Social Web Working Group Teleconference
14 Apr 2015
See also: IRC log
- Ann, cwebber2, +1.978.537.aaaa, aaronpk, Arnaud, ben_thatmustbeme, wilkie, tantek, Sandro, rhiaro, jperrault, Tsyesika, jasnell, eprodrom, AdamB
- elfpavlik, harry
Summary of Resolutions
- IRC and email and wiki are our canonical communication channels and if there are dropped balls we handled them as needed. For example, concern that not everybody was reading the mailing list which is fixed by bringing up things in the wiki.
- Reject issue 20 due to no actual implementation using JSON text sequences.
- Open issue 25 as something that Social API proposals should answer.
- Close issue 30 without action.
- Accept pull requests 96, 95, 92.
- Use Person as profile object to address issues 26 27. Ok to re-open with new information when available, still open to multiple personas e.g. via extension.
- Remodel "Connect" and "FriendRequest" as object types per issue 28
- Keep Generator, deprecate Provider from AS2 per issue 32
<Arnaud> if people like that, I'm happy to change our page to be more like that
<trackbot_> Meeting: Social Web Working Group Teleconference
<trackbot> Date: 14 April 2015
<trackbot_> Date: 14 April 2015
<tantek> double the trackbots?
<wilkie> double trackbot double the fun
<aaronpk> rhiaro summoned the bots a little too well
<Arnaud> trackbot_ bye
<Arnaud> trackbot_, bye
<jasnell> It's the Doublemint Trackbot Twins
<cwebber2> yeah zakim didn't report anyone on the phone
<cwebber2> did not yet start eh
<cwebber2> there we go
<jperrault> I am on the call, was invited by jasnell
<cwebber2> nobody really cares clearly
<jasnell> thanks for joining @jperrault
<jasnell> I'll dial in in a few minutes
<ben_thatmustbeme> jperrault++ I hear you are working on the test framework
<Loqi> jperrault has 1 karma
<wilkie> I can scribe
<tantek> scribenick wilkie
<AnnB> if you're planning to be at F2F in Paris, but haven't put your name on the list... please do so: https://www.w3.org/wiki/Socialwg/2015-05-04
<rhiaro> scribe: wilkie
<ben_thatmustbeme> scribenick: wilkie
<cwebber2> vs mixedCase :)
<cwebber2> looks like everyone's ID
<ben_thatmustbeme> tantek, case sensative too
<aaronpk> methinks Zakim needs a little work
approval of the minutes
<jasnell> no objections
<rhiaro> oh also the march 10th ones are up now too
tantek: does anybody have objections, fixes, etc to the minutes?
<rhiaro> they aren't approved yet?
tantek: let's go ahead and approval the minutes
tantek: let's put the march minutes for next week. rhiaro, can you go ahead and put that for the next week for people to review them.
... next meeting is next tuesday, Arnaud will chair.
Arnaud: I cannot actually
tantek: is evan available?
evan: I can
tantek: ok. evan will chair next week and I will be backup
... that takes us to the next f2f which is Paris
... AnnB was noting that we are requesting people sign up and looks like there are 13 participants, 1 pending visas and 2 questionable on funding. that looks like a healthy size of a meeting
... any other questions about the f2f?
AnnB: john can't come, a grad student is coming and is interested in vocabularities.
tantek: if you could add him to the regrets page that would be helpful
AnnB: ok. he just joined, not sure to which group, but I will add him to the regrets page.
tantek: he is listed as a remote participant though.
AnnB: if he says he is dialing in, then he is dialing in.
tantek: any other questions?
... I'll be in paris, but I have a conflict and won't be able to even remotely participate
... that's it for the f2f... didn't see anything else come up
... next is tracking actions and issues
<ben_thatmustbeme> tantek, in order to not take up too much of jperrault's time can we skip tracking of issues until after he gives his demo
<trackbot> issue-18 -- We need to know if there are any other products in this space and if there are any dependencies between ldp and opensocial planned in the future? -- raised
<trackbot_> issue-18 -- We need to know if there are any other products in this space and if there are any dependencies between ldp and opensocial planned in the future? -- raised
<tantek> bye trackbot_
Actions and Issues
<aaronpk> gtfo trackbot_
tantek: who raised this issue?
... harry are you on?
