advantage of arbitrary paths is that it's what social developers are used to

<cwebber2> we'll talk tomorrow about doing that

<elf-pavlik> in silo environments ...

<evanpro> Yep

<elf-pavlik> similar as symetric relations like friends, we will face some need to change our way of thinking in federated environment

<cwebber2> elf-pavlik: not sure where the arbitrary paths bit is being referenced from, but I agree it is important if I'm reading right

<elf-pavlik> but i do get your point about people currently having certain expectations to APIs

<elf-pavlik> cwebber2, we had this conversations yesterday, sorry for mixing conversations ...

<cwebber2> esp since projects that are trying to add federation later may have their own URL schemes

<cwebber2> ah :)

<cwebber2> apologies if I'm jumping in unhelpfully

<elf-pavlik> cwebber2, i recommend watching https://www.youtube.com/watch?v=UkAt9XSOfaE

<elf-pavlik> also https://www.youtube.com/watch?v=KB94dIamAQc

<cwebber2> elf-pavlik: okay, will watch... maybe tonight when unwinding :)

<elf-pavlik> much shorter :D

<elf-pavlik> second just 7min and IMO very relevant to MediaGoblin!

<elf-pavlik> cwebber2, James will talk about Actions during next session - this 7min vid can give you great intro :)

<cwebber2> elf-pavlik: aha, great! Tsyesika and I were trying to see how we might make use of them

<cwebber2> am I right that in some ways they could be useful as a kind of flip side to webintents?

<elf-pavlik> also Caley may solve your issues with relational database ...

<cwebber2> ie, webintents has a user say "I want to favorite with this" whereas this might mean between services "you can favorite this with our tool here"

<cwebber2> ?

<cwebber2> I guess I can wait for the talk :)

<oshepherd> How long until the next session? 15 min?

<elf-pavlik> https://www.w3.org/wiki/Socialwg/2014-10-27#Tuesday_28_October_-_With_breakouts

<oshepherd> Currently geting a bunch ofn oise over the talky. elf-pavlik, an issue on your end?

<elf-pavlik> no session now

<elf-pavlik> next one in ~3h i think 15:30 their time

<oshepherd> Yeah, thats why I presume you, they're offline :p

<oshepherd> Maybe try muting yourself?

<cwebber2> elf-pavlik: great video, thank you for sharing :)

<elf-pavlik> welcome :)

<elf-pavlik> relevant to GMG?

<cwebber2> highly, I think!

<cwebber2> elf-pavlik: one thing we have been figuring out how to deal with is that many mediagoblin sites can have different media types supported

<cwebber2> and those media types may be configured to have certain features on / off

<cwebber2> so you can't assume

<elf-pavlik> for example?

<cwebber2> this seems like the perfect way to broadcast what without assuming

<elf-pavlik> with hypermedia you just include supported controls so client app will not expect features you don't provide!

<cwebber2> elf-pavlik: well, the different media types is an easy one: we have images/video/audio/documents/3d models

<cwebber2> ascii art ;)

<cwebber2> by default, only images is turned on

<cwebber2> but

<elf-pavlik> yes, so audio would have 'listen' control and video 'watch' for example

<cwebber2> right

<cwebber2> there are some features, like geolocation support, which aren't on by default also for images

<cwebber2> but you can turn them on

<cwebber2> geolocation is a bad example, there are better, but I'd have to look at the config specs since I'm very jetlagged :)

<cwebber2> but you get the idea.

<elf-pavlik> cwebber2, do you have some form of collections / albums ?

<cwebber2> elf-pavlik: we do, they're a bit crude, but we have a collections feature

<elf-pavlik> BTW i think we SHOULD extract some use cases based on GMG features!!!

<cwebber2> :D

<elf-pavlik> expecially if you will work on implementing stuff and providing much needed feedback

<wilkie> yes, real-world implementation use-cases are great

<cwebber2> elf-pavlik: http://mediagoblin.com/u/cwebber/collection/mediagoblin-rewards-2014/

<cwebber2> a collection

<elf-pavlik> wilkie++

<Loqi> wilkie has 2 karma

<cwebber2> could be better

<cwebber2> but they're there

<cwebber2> elf-pavlik: I've somewhat wondered how we're going to express different media types with federation

<cwebber2> we already have some... I mean, audio / video are represented

<cwebber2> but stuff like presentations aren't.... I've figured json-ld will help make this cleaner

<cwebber2> but one thing that the actions stuff seems to make really appealing is that it removes the assumptions/guesswork from stuff like audio / video playing

<elf-pavlik> we can create types for them, for example sub types of http://schema.org/CreativeWork

<cwebber2> elf-pavlik: ah great :)

<wilkie> they should be extensible enough. should be able to create an arbitrary object type. "image", "displayName" and "content" are all generic for all objects in AS and can be used to kinda represent them on systems that don't know what those object are.

<wilkie> and they you'd have a link to the actual content which is navigated in the normal way or something

<cwebber2> so actually

<cwebber2> one thing Tsyesika and I are discussing that I think is going to be cmplex

<cwebber2> complex

<cwebber2> is assuming you do something like youtube-style "subscriptions" in mediagoblin

<cwebber2> or even adding media on different domains to collections, etc, other federation examples

<cwebber2> if you click a thumbnail to a federated video, for instance

<cwebber2> will that destroy navigation locally?

<cwebber2> there are a lot of UI decisions we're going to have to make... at the moment, I've assumed those are outside of the WG

<elf-pavlik> cwebber2 i think James will present embeded experiences during next session :)

<cwebber2> but maybe if other people are thinking about these things

<cwebber2> oh great!

<cwebber2> stellar

<elf-pavlik> http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams2-actions.html#dfn-embeddedview

<cwebber2> elf-pavlik: btw, you may be interested

<wilkie> oh yeah, embedded views are possible. but even in the low-tech minimal sense, a thumbnail that navigates is not that bad I don't think.

<cwebber2> that Tsyesika is collecting her research as she goes at https://theperplexingpariah.co.uk/GMG/mediagoblin_research.html

<elf-pavlik> from old draft: http://elf-pavlik.wwelves.org/w3c-socialwg-activitystreams/activitystreams2-actions.html#embedActionHandler

<elf-pavlik> cwebber2, thanks for link!

