- DRAFT -
Social Web Working Group Teleconference
24 Mar 2015
See also: IRC log
- AdamB, ben_thatmustbeme, Ann, +1.514.554.aaaa, tantek, eprodrom, Arnaud, jasnell, cwebber2, Sandro, rhiaro, elf-pavlik
<trackbot> Date: 24 March 2015
I can scribe
since I can't talk anyway
<eprodrom> scribenick cwebber2
eprodrom: first order of business for today is to approve the minutes from 10th of march, unfortunately those minutes seem to not be prepared
... tantek, do you remember who scribed?
tantek: let's see if I took notes
... looks like it was aaronpk
eprodrom: great, I am going to open up an action
<eprodrom> ACTION aaronpk upload the 2015-03-10 minutes to the wiki
<trackbot> Created ACTION-55 - Upload the 2015-03-10 minutes to the wiki [on Aaron Parecki - due 2015-03-31].
eprodrom: okay unfortunately we can't confirm those
... hopefully we can move on
... next item is approval for minutes for face to face
... harry was putting those together but is not on the call
... I will do another action on harry to assemble the minutes from the face to face
<eprodrom> ACTION harry Assemble the minutes from March F2F
<trackbot> Created ACTION-56 - Assemble the minutes from march f2f [on Harry Halpin - due 2015-03-31].
AnnB: the minutes from the 10th of may are messed up
... minutes from the 17th, sorry, not 10th
cwebber2: probably my bad
AnnB: the irc log is correct, but the w3c has a bot that makes the minutes, this is the part that's screwed up
... something seems wacky
... sandro, do you know more?
sandro: don't know, sorry, I think harry said he will clean them up manually
... there's a problem where if a group meeting goes past midnight
<eprodrom> rhiaro: can you mute please? I think you're getting some feedback
AnnB: oh no the midnight problem
sandro: harry claimed that problem
<Loqi> Eprodrom made 1 edit to Socialwg/2015-03-24 https://www.w3.org/wiki/index.php?diff=83096&oldid=83050
<Loqi> Tantekelik made 1 edit to Socialwg/2015-03-10 https://www.w3.org/wiki/index.php?diff=83097&oldid=82751
<Loqi> Tantekelik made 1 edit to Socialwg/2015-03-24 https://www.w3.org/wiki/index.php?diff=83098&oldid=83096
eprodrom: I think harry's on minutes from the march f2f, harry's on march 10th, so hopefully that can be done
<AnnB> past midnight <somewhere in world> .. can't remember .. (we certainly did not meet past midnight in real life)
eprodrom: minute approving is everyone's favorite part but we gotta skip it
... next meeting I think we have no conflicts, arnaud will chair
... that brings us to the next issue on the agenda which is the next f2f
<tantek> I've added links to our IRC logs for the f2f and redlinked where the minutes will go here: https://www.w3.org/wiki/Socialwg/2015-03-17#Social_Web_WG_Face_to_Face_Meeting_at_MIT_.28F2F2.29
eprodrom: we had an idea come up in the f2f where we definitely will do a f2f at the next tpac in september, but would like to another between now and then
... there are a couple of candidates for times that would work, one is the upcoming meeting in paris, another is indiewebcamp in edinborough at the end of july
eprodrom: we have an open doodle poll with unfortunately the wrong dates, if you responded before you may need to do it again
eprodrom: I'm not sure, do we want to close this poll at this point?
... the big issue is we need to make a decision soon, if people make travel plans to paris in may, they need to do it in a week or two if not this week
... would like to close this by the end of the week
Arnaud: I'd like to move faster on it, I'm one of the people who have to make travel plans, it's hard to keep waiting
... there's not much debate, the main option seems to be the july one
... I'd be inclined to propose that we settle on the main dates in paris
... but do we have a host in paris?
<AnnB> no, cwebber2 .. the main option is MAY
<rhiaro> harry said he could book Centre Pompidou
Arnaud: (correction: the main option is in may)
cwebber2: thanks annb
Arnaud: harry said there was a place we can go to
... we can confirm later
... I encourage others to respond right now
... it seems to be a pretty clear cut to me (based on yes/nos)
AnnB: even if there was an informal gathering, seems worth it
Arnaud: maybe we can give a few minutes, but then we can do 5 minutes then we can close on this?
