Social Web Incubator Group Teleconference -- 03 Nov 2009

<trackbot> Date: 03 November 2009

<rigo> use cases for social web from primelife: http://www.primelife.eu/images/stories/deliverables/h1.2.5-requirements_selective_access_control-public.pdf

<rigo> see http://www.primelife.eu/

<FabGandon> Claudio Venezia introducing himself

<FabGandon> Tom (?) from NTT communications

<DKA> SWXG Wiki: http://www.w3.org/2005/Incubator/socialweb/wiki/Main_Page

<FabGandon> Adam Boyet from Boeing

<FabGandon> Rigo from W3C dark side of the law

<FabGandon> Rigo presents Prime Life and mentions its use cases in particular the problem of taking down the walls of the walled gardens

<FabGandon> "PrimeLife - Bringing sustainable privacy and identity management to future networks and services"

<FabGandon> Adam presenting the inSite plaform internal tool of Boeing

<FabGandon> Vagner talking head of W3C Brazil office

<hajons> Håkan Jonsson (hajons), Sony Ericsson

<VagnerW3CBrasil> Vagner Diniz is the head of W3C Brazil Office, joinning thhis session to know what is going on, particular interest in social web in mobile web

<VagnerW3CBrasil> ... and deployment of social web in W3C Brazil Office

<FabGandon> Yoshiaki internal student of W3C/Keio Univ

<martin> Martin Higham - Ocasta Labs

<rigo> I presented Primelife use cases that _exclude_ tearing down the walls

<FabGandon> Christian de Sainte Marie (RIF WG) just joined as an observer.

<FabGandon> Scribe: FabGandon

<rigo> scribenick:FabGandon

DKA flipping the agenda ; want to start with the use case documents.

User stories : http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories

<DKA> http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories

<DKA> DanBri let me know if you want to skype in to our lovely session.

user stories are split e.g. http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories/BusinessIntelligence

<danbri> it's dinner time here soon unfortunately, perhaps I watch by IRC and jump in if something specific crops up?

http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories/DragAndDrop

http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories/AcceptingAFriendshipRequest

latest template http://www.w3.org/2005/Incubator/socialweb/wiki/Template:UserStory

DKA: users Users stories are to create a picture of the social web for the users ; looking at it from a user experience

Rigo: to get momentum we need to also include the social web makers in the picture

Fab: +1 for scenarios including all the people impacted by the stories

DKA: yep, service providers must be included too in the scenarios

Hakan: should we include third-party tools developers?

Rigo: PrimeLife developed Persona we can reuse.

DKA: Actors: people (end-users), service providers, developers (3rd party)
... looking at the use cases one by one.

<DKA> http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Download_your_data

DKA reading use case "Download your data"

DKA: we didn’t mention taking the data and uploading them in to another service ; should it be part of it?

FabGandon: is this part of another Use case ? if so is there an opportunity to merge them?

DKA: is this about having a unified API to the social web ?

FabGandon: API (such as in the CRUD scenario) + Format and model (as in FOAF + Relationship)

DKA: Difference between having interest in data and having the ownership of data.

<rigo> the PrimeLife personae are published at http://www.primelife.eu/images/stories/deliverables/personas_primelife.pdf

Rigo: data self determination says person data is mine ; every where except in the US

Information self determination : http://en.wikipedia.org/wiki/Informational_self-determination

DKA: Alice becomes a member of a social network and populate here profile ; Bob connects and see here phone number ; is the phone number Alice’s data or also Bob's ?

Rigo: Alice's
... if you have spread data you have a right to get it back, modify, etc.
... data is mine and have a say on this data

FabGandon: application of the CRUD scenario http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#CRUD_Operations_on_Social_Data

DKA: Data copied from Alice’s by Bob are still Alice’s data?

Rigo: if its only Bob, this is private data of Bob so there is nothing you canb do
... if the data are captured and kept by a Telco company then there is a problem ; no longer private copy of someone.
... in the case of Bob's copy there is no legal aspect to it.
... in the case of a Telco then there is a legal aspect to it.

DKA: DRM ?

Rigo: no privacy issue about this particular use case.

<rigo> I need to go to a telco

DKA: data + policy

FabGandon: see also Intransitivity of Policies Applied to Social Network Data http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Intransitivity_of_Policies_Applied_to_Social_Network_Data

Hakan: “downloading” does not describe the intent of the user.

Adam: the use case even talks about aggregating

Rigo: opportunistic use of data ?

DKA: Changing the title to "Reuse your data" quote "I have made the change because I am God"

Now talking about Drag and Drop http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Drag_and_Drop

Ann: not really different from previous use case

Adam: it is about adding information.