AnnB: he has the flu
<cwebber2> sorry, got disconnected
<cwebber2> reconnecting in a sedc
<jasnell> +1 to closing
tantek: I'm going to propose summary closure of this issue unless somebody has a good reason to keep it
<ben_thatmustbeme> +1 to reject
<AnnB> (helpful to read email for last minute statuses :-)
tantek: looks like a personal todo list item
... let's go ahead and close it
<trackbot> issue-19 -- WG communication channel explosion -- raised
<trackbot_> issue-19 -- WG communication channel explosion -- raised
<ben_thatmustbeme> we are avoiding the issue
tantek: (laughs) is this part of the communication problem that we can't communicate on this issue
<jasnell> happy with letting the chairs figure this issue out
evan: my feeling on this issue is that IRC and email and wiki are our canonical communication channels and if there are dropped balls we handled them as needed
<jasnell> don't think it's worth opening the issue
evan: for example, concern that not everybody was reading the mailing list which is fixed by bringing up things in the wiki.
<ben_thatmustbeme> I second jasnell on that one
<tantek> Proposal: IRC and email and wiki are our canonical communication channels and if there are dropped balls we handled them as needed. For example, concern that not everybody was reading the mailing list which is fixed by bringing up things in the wiki.
<tantek> RESOLVED: IRC and email and wiki are our canonical communication channels and if there are dropped balls we handled them as needed. For example, concern that not everybody was reading the mailing list which is fixed by bringing up things in the wiki.
<trackbot> issue-20 -- Represent Collections using JSON Text Sequences (RFC 7464) -- raised
<trackbot_> issue-20 -- Represent Collections using JSON Text Sequences (RFC 7464) -- raised
<Loqi> Rhiaro made 2 edits to Socialwg/2015-04-21 https://www.w3.org/wiki/index.php?diff=83609&oldid=0
<wseltzer> trackbot_, bye
tantek: ok. issue 20. I feel like we already talked about this but didn't open it or something. raised by erik
jasnell: this came up in the github tracker as well
tantek: we don't have to resolve it. just whether or not to keep it
jasnell: there are no current implementations that use json text sequences. unless we have actual implementations we shouldn't open it.
tantek: so we should reject this issue because there are no implementations?
jasnell: yes. if an implementation comes along later, we can revisit
<tantek> PROPOSAL: Reject issue due to no actual implementation using JSON text sequences.
tantek: any +1 0 -1?
<AnnB> abstain ... deferring to geeks
<tantek> RESOLVED: Reject issue 20 due to no actual implementation using JSON text sequences.
<trackbot_> issue-25 -- What syntax is (syntaxes are) to be used in the social api (eg microformats vs json-ld; form-encoding vs json-ld) -- raised
<trackbot> issue-25 -- What syntax is (syntaxes are) to be used in the social api (eg microformats vs json-ld; form-encoding vs json-ld) -- raised
<cwebber2> +0, seems fine
tantek: the last raised issue: 25.
... great question. don't really know until we see proposals. seems like something the social api will have to answer.
<jasnell> +1 to open
tantek: though elf isn't here, seems like one that is reasonable to open as it is a question we need to answer for the social api
<tantek> PROPOSAL: Open issue as something that Social API proposals should answer.
<ben_thatmustbeme> +1 to open
<Tsyesika> +1 to open also
<tantek> RESOLVED: Open issue 25 as something that Social API proposals should answer.
<tantek> trackbot, open issue-25
<trackbot_> Sorry, tantek, I don't understand 'trackbot, open issue-25'. Please refer to <http://www.w3.org/2005/06/tracker/irc> for help.
<trackbot> Sorry, tantek, I don't understand 'trackbot, open issue-25'. Please refer to <http://www.w3.org/2005/06/tracker/irc> for help.
jasnell: I have three items that I marked as pending review.
... I assumed they would be covered under tracking actions/issues
<trackbot> issue-29 -- Removing Activity Types not used by User Stories -- pending review
<trackbot_> issue-29 -- Removing Activity Types not used by User Stories -- pending review
<trackbot_> issue-30 -- Deprecate "actor" in favor of "attributedTo" -- pending review
<trackbot> issue-30 -- Deprecate "actor" in favor of "attributedTo" -- pending review
<trackbot> issue-35 -- Simplify the Actor Types -- pending review
<trackbot_> issue-35 -- Simplify the Actor Types -- pending review
<tantek> PROPOSAL: withdraw issue 30 without action.
jasnell: for issue 30 my solution is to simply withdraw it
... this was to deprecate actor for attributeTo and folks didn't seem to like this.