... but I would like to have a decision
tantek: I think you need a confirmation from the host
Arnaud: we can close on that pending confirmation from harry
tantek: yes we can do that
eprodrom: yeah, so we'll put this off to the end of the call so all of you can open your airline of choice in another browser window and see if you can make it oo the meeting. Let's see what we can do. I will push it off to the end of the hour, ok harry?
scribe: (okay arnaud)
Tracking of Actions and Issues
eprodrom: we have some perpetual issues
... any activity on open issues/actions that we can close today?
... take a moment please to look at them and see what need to be closed
- click clack*
scribe: okay if we have nothing
tantek: looks like on 26(?) that I think that's been merged
<jasnell> tantek++ for PR
tantek: so now it's back to see about more pull requests, thanks james for that merge
<Loqi> tantek has 170 karma
tantek: will keep going to see how far I can get with the examples
<trackbot> action-26 -- Tantek Çelik to Review microformats examples in AS2.0 specs -- due 2015-03-17 -- OPEN
eprodrom: harry's got an action open, 37 for testing
<trackbot> action-37 -- Harry Halpin to Put forth the test suite plan using standard js tools -- due 2015-02-10 -- OPEN
eprodrom: i've kind of taken this on myself, I'd like to edit this action and have it defer
... I think there's no easy way to say it's closed
<trackbot> action-50 -- Evan Prodromou to Extract the examples from the main documents, pick properties and put together a test for those examples -- due 2015-03-24 -- OPEN
eprodrom: because action 50 supercedes it
... but i will write it in the notes
cwebber2: hopefully me too :)
eprodrom: any other actions we need to get moving?
... is henry on?
<eprodrom> bblfish: Are you on the call?
eprodrom: henry / bblfish mentioned doing ontology examples for the activitystreams testing
<bblfish> ah I am on another call too
eprodrom: it would be great to see what tools we could use
<bblfish> pretty complex
eprodrom: oh okay sorry about that bblfish
<Loqi> Tantekelik made 1 edit to Socialwg/2015-03-17 https://www.w3.org/wiki/index.php?diff=83099&oldid=83035
eprodrom: given that I'd like to close up the open actions, we have a pretty big pile of issues on the agenda for today
<tantek> are we looking at http://www.w3.org/Social/track/issues/raised ?
eprodrom: I am going to try to get these, these are all on the f2f agenda, a few came over
<tantek> to decided whether to accept / open them or not?
eprodrom: we have a number for AS 2.0
<elf-pavlik> eprodrom, Raised issues?
eprodrom: I'd like to review our strategy for next steps for social api by end of the call
<tantek> I recall Arnaud saying we should review raised issues
Activity Streams 2.0
eprodrom: I count 1, 2, 3, 4, 5, 6 top level items here on AS 2.0
... jasnell brought up an issue around profiles
... maybe we should address it, let's throw it at the end
... let's start with the top one and move down
... we have a linked issue in actions, the issue is about as-link
... there's an action to gather options about media objects
<eprodrom> elf-pavlik: yes
eprodrom: elf-pavlik, are you on the call?
... okay elf-pavlik is on the call
... how are we doing on action 42
<elf-pavlik> who reviewed https://www.w3.org/wiki/Socialwg/MediaObject ?
<tantek> FWIW I don't know what use-case / user-story is used to justify as:link so I'm in favor of dropping it.
eprodrom: this is the issue on media object
cwebber2: btw, Tsyesika and I added info for mediagoblin
... eprodrom, fyi
<tantek> though I'm really confused by the examples in issue-14
eprodrom: is the question that we're going to change the activitystreams 2.0 object
... and what specifically will get changed?