Ann: another use case would be filling the blanks of some data I have.
... merging with data I have

Hakan: isn't it ease of use use case?

DKA: ok we should merge both use cases.

Liking to a remote Friend http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Linking_to_a_remote_friend

Adam: connection request across networks.

Ann: “establishing a connection across networks” since the notion of request is dependent on some application.

Martin: follow and friendship are very different.

FabGandon: slash dot has "foes" that you won't find in Facebook.

Martin: the application you use to set the link will impose the type of link you set.

DKA: different kinds of links.

Martin: social network B will impose its rules.

DKA: Social network B needs to know the request and decide to grant it or not.

Adam: Joe will thus appear as a User of social Network A although he does not use it?
... the social networks are actors.

DKA editing the page as we speak.

<DKA> http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Linking_to_a_remote_friend

DKA: missing use cases : foe, blocking, ...

Adam: current text is in conflict with post-condition

Hakan: post condition should state what the social networks A and B claim about Alice and Joe

DKA writing alternative paths.

Adam: if I am on several network which one should deal with my request?

BREAK until 11AM

<mischat> right i am to treck to wimbledom to lug boxes half away around london

<mischat> hehe

<mischat> wrong window:)

<mischat> sorry

<danbri> btw if anyone is in Montreal, identi.ca / statusnet are hiring --- http://jobs.status.net/

<scribe> scribenick:Adam

<FabGandon> Scribe:Adam

dka: let me tell you where my cursor is ... alternative path 2
... alternate path 2 needs to be filled out a little bit more so that SN B needs to find out more about SN A

martin: Alice request on SN B to connect to Joe (who is on SN B), she provides her SN A identity when doing this
... SN A asks/confirms that she did make the connection request on SN B. Alice confirms with SN A. Joe appears on Alices connections in SN A
... thats the basic course of events .... then there the case where SN B asks Joe if he wold like to connect to Alice as well

rigo: the desire of the business model is that SN A typically wants to rope in the user from SN B

martin: thats not the desire from the user

dka: captured that as an alternative path #2

http://www.primelife.eu/images/stories/deliverables/h1.2.5-requirements_selective_access_control-public.pdf

oops, wrong URL

http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Linking_to_a_remote_friend

<DKA> http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Linking_to_a_remote_friend

dka: the thing that makes connecting connections across social networks (SN).. both SN providers need to be aware
... even in an asymetric relationship, there is some symetry that SN B needs to know about that request, even if it doesn't need to approve
... thats the kind of leap we've been making

timbl: one way alice can tell sn b, or a third party could crawl the web
... you could ask who's friending me, you could do that without coperation of the sites

dka: so you're saying another alternative path that SN B could only find out after the fact

timbl: another thing, you could send the requst by public email, alice could send it to bob's email

<timbl> Currently for example qdos.com has a reverse friend lookup serviuce over the FOAF graph

rigo: the notion of friend is a central one ..... in that the thing you attch to when you arrange a whole set of policies
... the act of making friends means that Alice can now set an access policy, that means that it would have to go in to the access control system of SN A .. in this case SN A would have to allow Joe to see it

dka: SN B is able to request from SN A are there any pictures that i can display from alice

rigo: if you tear down the silo of the walls, then you have to make sure you see the notion of friends ... then SN A has to be able to limit

dka: if joe goes to SN A to see Alice profile he says let me see the pictures i can see, then SN A needs to be able to identify joe

<danbri> (aside, when you say SN, ... increasingly it might just be 'site'?)

rigo: there are some providers that federate like plaxo like an rss feed ... but then how do you do access control

dka: i think this means we need to add a use case taht i don't think is accounted for

<danbri> re lookup services, also consider http://socialgraph.apis.google.com/

dka: joe visits alices profile page on SN A

<danbri> it is used by http://googleblog.blogspot.com/2009/10/introducing-google-social-search-i.html (and based on xfn and foaf)

timbl: there is some initial interaction that seems to be missing

rigo: the reported in the social identity ???? they discussed unique identifiers and email address as the unique identifier
... they discussed web finger

Haken: its a discover step that we're missing

dka: this use could be me following steven fry, he doesn't know who i am .... it's symetric in the sense that he knows that i'm following him
... in this case, joe (SN B) knows about the relationship

Ann: i wonder if the title is misleading
... maybe connection is misleading

timbl: there could be lots of ways that alice knows about joe
... could be email, or could have found each other ... eventually they establsh this connection

dka: in the course of events its implicit that if joe doesn't have alices identifier, by the time this is done joe definately does
... if alice makes a request, then joe knows she made the request

martin: there is a post condition that joe knows about alice

<timbl> Missing step 2: Alice sees joe as an unconfirmed friend, Joe get extra rights to see Aice's stuff.