<tantek> any objections?
<tantek> RESOLVED: close issue 30 without action.
tantek: did you want to give the group (wrt 29, 35) some time for review
... is one week enough time?
<Arnaud> trackbot_, bye
<tantek> ACTION ALL, please review issue 29, issue 35, and respective pull requests by 2015-04-21.
<trackbot_> Error finding 'ALL,'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.
<trackbot> Error finding 'ALL,'. You can review and register nicknames at <http://www.w3.org/Social/track/users>.
jasnell: yep. pull requests are on github. the links aren't right.
tantek: you can add the links to the pull requests to the issues
jasnell: a week should be long enough
tantek: Does anybody want to claim success on actions/issues?
tantek: that takes us to activity streams 2.0 in the agenda
Activity Streams 2.0
tantek: first thing is the test harness demonstration
<tantek> Test Harness Demonstration (James Snell and JP Perrault, IBM)
jasnell: jperrault has not participating in the WG before but joined the call today
<Loqi> Alehors made 1 edit to Socialwg https://www.w3.org/wiki/index.php?diff=83611&oldid=83606
<Loqi> Rhiaro made 1 edit to Socialwg/2015-04-14 https://www.w3.org/wiki/index.php?diff=83612&oldid=83598
jasnell: week before last jperrault started working on a basic testing harness that does basic validation of an AS document
... I can hand it over to him
jperrault: if you go to the link, what we have here is a node-based service similar to the w3 validator except it validates against the AS2.0 spec
... intent was 3-fold: to validate the json that comes back from a uri you provide, to validate json format, and validate terms from the spec
... it gives you links that give back valid/invalid json and you can click on them and it will returns errors
<eprodrom> Thanks, sorry about that
... not much to say except the intent would be to modularize the test to test each of the terms and test specific use cases in the spec
tantek: thank you for the demo, can we go to the queue for q/a
<Zakim> aaronpk, you wanted to resolve action-55 thanks to rhiaro
aaronpk: this is unrelated I can wait.
eprodrom: this is interesting. couple questions. one of the things we want to do is test implementations and this seems testing available data, which I guess we could use to test output from libraries or tools we have.
jasnell: this is a start that uses that implementation underneath. the intent is to give us something to start with expecting it would need to evolve with api discussions and use cases
... this at least gets us a path to a way of testing at least a AS document
<cwebber2> eprodrom: I think you might be dropping in and out
eprodrom: one way to see this is if I were to take an implementation and put it so it could be checked with this framework and see the output is valid
jasnell: if you have an endpoint up that publishes AS at a public url, you could give that url and at least see the format is correct
... to check behaviors we would have to change it quite a bit. we could do it. this only checks the format, right now.
... the intent is to throw this up on github and make the tests modular so we can add tests as we move along.
<tantek> like a textarea?
AdamB: what if I don't have a public url to give you. can I run this myself?
<aaronpk> could you paste in a chunk of json in a text area?
jperrault: right now it is a node server. you could pull it down and serve it locally.
<Arnaud> another option would be to allow for a file upload
<aaronpk> example microformats parser that supports both: http://pin13.net/mf2/
<Loqi> Alehors made 1 edit to Socialwg https://www.w3.org/wiki/index.php?diff=83613&oldid=83611
jasnell: the github url doesn't exist yet. we don't have the space under the w3 area. we could throw it on our personal space for now.
<Arnaud> that's what the w3c validator does
tantek: don't wait for process. just ship it.
<aaronpk> is there an issue publishing it under a personal account?
jperrault: we do have an IBM space. we have to go through process.
AnnB: does it have to be harry that does that?
jasnell: not sure. our internal process should be good and we can get it out this week.
... very simple vanilla node app. installing it locally should be straightforward
tantek: comment by aaronpk if you could add a textarea for pasting json
jperrault: of course
<ben_thatmustbeme> jperrault++ thank you
<Loqi> jperrault has 2 karma
tantek: that could help private json validation without a url.
<Zakim> tantek, you wanted to note that if the validator could have direct <a href> links to the test examples, then a consuming application could use those as test cases
<cwebber2> thank you jperrault !