... that's my main question
... if this work is started I'd like to close this
tantek: I think it's related but not the same
<trackbot> issue-14 -- as:Link adds a lot of complexity, if we keep it we need to clarify consequences of using it instead of as:Object -- open
tantek: issue 42 is related to issue 14 but not the same thing
<ben_thatmustbeme> action 42 was to give us background on issue-14 i believe
<trackbot> Error finding '42'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.
eprodrom: right, 14 is about links and 42 is about objects related to AS 2.0 vocab, I'm wondering why we clumped them together
tantek: I have a hypothesis
Arnaud: no good to speculate, elf-pavlik put it for the f2f, let's let him explain himself
<elf-pavlik> currently as:Link mostly serves as pattern for MediaObjects
tantek: let me quote from elf-pavlik in the related notes: the as2 spec only contains as:link or related info about media objects(?) so for other thing as-link not needed
<Loqi> tantek has 171 karma
tantek: so if we can remove as:link from media objects, then no need for as:link
<elf-pavlik> me sorry guys, i made sure to have good audio for today but then couldn't dial code on pad of this softphone :(
eprodrom: excellent point, but as:link particularly useful for link relationships, so the various link relationships defined for a number of items, so they're used in html as in on a tags and on link tags
... also used in atom
... a number of them used
... in activitystreams 1.0 used quite a bit
<jasnell> +1 for keeping
eprodrom: so the link element is pretty useful even if only in the as 2.0 system for media objects, I think they're important to keep in the system
tantek: that would help a lot
<elf-pavlik> i'll provide more examples where it makes processing much more convoluted
tantek: because for now we don't have that information
eprodrom: more convoluted than..... what is the alternative, not having it?
<elf-pavlik> imagine as:Link as member of Collection for example
eprodrom: any time we convey info in a system it's more convoluted than not having it in the system
... but then we don't have that info
tantek: I think elf was trying to provide examples of conveying info without as:link, but that's in related examples
<elf-pavlik> please review issue and wiki page and i'll get proper audo for next week!
eprodrom: okay, I'd like to ask jasnell, please take the floor and address this
jasnell: not totally sure I understand what elf-pavlik's argument is
... if we have as-link in there or not, we have to convey the same info
... so I don't see the argument.
... I guess I need more examples to work with
eprodrom: seems unlikely to close this issue today, I suggest we leave it open and move on
eprodrom: next item is also media object, it's a related issue
... if the issue has been updated I'm going to close this action since it's specifically about the wiki page
... would like to keep moving on
... we have action 44 which is about collections
... collection comparison
eprodrom: there's different ways to represent collections in as 2.0
<trackbot> action-44 -- Pavlik elf to Collection - compare AS2 design with LDP, Hydra, Schema.org etc. -- due 2015-03-17 -- OPEN
eprodrom: and it looks like the question is...........
<elf-pavlik> someone offered adding LDP example to the wiki but never happended
eprodrom: there's a collection comparison
... and with quite a few different collections
<Loqi> Eprodrom made 3 edits to Socialwg/2015-03-24 https://www.w3.org/wiki/index.php?diff=83102&oldid=83098
eprodrom: I think the idea is we'll change the collection structure in as 2.0
eprodrom: since the comparison is done I'll close this
tantek: elf-pavlik says no
<elf-pavlik> i would like if someone adds LDP examples
tantek: I htink
eprodrom: ok, can someone familiar enough with ldp examples to put it in the wiki page?
<elf-pavlik> also we can discuss using ldp:Container instead of as:Collection if we use LDP as part of API
eprodrom: ok, so... great.
... so can we or should we defer this?
... I'm going to close this option then we can move on with the rest of the call
... sorry, trying to balance out all the stuff we have here
<tantek> is erik here?
eprodrom: we're moving through issues but without actually resolving htem
... that may not be in our best interest
tantek: perhaps we should skip issues if person assigned/raised is not present
... we can't seem to be effective without their presence
<elf-pavlik> we CAN'T work only 1h a week during the call ...
eprodrom: and some of these have been raised for a long time. I'm not sure we're closer to closing them than when we raised them.
... perhaps part of the effort is to get them closed
<AdamB> do we need to get all the data in one place in order to close them ?
<AdamB> to have the right conversation about it
tantek: we can worry about that when closer to last call
eprodrom: yes but some of them are fundamental, the core vocab around activitystreams, it will be hard to close these without examples
... it will be hard to do an implementation without these stabilizing
... we should stabalize these if moving towards implementations and testing them
... I'll pass on these, maybe we can move on
... question of separating grammar/vocab, I believe erik brought it
... question on identity/agent/persona/account
<trackbot> issue-17 -- Identity, Agent, Person, Persona, Account etc. need clarifications -- open
<jasnell> we can discuss that one briefly
<elf-pavlik> also actor/author
<jasnell> the identity, agent, account, persona, profile bit
eprodrom: uhhh, previous conversations on... agent persona account profile?