<cwebber2> elf-pavlik: yup

<cwebber2> maybe we should be getting more in the habit of Tsyesika and I discussing some of the federation implementation issues in here as we run into them

<cwebber2> as well as on the list, obviously

<cwebber2> :)

<elf-pavlik> cwebber2, please do!

<elf-pavlik> otherwise if you two plan to figure it all out on your own, please let us know so we stop wasting our time and find something else to do ;)

<oshepherd> cwebber2: BTW, as somebody familiar with the Pump API, you might like my WIP soon-to-be submission: http://oshepherd.github.io/activitypump/ActivityPump.html . Comments very much invited :-)

<cwebber2> elf-pavlik: haha :)

<cwebber2> no, I think we'd prefer to work together :)

<oshepherd> elf-pavlik: I'm not sure if you're aware but GMG is going to implement the Pump.io API at the present moment in time

<elf-pavlik> cwebber2++

<Loqi> cwebber2 has 2 karma

<cwebber2> we're already in process of implementing it

<cwebber2> or rather, Tsyesika is in process

<cwebber2> but we intend to support the social wg specs, assuming all goes well :)

<elf-pavlik> IMO implementing Pump.io API as its stands now can give very useful input to Social WG work!

<cwebber2> enjoy your typing break ;)

<evanpro> oshepherd: I'm interested in your thoughts on https://www.w3.org/wiki/images/5/54/Social-api-layers.png

<oshepherd> evanpro: My concern with all resource-based direct updates is when do these turn into activities, or, more importantly: do I do a thing and that generates an activity, or do I post an activity and that does a thing?

<evanpro> Yep

<evanpro> It's confusing!

<oshepherd> I kind of feel that it should be "You do something, that generates an activity", because that makes all activities not special

<evanpro> At least with pump.io a POST to /user/evan/following will generate an activity

<oshepherd> (hard to concentrate on things when new toy is screaming at you "play with me!" :p)

<evanpro> No way

<evanpro> Cool!

<evanpro> What do you do with it/

<evanpro> ?

<oshepherd> evanpro: Its' only just arrived, so mostly set it up so far :p

<evanpro> oshepherd: I find it really interesting but don't know what I'd do with it

<oshepherd> Maybe if I ever find the free time (HAH!) I'll develop something for it...

<oshepherd> In the meantime its' squarely in the "Play" category

<jtauber> I haven't done anything interesting with either of mine (DK1 and DK2), but here are my (ongoing) notes in case you're interested (although most of them are about the DK1): https://thoughtstreams.io/jtauber/oculus-rift/

<oshepherd> Oh FFS energy companies are useless "Heres your bill. You can pay it online" Great! Whats my account number? -_-

<oshepherd> Like, seriously, why is my account number not on the bill?!

<oshepherd> Hurrah! Of course its' hidden in the small print about paying by telephone

<rektide> thanks for the great post, departed evanpro

<oshepherd> jtauber: I'm fortunate enough not to wear glasses

<jtauber> yeah, it's tough with them

<oshepherd> One of the game demos I tried put a reflection of your face in the field of view. That was quite cool

<oshepherd> (your avatar is wearing a helmet, so it makes physical sense)

<rektide> evanpro: thanks for the great diagram! really enjoyed seeing something

<evanpro> rektide: I hope it informs the discussion

<evanpro> rektide: by "mentions" do you mean the type of thing where I say "@rektide good idea!" and it goes to a special stream that shows all posts that mention you?

<wilkie> that would get filtered through the general inbox for activities?

<wilkie> although you probably want a route to pull all mentions that have been filtered "/evan/mentions"

<evanpro> Right

<wilkie> seems like an optional thing for implementors

<wilkie> since it is more server->client than server->server syndication stuff

<evanpro> So it's probably worth breaking this kind of thing up

<oshepherd> I think filtering should be orthogonal to the general API

<wilkie> I agree

<oshepherd> /evan/inbox?type=Mention&type=Like or whatever

<evanpro> That's an interesting idea

<wilkie> it's possible that "mentions" is a completely separate feed, and thus a whole other object and thus a unique resource

<oshepherd> You should be able to request filtering from any API endpoin (and the server should also be free to ignore you)

<evanpro> I don't know where to capture this stuff

<wilkie> but filtering of any particular feed, yeah, would be good

<evanpro> OK I started here https://www.w3.org/wiki/Socialwg/API/Layers

<wilkie> it's the kind of thing that you'd want to subscribe to... "/evan/activities" where activities are posted articles

<wilkie> so in that case, it is server->server facing hm

<rektide> evanpro: yeah, external resources with inbound linkage

<rektide> wilkie: i think /evan/mentions might be a candidate collector for mention resources, but otoh i'm not sure why individual resources wouldn't expose their own mentions collector

<wilkie> if you want to filter a feed, you'd do it as a query like oshepherd suggested

<evanpro> Mentions really breaks down to two different important actions

<oshepherd> TBH I think that "mentioning" somebody is really just cc:ing them

<evanpro> One is sending someone a note or image or whatever "@oshepherd check this out!"

<evanpro> And the second is search and monitoring, "Who is talking about me?"

<rektide> i don't feel like /evan/inbox is a good permenant resource for the mention... where does it really get filed?

<oshepherd> rektide: In your inbox. Everything gets filed in your inbox

<evanpro> Not all social software conflates the two

<rektide> oshepherd: ahh ok. sorry, not quite up to speed /w ya'll across the board.

<rektide> evanpro: agreed with your breakdown

<oshepherd> rektide: Just think of your "inbox" as your feed (what you see on the homepage of facebook or twitter or whatever)

<wilkie> a "timeline"

<wilkie> or whatever

<evanpro> Right

<evanpro> "Stuff by people I follow, or in groups I'm a member of, or sent directly to me, or by me"

<evanpro> Roughly

<elf-pavlik> next session in 20min ?

<evanpro> I don't know if anyone else saw this great infographic on Web Annotations

<evanpro> http://www.w3.org/annotation/diagrams/annotation-architecture.svg

<evanpro> But I really like it

<jtauber> that's great!

<elf-pavlik> evanpro, when do we start?

<evanpro> elf-pavlik: I think 3:30PM PDT

<evanpro> Roughly

<elf-pavlik> in ~10 min

<KevinMarks> now

<cwebber2> should we be joining talky.io?

<cwebber2> I see a potted plant, at present.