dka: right but is that true that we always want to have that, is it meaningful that alices identifier is not known by joe

rigo: typcially you have a public profile where it gives you a teaser to get you to join the SN

Ann: the profile could be obscure too

rigo: the web has unlimitted number of possibilities to discover people
... in this case there is no back identifier

timbl: step that is missing, after alice requests SN B (joe) .... when she stated joe is her friend, she see joe is listed as her friend. second part is if he ever comes in to facebook now, he gets rights to see more about alice

<timbl> Missing step 2: Alice sees joe as an unconfirmed friend, Joe get extra rights to see Aice's stuff.

dka: adding that notion to the user scenario

refresh user scenario page

yes, http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Linking_to_a_remote_friend

<oshani> thanks

dka: does that mean we should delete the Accepting a Friend use case

<danbri> (another aside from me - I hope you can talk a bit about lists/groups, eg. the recent twitter 'list' feature is making quite a stir... and identi.ca/statusnet have groups)

martin: its seems to be talking about http ... low level stop

Ann: seems that there are components of accepting and rejecting that might be good to keep separate

dka: seems that rejecting is a different alternative path and that blocking may be a separate use case

FabGandon: maybe the blocking could be part of the CRUD one

Ann: one aspect of block that you just block people you don't know but then the case about a stalker, you are being stalked and want to block that person

dka: editing wiki, strying to add blocking
... and then lets talk about that ... i think the relationship is binary but the meta-information is on top of that

timbl: thats how it is on facebook, but they've had to start adding classification to it
... but whether i classify you as a friend or a foe, is not visible to you

dka: i still like thats classifications of your social graph
... i have linkage but i hang metadata off of that linkage

Hajons: but do you have to accept to be a foe

<danbri> all the services will add private and public categories, i'm sure of it...

dka: is there a case where anything more than the binary connection is a personal matter

<hajons> Håkan = hajons

thanks!

dka: if listing that we, individually say we are w3c colleagues that doesn't mean as much as if we both agree to say that. there additional value in that

<danbri> here's an example i started btw, putting people into groups --- http://wiki.foaf-project.org/w/FOAFLists

Ann: what if you want to remove / revoke the relationship

martin: establishing the connection and then another could be managing the connection

dka: now we have establishing, managing, and blocking

Håkan: I'm thinking about the metadata. we're talking about establishing the relationship ... what about the rules of the relationship

dka: if SN A really didn't know about SN B, then somehow SN A would have to allow Alice to agree to SN B term and conditions

martin: with status.net all the content is pushed out to the network ... so alice wouldn't have to agree to SN B terms and conditions cause she's not creating any content on SN B

Håkan: providing some type of informationa bout the type of relation since there are no standards for this

scribe: but there is no way for SN B to consider different types of relationships
... it can't apply policies

claudio: later on in the privacy section of the use cases, there are some things about different types of identities
... right now we are decoupling the identify from the SN account. what happens if we manage the different relationship profiles with these different identities ... are we over complicating the situation

dka: i think one of the preconditions is that SN A has already verified the identity of Alice

rigo: don't even get in to the game of determining what an identity is
... we hit that problem in many areas .... now more than 6 years of research in this areas is a fan of multiple identities
... do not try to identify a physical natural person
... here we shouldn't think about identifying a natural person but should take some kind of virtual identity and person can have more than one

<AnnB> s/sitation/situation/

dka: agree we don't want to get in to solving the identity problem

<rigo> scribe:Rigo

<scribe> scribenick:rigo

dka: adding to preconditions that SN A has added Alice as an identiy

tbl: lots of personae around, some people can correlate others don't

dka: one use case missing, relationship between different profiles to say "this is me"
... is a kind of managing personae idea
... actually there is no way to say that "this is me"

<danbri> re this is me, XFN handles that very concisely: rel=me

<danbri> i've been looking at foaf/xfn integration - via groups:

<danbri> here's an example i started btw, putting people into groups --- http://wiki.foaf-project.org/w/FOAFLists ... also as a plan for bridging foaf with xfn

dka: use case is e.g. to say to my phone that all of those different entries is me

CSM: is that related to the use case this morning?

dka: yes

??: can do the 'me' thing

<hajons> ?? = Håkan = hajons

<AnnB> also, TBL said FOAF can identify "me"

<danbri> FOAF has a few tricks for figuring out identity from descriptions (plus simple use of URIs and owl:sameAs)