... I'm not sure. Are these part of our vocabulary for activitystreams?
<jasnell> we have objects in the vocabulary that deal with these
eprodrom: I'm not sure we must address these if we aren't encoding them.
... we need clarification
<jasnell> Actor, Person, Identity, Role, etc
eprodrom: it would be nice if when we have items on the agenda, we have more stuff ready
<elf-pavlik> i proposed in reply to agenda email to move inference to next week
eprodrom: I'm sorry, unless anyone has something to say one of these, I have nothing to move us forward
... they are all complex issues
... maybe since we do have one person with voice on the call since it's this last one
... maybe we should discuss it
... jasnell discussed issue of structured profiles
<trackbot> issue-26 -- Representing profiles in Activity Streams 2.0 -- raised
jasnell: it's pretty simple, in many of our examples, social platforms we look at, the profile is pretty prominent
... but we lack profile in our vocabulary
... and it's not clear how to model that
... so taking a stand on user stories as example
... so maybe model profile as type of collection
... so tim added a profile, added items to it
... so maybe it makes sense to manage it as a collection
... so we could have a profile object
... we have an existing object, person
... is this sufficient as a profile
jasnell: recognizing that anyone can have multiple profiles, and profiles can represent community
eprodrom: I would strongly say that profile is a person
... maybe a person object, or maybe something more abstract like a product is a type of thing that maps very closely to a person, but with brands taking that organization role
... I think getting really abstract with profiles would be a bad thing
... just treating a profile as a collection of name/value pairs might be worse than a profile object
tantek: yes I'm curious what's driving this
jasnell: the user stories, I'm mapping the vocab to the user stories, to make sure those are coinciding
... so we have many objects in the vocab based on a number of use cases
... those are not mapped to the use cases themselves
... we have a distinct version of profile
... with no idea what a profile is
tantek: good point, I see sandro on the queue fyi
sandro: sure, so I think the profile as a container is kind of a problem
... I can't figure out the logic of what it would be the container of other than a reified object of (?)
... I mean, we could make it work but
... I think of a profile as a web page that has some info about the person or the object
... as long as we have ideas about the person or the profiles
... as long as we have different access control
... eg internal company profile and public profile
<elf-pavlik> I remember sandro discussing it on WebID list while ago
sandro: and whether it's json or h-entries or rdf triples, we can do any by a web page
... plus I have a problem with a profile being a person
... it's not
... it manifests in the case of a person having multiple distinct profiles
<rhiaro> +1 profiles != persons
sandro: it follows logically that there's different profiles, you have to distinguish between profiles and the person raising htem
eprodrom: multiple profile per user seems awesome, we have no user stories that manage that, but it's a big pain to manage
<Loqi> multipleprofiles has 1 karma
eprodrom: so you have multiple profiles on facebook?
<elf-pavlik> disussion from public-webid https://lists.w3.org/Archives/Public/public-webid/2014May/0103.html
cwebber2: (who's speaking aside from eprodrom ?)
sandro: there's a page that's stuff about yourself, and there's a page that others see
<rhiaro> multiple profiles is against facebook's TOS, but people do it because it's necessary for managing their social interactions
eprodrom: so you see that as multiple profiles?
tantek: no it's the same profile, but different acl's
<elf-pavlik> i know people who have multiple accounts on both facebook and twitter
eprodrom: it's complicated, I can see doing it, but it didn't come up in our user stories
... tbh I'd rather get something done than do something complicated
sandro: hr has a different profile than the person
eprodrom: I could see having multiple proifles as the general case, but it's def the most complicated case
<elf-pavlik> rhiaro, could you document how it worked in BBC with named graphs?
eprodrom: would be interesting to see if any apis we have can do it
... obviously it would be a reparesentation of the person in json
<eprodrom> person class in as 1.0
eprodrom: there's a person classin AS 1.0
... that's what I'm referring to
sandro: I might need two distinct profiles
... one for personal stuff one for work
... those are logically the same
eprodrom: that's possible, will go to the queue
jasnell: so I would note that google plus as an example, as a system for multiple profiles
... every google employee has an internal profile
... as well as public
<elf-pavlik> sandro, maybe worth adding it to https://www.w3.org/wiki/Socialwg/Social_API/More_user_stories
jasnell: every user can have multiple profiles
... one person with multiple identities
... kind of going back to original question, "how to model profile?"