<KevinMarks> we're in the other room - the #annotation one

<evanpro> So we just moved into #annotation

<elf-pavlik> cwebber2, yes i wait for someone to start it in the room

<evanpro> People who are participating remotely may want to join that

<evanpro> Since someone is scribing there

<cwebber2> great

<bigbluehat> hi new friends!

<KevinMarks> can you scribe in 2 rooms at once?

<elf-pavlik> hi #annotation :) hey tilgovi!

<KevinMarks> I could noterlive...

<evanpro> We're not scribing here

<KevinMarks> slides for #annotation http://www.w3.org/2014/annotation/diagrams/annotation-architecture.svg

<elf-pavlik> could someone start https://talky.io/socialwg ?

<evanpro> So please come to #annotation

<tantek> we're asking to scribe in here

<tantek> we are scribing here

<KevinMarks> who will annotate the annotators?

<azaroth> ScribeNick: nickstenn

<evanpro> !!!!

<tantek> because #social folks are split across rooms

<tantek> physical rooms

shepazu: we have a charter that contains a number of things
... some deliverables we hope we don't have to make and #social will make instead
... want to explain what we're trying to accomplish overall for annotations
... [presents diagram of "Web Annotation Architecture"]
... this is an attempt of a description of what we hope to see as the overall architecture
... everything starts with someone creating some content

<KevinMarks> http://www.w3.org/2014/annotation/diagrams/annotation-architecture.svg is diagram if you just joined

shepazu: content creator sends content to a publisher (maybe the same person) who publishes it
... footnotes and comments may be added to the document
... comments aren't in separate documents
... another role -- a reader -- who can annotate documents
... can annotate footnotes or comments
... they target a discrete selection within the document
... support for overlapping annotations
... annotations aren't supposed to live on the same document
... when I submit a comment to someone's blog, I'm implicitly entering into a creator-publisher relationship with the blog owner
... we think it would be more interesting if an annotator can publish their own annotations without the say-so or involvement of the publisher/creator of the document
... we also want the ability to notify a publisher of annotations made on their documents

<raphael> s/Topic: Joint meeting Social-Annotation WG/Topic: Joint meeting Social/Annotation WG

<Loqi> raphael meant to say: RRSAgent, draft minutes

shepazu: annotation services provide a way for readers to discover and subscribe to annotation content on the web
... the aim is a distributed mechanism for creating, storing, and discovering annotations
... we introduce another character, the "sharer"
... people can share the annotations [as first-class entities]
... and there's a chain of attribution

<Loqi> nickstenn meant to say: ... and there's a chain of attribution

James Snell: this is a perfect use case for the stuff we're looking at

scribe: this flow here fits in perfectly with what we've been talking about

??2: i'll run through a status update on what [social] have been working on

<harry> s/??e/Evanpro

<KevinMarks> ironically can't annotate that presentation with fragmentions as it's in SVG

<evanpro> http://evan.prodromou.name/files/TPAC/

harry: mark crawford is chair of interest group

mark crawford: we have a template for use cases

scribe: on the wiki
... you can fill that out, add any figures you want

evanpro: social ig/wg chartered in july

<Loqi> tilgovi meant to say: I don't have to escape those ?s do I? This isn't regex?

evanpro: social ig is working on use cases and issues around social and identifying ways the w3c can be helpful in this space
... three main deliverables

<KevinMarks> Loqi is logging

<shepazu> Social data syntax

evanpro: 1) social data syntax

<shepazu> Social API

<shepazu> Federation protocol

evanpro: 2) social api
... 3) and a federation protocol
... 1) is a repr of activities (status updates, etc.) that happen in a social context -- changes to the social graph, content creation, content replies, etc.

<tilgovi> This link (appears first above) is to slides being summarised right now: http://evan.prodromou.name/files/TPAC/

evanpro: it's a syntax that's JSON(-LD)
... just went to FPWD with activity streams 2.0
... activity streams is a subject-verb-object structure "Evan - posted - this image"

<KevinMarks> Loqi is the social logging bot for indieweb et al - generates nice HTML logs http://socialwg.indiewebcamp.com/irc/social/2014-10-28#bottom

<raphael> Activity streams 2.0 FPWD at http://www.w3.org/TR/activitystreams-core/

evanpro: 2) social API -- a client/server HTTP based API that will use the social data syntax

<AnnBassetti> elf, setting up talky right now

evanpro: there's a client embedding API

<Loqi> Loqi has 314 karma

evanpro: we are currently in the process of developing this social API
... working off existing work that came out of open social, and a number of other social APIs
... goal is to get social API to a WD this quarter
... 3) federation protocol -- a syndication of activities across different security domains
... if I generate a number of activities that I'm storing in a social store
... people will be able to follow me remotely and receive updates on those activities -- under my control
... aim is to have a well-documented mechanism to transfer social data between different locations
... we could walk through the Web Annotation Architecture, but a lot of people in the room can probably see where they fit pretty quickly

shepazu: everything that's not at the document level is a pretty obvious fit
... i want to call out the work we're doing that you're not
... we're defining a data model that says what an annotation is
... also working on serializations

<KevinMarks> people like JSON-LD?

shepazu: working on the piece that links to a specific part of a document -- what we're calling robust anchoring

<cwebber2> yes, I like JSON-LD

tantek: some of the work on webmention may have overlap on that as well

<cwebber2> it solves a major problem for us

james snell: we're not modelling the specific content types so this fits very well for us

<elf-pavlik> thank you AnnBassetti :)

<Loqi> AnnBassetti has 1 karma

AnnBassetti: [introduces people from Europe on a webrtc link]

fjh: I want to know what the actions are when we're done discussing

shepazu: at the very least we should be giving [social] some use cases

james snell: we're looking for feedback on our FPWD

<tilgovi> elf-pavlik et al, video laptop just lost power, coming back momentarily

tantek: hi, I'm the co-chair of the social WG
... one of the inputs to the federation protocol is webmention
... [demos a note, with a bunch of comments that have come into the site via the webmention protocol]

fjh: is this an open-source implementation?

tantek: [demos an example of a reply made using an open-source implementation hosted on someone's own site]
... the other example you're probably more interested in is marginalia

<AdamB> sorry guys, lost power on the webcam

tantek: Aaron Parecki combined webmentions with fragment identifiers to create "fragmentions"

<raphael> Second demo link is https://kartikprabhu.com/article/marginalia

<evanpro> Aaron Parecki

tantek: [shows example of inline margin comment, published again on someone's own site, but which makes reference to a specific location in a published document]

shepazu: I'd like to see what the next steps look like

<tantek> elf-pavlik: so far no JSON-LD needed for webmention comments, or marginalia, or fragmentions

<cwebber2> may very well be needed for adding new media types in mediagoblin, etc

<cwebber2> or, at least, keep various identifiers in the json from being lonnnnng

fjh: why do you folks need use cases from us?