<Loqi> jperrault has 3 karma
<Loqi> jperrault has 4 karma
<eprodrom> Thanks jperrault !
tantek: how do we use this to test consuming applications. if all the links you have could just be a-hrefs you could just use those for consuming applications as well
jperrault: I was having too much fun with jQuery
<ben_thatmustbeme> i think we can save feature requests for the github repo :)
tantek: always a dangerous thing
... any more questions about jperrault's validator?
... great work. looking forward to seeing it on github. let's go to the next item.
<jperrault> diconnecting - thank you all.
tantek: open pull requests. jasnell... all these are from you? these need the group's approval?
tantek: these were in the agenda for a while for people to review?
tantek: I think we talked about this last week. unless objections, we should just say you move forward on these.
<tantek> PROPOSAL accept pull requests 96, 95, 92
<eprodrom> 30 se
<eprodrom> 30 seconds
tantek: eprodrom, you want more time to review?
<cwebber2> oh thre we go
<eprodrom> Sorry, I just wanted to make sure we were doing the ones I though
tantek: alright, we can declare this resolved (laughs)
<eprodrom> +1 for all three proposals
<tantek> RESOLVED: accept pull requests 96, 95, 92
<eprodrom> +1 for all three pulls I mean
tantek: next item: open issues from the editor. that's you again jasnell
<tantek> now Open Issues from the Editor:
<trackbot> issue-26 -- Representing profiles in Activity Streams 2.0 -- open
<trackbot_> issue-26 -- Representing profiles in Activity Streams 2.0 -- open
<trackbot_> issue-27 -- How do we represent changes to a profile in an activity -- open
<trackbot> issue-27 -- How do we represent changes to a profile in an activity -- open
jasnell: we need a resolution. do we want to add a 'profile' object to the core vocabulary or not
... we already have 'person' in the vocab. do we need a 'profile'
eprodrom: I feel like the Person is enough to justify a Profile, or if there are other kinds of profile, we could use other objects. such as Corporation etc having profiles for.
... the biggest issue that a Person having a Profile that people may have other aspects or personalities
<rhiaro> a person can have multiple profiles, but maybe persons and profiles don't need to be modelled separately in the vocab (unsure)
eprodrom: for instance, my interest in baseball or poker may have different personalities facing different social areas
... I feel like it is a complicated and fine-tuned type of concept that can be mapped on multiple accounts
... my feeling is that requiring that all users have multiple profiles is unnecessarily complex for the general case
tantek: do you have a specific proposal to help resolve the issue
<eprodrom> Person class is profile class
eprodrom: my proposal would be to use Person as profile object
<tantek> PROPOSAL: Use Person as profile object to address issues 26 27
<rhiaro> -0 not sure :/
<eprodrom> rhiaro: let's discuss now!
<eprodrom> OK, that's good too
<cwebber2> -0 not really sure either, the thing I'm concerned about
<rhiaro> I'll come up with a counterproposal
<cwebber2> is the "multiple personas" type thing that rhiaro brought up on-list
<rhiaro> cwebber2: maybe you can help :)
<eprodrom> cwebber2: could you bring it up now?
eprodrom: I would encourage discussion
<cwebber2> I'm *not* -1 btw :)
jasnell: the multiple personas thing needs to be considered but gets complex
... lots of different ways to resolve this issue and it is not clear to me that there needs to be one way to resolve this issue
<tantek> Updated PROPOSAL: Use Person as profile object to address issues 26 27. Ok to re-open with new information when available, still open to multiple personas e.g. via extension.
<cwebber2> I don't have a specific proposal right now.
<cwebber2> I would like to talk with rhiaro
<cwebber2> and find out if we could put one together
<cwebber2> yes tantek
<cwebber2> thank you
<tantek> RESOLVED: Use Person as profile object to address issues 26 27. Ok to re-open with new information when available, still open to multiple personas e.g. via extension.