<rhiaro> Also G+ stuff ties into youtube, with multiple channels connected to one user account
jasnell: in every one of these examples, a profile is kind of a container
... a variety of things which are sub-categories of the profile
... managed independently, who i give htese with
<elf-pavlik> can someone take ACTION to provide alternative to container based strategy proposed by jasnell ?
jasnell: same as g+, twitter, each of these has a container-like model which is there
... so it might not make sense to model it as a *collection*, but model it as a *container*
... the question is: if I'm going to model an activity as a container, what's the right way to model that in the syntax
<elf-pavlik> NOTE very relevant to work happening in Credentials CG http://opencreds.org/specs/source/identity-credentials/
jasnell: as for how the api is concerned, how will we use the profile in the ??
... as for whether someone has more than one of not, it's a basic modeling problem
AdamB: we don't see that as a web page we see that as a something with a different ???
<ben_thatmustbeme> This seems to overlap with profile and implies access controls on parts of profile Socialwg/Social_API/User_stories#Contact_Info
we have different profiles on multiple systems, one additional profile with multiple systems, the example is a main profile with multiple systems
scribe: so we can figure out how to move that between things
<elf-pavlik> tantek, any comments based on rel="me" experience?
scribe: we provide a bit of background on that, users can't change their hr data, but they can change other things
... so that's where it comes from
<Loqi> Abasset made 1 edit to Socialwg/Social API/Sorting user stories https://www.w3.org/wiki/index.php?diff=83103&oldid=82669
AnnB: I would give those same examples, in addition we have multiple tools inside the company with people wanting to set up profiles, eg people setting up profiles in sharepoint and ?
... as well as hr data
... and our challenge is to pull them together in a cohesive manner
... and establish their profile in one place, to handle it, how to represent it
<jasnell> I would also note that systems that allow me to use third party authentication to create an account generally use a multiple profile model
AnnB: and not recreateit over and over
... that's our issue about federation
<jasnell> for instance, I use facebook to log into spotify but my spotify profile is separate from my facebook profile
<jasnell> even tho they represent the same person
eprodrom: james, I just want to ask, assuming we have a class something like as1.0 person, why not just do CRUD on that
<eprodrom> CRUD on an AS 1.0 person
eprodrom: eg, setters/getters on json document
... put it, get it, delete it, etc
... why isn't that about what we want to do?
jasnell: if you look at first user story, you see that *foo* creates distinct profile adding hometown, blah blah
... how much to do actor kim updated actor kim, or does she update profile
<elf-pavlik> jasnell, can you please explain
jasnell: and how to update granularity of adding profile, phone number, occupation
<elf-pavlik> "@type": "foaf:name",
<elf-pavlik> "displayName": "Kim Smith"
jasnell: we want to add ? to the profile
... in whihc case we want profile as a unique thing
... eg adding profile to the object
sandro: and how to update those parts
<elf-pavlik> http://opencreds.org/specs/source/identity-credentials/ proposes PATCH
jasnell: and how to update spotify vs facebook profile
... same ID to log into multiple services
... each ID may have multiple services
... kim may be on both identi.ca/facebook, but they're distinct profiles
sandro: I've noticed on facebook change, people can like the avatar change
jasnell: so yes, how much granularity to represent? the actions in the activitystream
<elf-pavlik> rel="me" and http://schema.org/sameAs take same approach IMO
eprodrom: how to manage the person updates one profile properties
jasnell: I'm asking do we need to
eprodrom: okay, a good question, can we capture that particular issue?
... I'd probably be... that's a finer point than the "how do we represent changes to the profile" not just how to represent it
... would like to raise it as an issue
<eprodrom> ISSUE how do we represent changes to a profile in an Activity
<ben_thatmustbeme> dog is going nuts, you wouldn't be able to hear me
jasnell: both issues are important, how to represent profile vs profile changes
eprodrom: we are closing up on top of the hour and I want to represent changes
... okay to close this and have multiple changes on it
<tantek> I think profile is an area that can get ridiculous complex / deep end.