James Snell: we can look at it and see if our stuff [Activity Streams] fits what you're doing

<tantek> In summary - I showed this post with favorites and reposts and comments federated via webmentions: http://aaronparecki.com/notes/2014/10/15/3/border-none

harry: we can formalise the relationship between the WGs if there are clear dependencies

fjh: there may be some of our standards effort that overlaps with social APIs, and some that doesn't

<tantek> and then I showed Marginalia: https://kartikprabhu.com/article/marginalia and the annotation on this fragmention: https://kartikprabhu.com/article/marginalia##So+if+you+sent+a+webmention posted on another site here: https://kylewm.com/2014/06/this-is-super-exciting-great-work-building-it-and-a-real-testament

<harry> Social IG for use-case discussion is Wednedays at 10:00 AM

<harry> Eastern (biweekly)

fjh: is fragmentions something you're looking to standardise?

<harry> Social WG (focussed on ActivityStreams and specs) is Tuesdays Noon Eastern

<harry> See web-pages

<harry> www.w3.org/Social/WG

<harry> www.w3.org/Social/IG

<tantek> KevinMarks - came up with fragmentions after attending the Annotations Workshop in April

<tantek> see http://indiewebcamp.com/fragmentions for the spec

KevinMarks: came up with it after I Annotate 2014. it was the simplest thing I could think of which addressed the problem of identifying a part of a document.
... so far it's just on the IndieWebCamp wiki

<KevinMarks> it was the annotations meeting in April that inspired fragmentions: http://www.kevinmarks.com/w3cannotation.html#quotation+from+the+work http://www.kevinmarks.com/fragmentions.html

TimCole: there are potentially issues with fragmentions (IP, overloading the meaning of fragment identifiers, etc.)

shepazu: the way fragmentions and what we're looking to do dovetails is as follows
... you currently need a javascript library to make that work
... one possibility is that we standardise a "find text" in page API

tantek: hasn't been proposed for standardisation, but it is CC0, on the web

<tantek> and interoperably implemented

<tantek> see http://indiewebcamp.com/fragmentions#Open_Source for implementations

<KevinMarks> there is a prior art discussion at http://indiewebcamp.com/fragmention#Related+work

<AnnBassetti> Lloyd Fassett

Lloyd_Fassett: there was a mention of if someone liked "Ice Cream" they could follow "Ice Cream" ... what are the proposals for that?

someone: probably a throwaway comment we can ignore

<AnnBassetti> elf , did you want to be on queue?

<Loqi> nickstenn meant to say:someon: not sure that we have an answer for this

<elf-pavlik> AnnBassetti, just Zakim not present when fjh tried get on queue...

<Loqi> fjh meant to say: q?

<AnnBassetti> aha

sandro: is there a nominated liaison between the two groups?

<elf-pavlik> +1 liaison

azaroth: we have no invited experts at the moment, but this may be an opportunity to add one

<scribe> ACTION: identify a liaison between the two WGs [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action05]

<trackbot> Error finding 'identify'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<harry> I believe we just nominated Kevin Marks

<harry> or as he prefers to be called in IRC, KevinMarks

<scribe> ACTION: KevinMarks to act as liaison between the two WGs [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action06]

<trackbot> Error finding 'KevinMarks'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<elf-pavlik> KevinMarks, your w3 nick?

<KevinMarks> kmarks2

Arnaud: there's a pretty big overlap in general, not restricted to robust anchoring

<elf-pavlik> ACTION: kmarks2 to act as liaison between the two WGs [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action07]

<trackbot> Error finding 'kmarks2'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

fjh: maybe just Evanpro and I should just talk and arrange the liaison we deem appropriate

<trackbot> Sorry, raphael, I don't understand 'trackbot hates nickstenn :-)'. Please refer to <http://www.w3.org/2005/06/tracker/irc> for help.

Arnaud: it's good we've identified the overlap early on

<trackbot> Sorry, elf-pavlik, I don't understand 'trackbot says Loqi s/LOL/:D/ ??'. Please refer to <http://www.w3.org/2005/06/tracker/irc> for help.

tantek: it'll be interesting to see how the overlap pans out, as the social WG is focused heavily on use cases, user scenarios, whereas you are starting from an abstract model

azaroth: we have discussed use cases at length in the CG, it's simply that we're not going to be rediscussing those in the WG

<AdamB> http://www.w3.org/community/openannotation/

James Snell: I can take an action look at the existing data model and see how it fits

<KevinMarks> this wiki looks a bit blank: https://www.w3.org/wiki/Web_Annotations

fjh: we'll get back in touch before our next call

<AnnBassetti> ACTION: James Snell will look at web annotation model and figure out how it maps into social work and Activity Streams [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action08]

<trackbot> 'James' is an ambiguous username. Please try a different identifier, such as family name or username (e.g., jsnell, jktauber).

<bjdmeest> https://www.w3.org/annotation/wiki/Main_Page

azaroth: I'll take the corresponding reverse action to look at Activity Streams 2.0

<azaroth> ACTION: azaroth to look at the Social WG FPWD [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action09]

<trackbot> Error finding 'azaroth'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

Arnaud: where are you in terms of legacy, backwards compatibility problems?
... in social WG we have some issues around this -- that's why it's Activity Streams 2.0

<tantek> FYI: here is a permalink to the start of our joint meeting and minuting between #social and #annotation: http://socialwg.indiewebcamp.com/irc/social/2014-10-28#t1414535621356

<elf-pavlik> ACTION: jasnell to look at web annotation model and figure out how it maps into social work [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action10]

<trackbot> Error finding 'jasnell'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<Loqi> AnnBassetti meant to say: ACTION: James Snell will look at web annotation model and figure out how it maps intosocial work and Activity Streams

<elf-pavlik> ACTION: jsnell to look at web annotation model and figure out how it maps into social work [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action11]

<trackbot> Created ACTION-8 - Look at web annotation model and figure out how it maps into social work [on James Snell - due 2014-11-04].

azaroth: we're in pretty much the same situation -- two groups originally that joined forces in 20xx and published a joint spec for data model and serialization (no APIs)
... we have free rein for the HTTP API and client-side/server interactions
... looked at several platforms, including Annotea, which we eventually abandoned

<azaroth> http://www.w3.org/2001/Annotea/

<Zakim> sandro, you wanted to ask about existing deployed protocols

sandro: are there protocols that are currently being used being by products in this space?

azaroth: nothing beyond basic REST APIs afaik
... multiple ways of doing search

<tantek> in the social realm, we have interop across many implementations with Webmention http://indiewebcamp.com/webmention

<tantek> in terms of a protocol

sandro: we'd like to know what you need from the protocol in this context?