<trackbot> issue-28 -- Remodel "Connect" and "FriendRequest" or Remove them? -- open
<trackbot_> issue-28 -- Remodel "Connect" and "FriendRequest" or Remove them? -- open
tantek: next issue is 28
<cwebber2> Arnaud: :)
<cwebber2> Arnaud: probably true, in the long run
<AnnB> yes, Arnaud
<cwebber2> let's not press tantek into extending this call though :)
<rhiaro> facebook still does that, twitter does with following too
tantek: I have seen connections/friendship/etc as permalinks on things like facebook and seen them on timelines
<cwebber2> AnnB: ;)
jasnell: the way it is currently modeled is that friend-request is the verb, which doesn't make much sense. as opposed to creating a connection between me and somebody else.
tantek: there is 'followed'
jasnell: there is a subtle difference and worth exploring. my recommendation is to change them from activity types to object types. instead of friend-request, they'd be a friendship object.
<tantek> PROPOSAL: Remodel "Connect" and "FriendRequest" as object types
tantek: seems like a reasonable improvement
<cwebber2> hm interesting
<rhiaro> +1 to connect object I think
eprodrom: the main thing I like about this proposal is that that relationship object can have a whole CRUD lifecycle. it can be created/updated/deleted and that's a nice part of that connection.
<cwebber2> having a CRUD lifecycle to an object would be nice
<cwebber2> to a connection
<cwebber2> AnnB: I agree
eprodrom: it can even have properties that are represented. work-relationship, wife, father, etc.
<cwebber2> fwiw I think this meeting is going pretty well
<tantek> eprodrom +1 agreeing with CRUD aspects of relationships as object types
<cwebber2> and I know I've been complain'y about us dragging on communicating about communicating in meetings, I think this one has been hitting important points :)
eprodrom: the only problem I would have is that in general a lot of the semantics of AS is in the activity types / verbs, and would we be losing some part of the social software by doing this.
... I think the up-sides outweigh the down-sides.
tantek: I'll take that as a +1. the proposal still stands. no objections?
<tantek> RESOLVED: Remodel "Connect" and "FriendRequest" as object types per issue 28
<cwebber2> I'll like the reshare of the connect object :)
<trackbot> issue-32 -- Deprecate "generator" and "provider" -- open
<trackbot_> issue-32 -- Deprecate "generator" and "provider" -- open
<tantek> PROPOSAL Deprecate "generator" and "provider" in AS2
jasnell: I've only seen [generator/provider] used for metadata. I want to simplify and move these down. deprecate them. want to see what people think.
eprodrom: I think the big thing about generator/provider.. I'm less concerned about provider.. the generator one plays the same role as an application generator in a facebook stream does. helps identify the application that generated the activity.
<jasnell> so maybe just deprecate provider?
eprodrom: super helpful for the app. if you are farmville, then this is important to you.
... would be ok with dropping provider, not generator. wonder why jasnell think's it is not important.
<cwebber2> could se way this is something that could be an extension?
<cwebber2> we say
<cwebber2> not se way
jasnell: I just haven't seen it used in implementations. I've fine if your experience is different. provider is largely redundant.
<tantek> updated PROPOSAL keep Generator, deprecate Provider from AS2
<ben_thatmustbeme> i think its a pretty useful item, certainly something we see examples of (facebook does with "via" links)
<rhiaro> yeah, I'm going to add crediting applications to my posts, too
aaronpk: I agree with eprodrom on this. the generator is useful to credit the application that was used. I use this to credit applications on my own streams. think it is useful to keep it.
<AnnB> cool, aaronpk
<cwebber2> +1 sure why not :)
<eprodrom> Finger slippage
<tantek> RESOLVED: keep Generator, deprecate Provider from AS2 per issue 32
<Tsyesika> i agree
tantek: we have one remaining issue we will move to next week. we have an open issue from elf, but he isn't here. so I say this is a reasonable progress for the hour. we should go for an hour next week as well
tantek: if no objections
<cwebber2> +0 :)
<eprodrom> Great! Thanks wilkie and tantek
tantek: great. with that, the call will be closed.
... (some vague praise for wilkie)
<cwebber2> great meeting everyone, thanks for chairing tantek
<cwebber2> and thanks for scribing wilkie
<eprodrom> Thanks all
<AnnB> could you guys look at https://github.com/w3c-social/social-ucr/issues ... make your comments to help resolve objections?
<Loqi> tantek has 180 karma
<Loqi> wilkie has 8 karma
<AnnB> thanks a lot tantek .. good chairing
<Loqi> wilkie has 9 karma
<Arnaud> trackbot, end meeting
<Arnaud> two bots and that didn't even work
<tantek> Arnaud, that URL looks reasonable