<tantek> it's an area that needs simplification, not more options
eprodrom: where we left social API at end of our F2F, we have a # of different approaches to APIs that we are coming at from different directions
<eprodrom> micropub, AS 2.0-based JSON-y system, LDP adaptation
eprodrom: one is micropub, another is an activitystreams 2.0 JSON'y system, another is an LDP implementation
<tantek> I thought the three were micropub, pump.io, LDP ?
<tantek> this is the first I've heard of "AS 2.0-based JSON-y system"
<tantek> ok oh
eprodrom: our interest was to solicit drafts
... json'y activitystreams 2.0 thing is probably pump.io updated to AS 2.0
<tantek> I'd expect all the API candidates to be somewhat compatible with AS2
<tantek> so I'm not sure how it's helpful to frame one in particular as "AS2.0-based"
eprodrom: we want to solicit drafts, I don't think the other chairs will disagree with me in that we will probably not persue if we don't get drafts in for other candidates
<eprodrom> micropub, pump.io API, LDP
<jasnell> I've implemented an example server that derived from an LDP implementation but uses a AS 2.0 interface. I'll be open sourcing it soon and will have an example endpoint public hopefully later this week
<tantek> maybe I'm being optimistic :)
eprodrom: okay, sorry tantek, I will frame it as pump api as 2.0
... appreciate the correction
... we are seeking documentation on one or more proposals
... in everyone's interest to persue on our own
... tantek, Arnaud feel free to correct me, but if we don't get a proposal, we will not persue
<tantek> btw drafts may simple normatively reference an existing *open* spec (license etc.) and provide examples
eprodrom: tantek, Arnaud, is this fair?
someone: yes, beep
tantek: yes that's what we discussed at the f2f as something to consider
... an easier way for people to consider
... than just w3c submission style
... that's nice but not required
<Arnaud> the someone was me but I have nothing to do with the beep :)
tantek: as long as you have openly referenced doc
... as long as ??? convinces submit early, submit often
eprodrom: haha, agree that makes sense
... I encourage others look at oshepherd's activitypump 2.0 submission
eprodrom: would love to encourage pump people to look at that
... for those who would like to look at that, talk to oshepherd, persue further
... my point on social api: make proposals sooner than later
<tantek> I'd like to see submissions that cite an existing implemented API proposal, and provide examples of how one or more user stories are done with the API
eprodrom: talk amongst yourself about getting htem working
eprodrom: the other issue is talking about f2f in paris
... check ???
... we could do it as a proposed, with +1 voting, sound reasonable Arnaud ?
<eprodrom> PROPOSED We have a meeting of the SocialWG in Paris May 4-5
Arnaud: yeah sounds so, let's either doodle poll, or +1
<ben_thatmustbeme> I think the doodle poll is fine
eprodrom: sound about right?
<ben_thatmustbeme> +0 (i cannot make it)
eprodrom: okay, so I think that carries
<tantek> (also cannot make it)
eprodrom: looks like we have a resolution
sandro: let's make sure eveyrone understands ????
... that they're waiving their rights to 8 weeks notice
sandro: because we haven't had 8 weeks notice
<eprodrom> RESOLVED We have a meeting of the SocialWG in Paris May 4-5
sandro: any objections please note
... otherwise will mark as resolved
<tantek> pending host!
sandro: and I will look for potential flights to paris
<tantek> (just a heads up folks )
<elf-pavlik> eprodrom++ cwebber2++
<Loqi> eprodrom has 9 karma
AnnB: let me just say that the advisory committee there is a discount at novotel for w3c
... and this discount qulaifies
... about $180/room
<Arnaud> yes, we need to get confirmation from Harry that he can host ASAP
AnnB: I'll figure out how to sen dlink
<elf-pavlik> please ask me for help with finding couchsurfing in Paris if interested!
AnnB: you have to send in a form
<tantek> a fax?
AnnB: you get breakfast and wifi with the price
... fax works also
... it's south of the whatever, the eifel tower
sandro: very central to paris
sandro: we have to figure out where it takes place
AnnB: and it's only good until
cwebber2: this was a crazy call
<Loqi> cwebber2 has 22 karma
<ben_thatmustbeme> cwebber2: ++
eprodrom: call to close, thanks for handling
<eprodrom> trackbot, end meeting
<Loqi> cwebber2 has 23 karma
<Arnaud> For more information about the hotel, please see:
Summary of Action Items
[End of minutes]