Evanpro: we need to wrap

*: thanks all round

<AnnBassetti> talky guys ... I'll reconnect in other room

<elf-pavlik> #social as:<3 #annotation ;)

<EdK> scribenick EdK

<Arnaud> https://www.w3.org/wiki/TPAC2014#Session_Grid

<EdK> Wed morning 9:15 am unmeeting with schema.org

<EdK> jasnell: opensocial embedded experience was an input

<EdK> primarily built around a gadget model

<EdK> jasnell: not sure anyone is interested in continuing to use the gadget model

<EdK> potential actions like a share is something could to with content

<EdK> then a set of actions of what should happen

<EdK> for example call out a deep link in a mobile app

<EdK> a simple action like Share could be done with simple declarative syntax

<EdK> the user experience would be a buoon - a share or like button

<EdK> then it could open a view, like a browser view

<EdK> the media type will define what is embedded, like a video

<EdK> embedding a script tag is far more common

<EdK> jasnell: where different actions might be done, open a browser view, a native mobile action,

<EdK> want to embed with the content so the . can choose which to use

<EdK> it would be that the implemetner could leverage the options, or could ignore them and do their own thing

<EdK> these are the publisher's options, the things teh publisher says you could do with it

<EdK> decoupled this, comapred to Embedded Experience in OpenSocial which was tightly tied

<EdK> the details about whatr you are embedding isn't part of this

<EdK> could be web component, iFrame, scripts, a number of different ways

<EdK> in current editors draft is an "expects" property, which is a variety of additional metadata

<EdK> but this overlapw with other work like hydra

<EdK> dow we want to define or leverage other prior work

<EdK> can designate the sandbox policy

<EdK> the activity vocabulry defines the model for this embedded view, it does not define the share action

<EdK> reference yesterday can verbs be just nouns, would make this quite a bit easier

<elf-pavlik> linke to presentation?

<EdK> can have an activity statement like a share, currently activity being modeled different than a potential action

<elf-pavlik> http://tpac.mybluemix.net/as2.html#19 ?

<jasnell> http://tpac.mybluemix.net

<elf-pavlik> jasnell, cool! thx

<EdK> Sandro: difference wouldn't be noticed by user

<EdK> jasnell: right, the implementor determines

<EdK> this is something that needs to be worked out, should there be a ui ...

<EdK> jasnell - basically what potential actions can be done aganst a containing object, it doesn't matter what the object is (images or whatever)

<EdK> query the data store gets the actions along with it. the action data travels along with it.

<EdK> Sandro: how close to running code?

<EdK> jasnell: some is running now

<EdK> evanpro: support 3 basic actions share like and comment, this could be a big payload to pass along, perhaps not attach to each and every payload

<EdK> an event with rsvp options, a poll that could be replied to, games, whater

<EdK> s/whater/whatever/s

<Loqi> EdK meant to say: an event with rsvp options, a poll that could be replied to, games,whateve

<EdK> some security issues with accepting a url from someone else, action syas use this to open something else, etc.

<elf-pavlik> cwebber2 do you get audio/video?

<cwebber2> I'm getting audio, no video

<cwebber2> on the room

<cwebber2> I see you

<AnnBassetti> I had to reconnect .. lost you

<EdK> jasnessl: no standard way to do a Like, if we had a standard, thenw e could pass it without the identifiers. but there are too many to accomplish, so could do the basics and then detail teh others

<AnnBassetti> oh yeah? I don't see you

<AnnBassetti> more important you see the room

<cwebber2> I don't see the room but I see elf-pavlik

<AnnBassetti> rats

<cwebber2> AnnBassetti: thanks for handling the recording btw, greatly appreciated

<cwebber2> I'll reconnect

<AnnBassetti> my pleasure; it's a drag to be remote

<EdK> another alternative is to have a thing we can do like, share, comment. a get would say what these actions are

<EdK> could use whitelisting, haven't found a better solution yet

<cwebber2> there we go, looks fine now

<AnnBassetti> roger dodger

<AnnBassetti> except I don't understand why I have split screen

<AnnBassetti> oh well

<EdK> dret: I see repeating these actions over and over could be noisy, but if push to type, then can teh client infer that. if not, then they can't tell waht action to do for an activity.

<Zakim> tantek, you wanted to note existing implementations

<EdK> or ATOM pushes something into the feed, discussed that its been done in several ways, need to solve so we don't attach 20 actions and send around

<elf-pavlik> dret, have you looked at http://www.hydra-cg.com/spec/latest/core/ ?

<tantek> indiewebcamp.com/webactions

<EdK> tantek: in webcamp have a lot of experience, not theoretical, done over more than a year

<EdK> displaying page noted in irc

<Loqi> AnnBassetti meant to say: oh well

<Loqi> it'll be ok

<AnnBassetti> hmm ... now why?

<EdK> tantek: webintents had issues that translated into UI nightmare

<EdK> came up with a set of common actions, this is what people are deploying today.

<EdK> only 2 attributes "do", "with"

<cwebber2> webrtc seems to work well as long as you're willing to reconnect ;)

<EdK> even if you have no support for webactions, the page will still work

<KevinMarks> link for Tantek's discussion http://indiewebcamp.com/webactions

<EdK> if you are handling an indie action, user has specified what to do with the action

<EdK> demonstrated a few approaches linked from the above noted page

<EdK> berry frost used text and graphic/icon as a good use case example

<EdK> took some webcomponent magic, he used CSS

<EdK> demonstrated fallback action to twitter when not defined

<EdK> evanpro: when using feedreader approach, can you operate on teh actions?