<hajons> on the me rel tag: http://www.rexblog.com/2009/04/21/19358

<Adam> lots of discussion on how linking or saying this is me can cause concerns

rw: don't absolute IDs, just use relative identies wherever you find them

<AnnB> tlr: the "me" relationship might help one SN discover info from "me" in another SN

tlr: ...dka
... said

dka: how to see information flow between networks
... how this is better compared to the current situation
... bad scenario: how to manage all fragmentation

HJ: want to link my professional profiles together and my private, but independent from each other

vocab.org/relationship

<DKA> http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Managing_an_existing_connection_across_social_networks

dka: managing connections, we haven't anything yet
... if Alice and Joe have connected, SN A and SN B already know about each other

tbl: reciprocal?

dka: even in asymetric relations, a and b have to know each other

http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Managing_an_existing_connection_across_social_networks

dka: basic connection unless Bob has to agree

hj: connection only established after permission

MH: we are "managing"

HJ: we should include the possibility to reject a relation

dka: http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Establishing_a_connection_across_social_networks

CV: friend only discussion on business thinks
... normally you don't talk about nature of relations

HJ: not saying that we need always type of relation, but we should allow for it

dka: this can be serialized..
... if connection is established, the connection is categorized..

HJ: this is only about current things, not how this should be in the future

AB: friends that you don't care about

MH: there may be a level of control

AB: profile information is something that you totally control
... not true in a corporation environment, no choice, company policy. can't force people to put picture

rw: best practices document for corporate environment may be interesting

MH: establish connection with your business network, it is not only you that determines

AB: they struggle already with it
... already big warnings about not giving names
... most worried about phishing attacks

DKA: not talking about your secret company projects ...

<hhalpin> hey sorry I'm late

dka: capture some stuff on enterprise

<hhalpin> how is the meeting going?

FG: it would be good to have use cases from inside big corp viewpoint
... it would be 5.14

<hhalpin> Just ping me if you need any help, otherwise I'll monitor in and out via IRC.

FG: another one on the mailing list

AB: different in business intelligence, here rather somebody from inside the corp communicating outside corp

FG: we do not capture use of social networks inside corp
... on the intranet

<DKA> ACTION: Adam to write a use case about business social networks - on an intranet - e.g. when ownership or editorship of certain parts of your profile are not under the user's control. [recorded in http://www.w3.org/2009/11/03-swxg-minutes.html#action01]

<trackbot> Created ACTION-104 - Write a use case about business social networks - on an intranet - e.g. when ownership or editorship of certain parts of your profile are not under the user's control. [on Adam Boyet - due 2009-11-10].

RW: in corp environment, linking several instances of SN needs super social network with ID

=====================================

http://www.w3.org/2005/Incubator/socialweb/wiki/UserStories#Managing_an_existing_onnection_across_social_networks

dka: do we need to have specific relationship vocab

FG: recategorizing existing relation

Adam: categorization, should Bob be aware.

dka: different from requesting friendship to share information
... or some additional functionality

tbl: categorization of relation and categorization of persons
... facebook makes classes, related to me, "my friends". Box of students, want to be able to treat them the same as the list of students that MIT has published

<harryhalpin> to ask a quick off topic question in irc, any idea what happened to henry story? does he need any help, and is there anything I can do?

tbl: or class of people that have attended a certain conference
... some will be public, some private. Notion of Group is very general

FG: case: party with all colleagues except one... intersection, challenge to do that ergonomically

tbl: all my friends except this persons is difficult

AB: you have the list of friends, then you can unclick the preselected,

tbl: so procedural?

AB: yes

Adam: Tbl wants to use the groups consistently across different groups

tbl: to every relation, there is a class

MH: does this have an impact on later interconnect? We should focus on those

<danbri> are you talking about what i think you're talking about?

<danbri> groups/lists vs relations?

<harryhalpin> sounds like rdf domain and range...but not *required* class? Open world or closed world?

<Adam> yeah, kind of :)

<danbri> my current example is

<danbri> #danbri :homepage <http://danbri.org/>; :openid <http://danbri.org/>; :made #danbri-wouldliketoknowbetter .

<danbri> #danbri-wouldliketoknowbetter a :Group;

<danbri> :member [ a :Person;

<danbri> :homepage <http://tantek.com/>;

<danbri> :account <http://twitter.com/t> ] .

<danbri> ... but not clear how much logic to put into Group versus simply using OWL and subclasses of Person

lunch

<Adam> out to lunch all

- DRAFT -

Social Web Incubator Group Teleconference

03 Nov 2009

Attendees

Contents

Summary of Action Items

Scribe.perl diagnostic output