<EdK> tantek: is inside the hentry, currently wouldn't do something. certainly doable erhaps as simple as one class name

<KevinMarks> you mean e-action tantek?

<EdK> Sandro: what about if you dont' do what evan just asked:

<EdK> tantek: using register protocol handler. can register on your site. the site you are browsing doesn't know that. asks if you have a handler, if not it will fall back

<EdK> discussed "register the web action protocol" as teh method

<Zakim> tantek_, you wanted to point out indiewebcamp.com/webactions and what's been deployed

<EdK> jasnell: the action handler work, jasmes sees lots of overlap. the action handler type prevents more options, but is much simpler. roughly equivalent approaches, support of multiple fallback options, dealing with verbs, don't see thse as too far off

<EdK> tantek: doesn't say how to handle the verb. you can have an indei-action tag without a fallback

<EdK> s/indei/indie/s

<Loqi> EdK meant to say: tantek: doesn't say how to handle the verb. you can have anindi-action tag without a fallback

<EdK> evanpro: issue he'd liek to address, where will we handle in our schedule and process. ack that James is already working on. is a llot for the working group to address. getting the social api and teh federation protocol may want to happen earlier than in-browser experiecne

<cwebber2> +1 to federation / social stuff coming first... I do think actions stuff is something that will be critical to a lot of people, but federation stuff is more critical I feel

<tantek> http://indiewebcamp.com/webmention

<EdK> would like to talk about whether to include these actions. preference is to push off until after federation protocol, then can discuss using these actions in a federated environment.

<tantek> I meant to say:

<EdK> timing discussed, feeling is that should be toward end of process

<EdK> Arnaud: asked if these use cases mean we do have use cases.

<EdK> jasnell: has an editor's draft, will continue to work on it. on road to FWD doesn't mean have to talk a lot about it. suggests Tantek and James discuss. would like to at least look at in draft, even if it isn't going to end up being prioritized. see actions as an entrypoint into the API discussions

<EdK> provides use case insight

<EdK> ARnud: is concern about resources and would be delaying things, or is this about dependancies

<EdK> evanpro: both. limited bandwidth to apply to issues, schedule of others on teh roadmap will define dependencies. roadmap is on page, but not ordered in teh charter.

<Zakim> tantek, you wanted to suggest federation protocol before social API

<EdK> jasnell: not in a rush to be finished for this work, is a priority in own company. won't let it derail WG efforts, if it has to wait it will. want to start it going.

<EdK> tantek: would put federation protocol before api based on experience in his work.

<KevinMarks> +1 on webmention maturity

<EdK> maybe tehre are different scopes in mind for federation, but wrt webmention, seems mature enough to go somewhere. freinding, following, pushing out live updates is an area that needs a lot more work though.

<EdK> would like to see tha actions draft contain the superset, then can point to what is being deployed on the web today

<cwebber2> I think that part of federation stuff is foundational enough that I don't think federation can be considered partially done

<Zakim> evanpro, you wanted to ask if "action" suffices for our embedded experience API

<EdK> jasnell: propose to put heads together on these approaches, come up with one that will work.

<KevinMarks> cwebber2: you mean federation is unitary? Can't be incrementally implemented?

<EdK> evanpro: from charter point of view, if action indicators are in social data syntax, would that suffice for embedded experiecne api or would there be more work to do there.

<EdK> harry: problem is too many working drafts will mean can't finish. have to fit together. darwinian process... supports the general activity streams, api, federation, happy with embedded experience work toward the end

<cwebber2> KevinMarks: I'm having a hard time understanding how federation where subscribing/pushing updates/commenting/liking are not all clearly enough implemented where there is a solid enough of a foundation to do much else

<cwebber2> maybe I'm wrong, but

<EdK> jasnell: doesn't want to confuse Open Social Big E embedded experience with the work James is doing about little e experiences.

<EdK> sugarCRM doesn't want to do open social gadgets work anymore either,

<KevinMarks> the micropub/webmention split seems quite good in practice for update/notify

<EdK> harry: need http api out soon, federation. need to get webmention as working draft, no problem also working on embedded draft too.

<KevinMarks> subscription and updates is tricker though PubSubHubbub is a marker there as discussed yesterday

<EdK> main holdup will not be lack of good concepts, will be lack of editor's time in drafts. would focus on http api first.

<cwebber2> I disabled my video for a sec :)

<EdK> clarification there is no order in the charter, but there is a schedule. but it isn't a big deal to change that and inform people of there is a better order.

<cwebber2> sorry for the confusion

<cwebber2> thx for staying on top of it, AnnBassetti

<EdK> evanpro: feels federation is harder,

<KevinMarks> the indie-action technique could be a gadget replacement, but with user-chosen gadget binding

<EdK> jasnell: there is no proposal for the api draft yet, what is the plan to get on teh table for discussion.

<EdK> harry: need to get all tof them on teh table as soon as possible

<EdK> suggests everyone put out editr's draft as soon as possible as teh mechanism to drive timing and forward progress

<EdK> Arnaud: recapped the rather lengthy process to get through all the drafting and review processes, test suites, implementations and reportback of results. putting drafts is lower hanging fruit but the working group has a lot of things to do and can't push a lot of parallel thigns through

<EdK> harry: not concerned about meeting chrter, more worried on time

<EdK> tantek: ok with seeing multiple drafts to get many more eyes on it, even if it is multiple approaches.

<elf-pavlik> +1 staying open to multiple drafts

<EdK> evanpro: could we do with proposals to drafts, winnow down to effort

<EdK> tantek: can see multiple drafts expanding collaboration regarind potentially independent use cases.

<KevinMarks> http://indiewebcamp.com/SWAT0

<EdK> sandro: re SWAT0, asked the other day about whether it was implemented in indeweb

<elf-pavlik> ProjectDanube?

<EdK> evanpro: did in 3 implementatiosn, statusnet, rstatus, another I missed :-)

<evanpro> buddycloud

<KevinMarks> I say it is implemented; Tantek is quibbling about what "tagging somone in a photo" means

<elf-pavlik> http://buddycloud.com/ (XMPP based)

<EdK> AnnB: so you think it is achievable?

<EdK> tantek: yes.

<EdK> harry: coming to the end of the day. want to finalize with clear editors.

<tantek> http://indiewebcamp.com/micropub as well for a social API

<elf-pavlik> possibly Hydra and Linked Data Fragments !

<EdK> evanpro: seems like there are 3 pump.io, micropub, ....?

<tantek> see and add to https://www.w3.org/wiki/Socialwg/Social_API_candidates

<EdK> thanks tantek

<EdK> Arnaud: want to use last 1/2 hour to wrap up. are we done with actions discussion at the moment?

<EdK> KevinMarks suggests indieweb did (mostly) meet SWAT0

<EdK> Arnaud: great meeting over last 2 days

<EdK> what is important is have clear idea who needs to do what

<EdK> some actions notes, some issues not noted formally

<EdK> anything not related to a specific draft, broader should be recorded in tracker.

<tantek> jasnell: indiewebcamp.com/webactions and indiewebcamp.com/indie-config

<EdK> review different topics agree what to do next

<EdK> activity stream specs - are we clear?

<EdK> jasnell: next step is to get feedback.

<EdK> Arnaud: need 2 people to commit to reading and commenting on specs.

<EdK> Arnaud: are we near last call yet?

<EdK> jasnell: no

<EdK> jasnell: need one more working draft to add namespace, resolve questions like verbs as nouns, may need one more working draft early december to wrok out that. then can work on starting to do some tests and test suite

<EdK> evanpro: what would a test suite be evaluateing:

<EdK> jasnell: make sure parsing is handled the same way?

<elf-pavlik> o/

<EdK> Arnoaud: who is planning implementing - given all the other work.

<KevinMarks> is http://tools.ietf.org/html/draft-snell-link-method-10 meant to be a webmention equivalent, jasnell ?

<EdK> Sandro: what is implementation ...

<EdK> evanpro: would do parser for consuemr, but what else?

<EdK> jasnell: has some opensource, library not an application. java and javascript parsing and consuming

<AdamB> jasnell: where is the opensource library at today?

<EdK> need to have a JSON-LD to fully handle. All valid AS1.0 must be consumed.

<cwebber2> thx AnnBassetti

<EdK> tantek: use cases should drive tests

<EdK> for an implementation

<sandro> ISSUE: Are AS consumers REQUIRED to understand the pre-JSON-LD syntax?

<trackbot> Created ISSUE-7 - Are as consumers required to understand the pre-json-ld syntax?. Please complete additional details at <http://www.w3.org/Social/track/issues/7/edit>.

<EdK> cwebber2: right now implementing at mediagoblin, suggests will have more than 1 implementation likely.

<EdK> Arnaud: may be boring but figuring out what spec is first, helps people know what to implement.

<cwebber2> thx for handling that AnnBassetti

<EdK> tantek: finding existing implementations can make sure the call goes out to find out who will then implement the spec

<KevinMarks> we should invite Echo and Gnip to comment as they have existing AS implementations

<jasnell> KevinMarks: +1

<cwebber2> oh I should be off the queue

<EdK> harry: need to recognized that some implementers tend to wait.

<cwebber2> logical guess of queue management command worked ;)

<tantek> this is what I'm talking about: https://www.w3.org/wiki/Activity_Streams#Implementations

<EdK> Arnaud: one thing practically that can be done, sooner we have test suite sframework set up, tell people what a test looks like, then people can submit tests and therefore implement a test suite

<evanpro> ACTION: test suite for Activity Streams 2.0 [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action12]

<trackbot> Error finding 'test'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<elf-pavlik> I started already with testing RDF from examples in the spec draft, if that counts as useful test...

<evanpro> ISSUE: test suite for Activity Streams 2.0

<trackbot> Created ISSUE-8 - Test suite for activity streams 2.0. Please complete additional details at <http://www.w3.org/Social/track/issues/8/edit>.

<EdK> evanpro: what is action - to create test suite for AS2.0? who gets teh action? better create an issue

<wilkie> oh yay. good.

<wilkie> :)

<EdK> evanpro: since don't have a vocabulary, haven't adopted AS1.0, need an action.

<evanpro> ACTION: jasnell adapt Basic Schema from Activity Streams 1.0 to 2.0 [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action13]

<trackbot> Error finding 'jasnell'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<elf-pavlik> https://github.com/jasnell/w3c-socialwg-activitystreams/pull/32 first version of tests for expected RDF

<wilkie> tantek: that implementation list doesn't even have statusnet or rstatus on it heh

<evanpro> ACTION: evanpro share verbs and object types from pump.io with jsnell before adapting basic schema [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action14]

<trackbot> Error finding 'evanpro'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.

<tantek> wilkie heh - please edit it! :)

<wilkie> evanpro: try jsnell

<evanpro> ACTION: jsnell adapt Basic Schema from Activity Streams 1.0 to 2.0 [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action15]

<trackbot> Created ACTION-9 - Adapt basic schema from activity streams 1.0 to 2.0 [on James Snell - due 2014-11-05].

<wilkie> tantek: will do *salutes* :)

<tantek> wilkie++ thank you!

<Loqi> wilkie has 3 karma

<elf-pavlik> s/identity/profiles/ ;)

<evanpro> ACTION: eprodrom share verbs and object types from pump.io with jsnell before adapting basic schema [recorded in http://www.w3.org/2014/10/28-social-minutes.html#action16]

<trackbot> Created ACTION-10 - Share verbs and object types from pump.io with jsnell before adapting basic schema [on Evan Prodromou - due 2014-11-05].

<Loqi> elf-pavlik meant to say: https://github.com/jasnell/w3c-socialwg-activitystreams/pull/32 first version of tests for expected RDF

<elf-pavlik> who speaks?

<KevinMarks> hadleybeeman is speaking, elf-pavlik

<EdK> hadleybeeman: introduced herself. story about needing to make crx very clear in the beginning, to smooth discussions with the director

<KevinMarks> test or validation?

<EdK> jasnell: AS1.0 did not define conformance. now with implementations can define within test sutie for 2.0

<hadleybeeman> ...to smooth discussions with the director and to avoid discovering unexpected confusion later on.

<EdK> s/sutie/suite/s

<Loqi> EdK meant to say: jasnell: AS1.0 did not define conformance. now with implementations can define within testsuit for 2.0

<wilkie> KevinMarks: "test or validation" wrt AS?

<EdK> harry: hard to test based on user experience, so parser is default

<sandro> sandro: just testing whether the stream has valid syntax is not a test of a consumer.

<EdK> Arnaud: lets keep on going with actions, it is useful. haven't said anything about dederation, should we ahve an action

<EdK> s/dederation/federation/s

<Loqi> EdK meant to say: Arnaud: lets keep on going with actions, it is useful. haven't said anything aboutfederatio, should we ahve an action

<KevinMarks> sandro, so some kind of round-tripping?

<tantek> https://www.w3.org/wiki/Socialwg/Federation_candidates

<Loqi> sandro meant to say:sandro just testing whether the stream has valid syntax is not a test of a consumer.

<sandro> KevinMarks, I don't really have much idea, other than manual testing. It seems like a really hard problem.

<cwebber2> AnnBassetti: I don't now, I disconnected and reconnected

<cwebber2> thanks :)

<hhalpin> consumers can be judged by number of users, which can be documented rather easily if there are products with X customers or in open-source products where you can list number of subscribers/users.

<evanpro> ISSUE: need candidates for federation protocol

<trackbot> Created ISSUE-9 - Need candidates for federation protocol. Please complete additional details at <http://www.w3.org/Social/track/issues/9/edit>.

<evanpro> ISSUE: need candidates for Social API

<trackbot> Created ISSUE-10 - Need candidates for social api. Please complete additional details at <http://www.w3.org/Social/track/issues/10/edit>.

<tantek> https://www.w3.org/wiki/Socialwg/Social_API_candidates

<hhalpin> So, if we have a parser for a project with no subscribers or a product with no estimated number of users, then it doesn't count in my book.

<KevinMarks> is micropub also federation?

<hhalpin> That's why W3C didn't want to launch group till we had some clear products that were interested in the space.

<elf-pavlik> reminder: tomorrow meeting with schema.org - https://www.w3.org/wiki/TPAC2014/SessionIdeas#Schema.org_and_Social_WG

<hhalpin> Since its harder to tell with open-source projects

<hadleybeeman> hhalpin: that all looks fine to me. I'm not fussed how you measure conformance or what the CR exit criteria are — just don't want you all to end up frustrated with this later!

<hhalpin> micropub does indeed count as a federation.

<cwebber2> I need to drop off

<cwebber2> sounds like things are wrapping up anyway

<EdK> annB says her browser crashed

<cwebber2> thank you, all!

<hhalpin> I'm also totally happy with things not going to Rec if they don't have users.

<cwebber2> AnnBassetti: and thanks again for handling the recording

<elf-pavlik> cwebber2, come back for picture!

<elf-pavlik> AnnBassetti, no need!

<elf-pavlik> we take one next summer in Europe :)

<EdK> Arnaud: closes meeting, time for pictures

<elf-pavlik> great job everyone!!!

<hhalpin> trackbot, end meeting

<EdK> cwebber2, get back on video, eh?

<EdK> Ann'll take screen shot

<cwebber2> oh

<cwebber2> okay

<cwebber2> EdK: just in time :)

This is scribe.perl Revision: 1.139  of Date: 2014-08-26 14:08:24  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/<a couple other points>/Activity Streams 2.0/
Succeeded: s/SWOT 0 (??)/SWAT0/
Succeeded: s/SAT/SWAT/
Succeeded: s/consming/consuming/
Succeeded: s/question: yes,/elf's question: yes,/
Succeeded: s/<name?>/James Tauber /
Succeeded: s/micrpub/micropub/
Succeeded: s/follow the discussion./follow the nuance of the discussion.  But I think it's important to document the consensus of the group./
Succeeded: s/in include/to include/
Succeeded: s/person/persona/
Succeeded: s/right/write/
Succeeded: s/system/system implementation/
Succeeded: s/tempting /attempting /
Succeeded: s/simetric /symmetric /
Succeeded: s/Arnaut/Arnaud/g
Succeeded: s/Is your federated system considered an end to end type of relation?/Is your federated system considered an n to n type of relation?/
WARNING: Bad s/// command: s/!!!!/Topic: Joint meeting Social/Annotation WG
Succeeded: s/!!!!/Topic: Joint meeting Social-Annotation WG/
Succeeded: s/???/James Snell/
FAILED: s/??e/Evanpro/
Succeeded: s/julyt/july/
Succeeded: s/???/Aaron Parecki/
Succeeded: s/fjh/someone/
Succeeded: s/probably a throwaway comment we can ignore/not sure that we have an answer for this/
Succeeded: s/social work/social work and Activity Streams /
WARNING: Bad s/// command: s/whater/whatever/s
Succeeded: s/Adam pushes/ATOM pushes/
WARNING: Bad s/// command: s/indei/indie/s
WARNING: Bad s/// command: s/identity/profiles/ ;)
WARNING: Bad s/// command: s/sutie/suite/s
WARNING: Bad s/// command: s/dederation/federation/s
Succeeded: s/sandro:/sandro,/
Found ScribeNick: Annb
WARNING: No scribe lines found matching ScribeNick pattern: <Annb> ...
Found ScribeNick: AnnBassetti
Found ScribeNick: AnnBassetti_
Found ScribeNick: evanpro
Found ScribeNick: harry
Found ScribeNick: Ann Bassetti
WARNING: No scribe lines found matching ScribeNick pattern: <Ann\ Bassetti> ...
Found ScribeNick: AnnBassetti
Found ScribeNick: rhiaro
Found ScribeNick: nickstenn
Inferring Scribes: Annb, AnnBassetti, AnnBassetti_, evanpro, harry, Ann Bassetti, rhiaro, nickstenn
Scribes: Annb, AnnBassetti, AnnBassetti_, evanpro, harry, Ann Bassetti, rhiaro, nickstenn
ScribeNicks: Annb, AnnBassetti, AnnBassetti_, evanpro, harry, Ann Bassetti, rhiaro, nickstenn

WARNING: Replacing list of attendees.
Old list: Salon2
New list: [IPcaller]


WARNING: Replacing list of attendees.
Old list: [IPcaller]
New list: Salon2 elf-pavlik

Default Present: Salon2, elf-pavlik
Present: Salon2 elf-pavlik jeff

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 27 Oct 2014
Guessing minutes URL: http://www.w3.org/2014/10/28-social-minutes.html
People with action items: 1.0 activity adapt arnaud azaroth basic eprodrom evanpro from hhalpin identify james jasnell jsnell kevinmarks kmarks2 schema see streams test