W3C

- DRAFT -

Workshop on Social Standards: The Future of Business

07 Aug 2013

See also: IRC log

Attendees

Present
Regrets
Chair
SV_MEETING_CHAIR
Scribe
gkellogg

Contents


<aaronpk> RRSAgent: "Sorry, Insufficient Access Privileges"

<inserted> scribenick: wseltzer

<aaronpk> it says "Sorry, Insufficient Access Privileges"

<aaronpk> thanks!

<danbri> Hi folks. Sorry I couldn't make the trip, will lurk here...

Welcome

hhalpin: Welcome. Thanks to AppFusions for hosting; IBM and Open Mobile Alliance for sponsoring

<bret> Is anyone recording video or audio?

hhalpin: Harry Halpin, W3C, co-chairing with Mark
... W3C is standards org for the Web. We're excited to make Social a first-class tech on the Web.
... W3C is a consensus-driven, open, voluntary standard-setting process
... If you have questions about W3C, see any of the W3C staff here.

Mark_Weitzel: President of OpenSocial Foundation
... Excited to think about how Social is solving problems for enterprise
... Started from consumer side, now looking at bringing social models to the workplace
... Find any of our Board members for more about OSF

hhalpin: Agenda: http://www.w3.org/2013/socialweb/agenda.html
... First day, high-level overviews for level-setting; lots of Q&A, open discussion
... hope to understand visions, gaps, agreements and disagreements
... Second day, focus on next steps, standardization strategy.
... Breakout groups and discussion
... After the workshop, W3C and OpenSocial will prepare a workshop report
... Develop concrete outcomes from the workshop and beyond.
... calls for scribes

AnnBassetti: volunteers to show people around IRC

hhalpin: Thank you all, especially Ellen Feaney for setup, all for coming.

Keynote

Dion_Hinchcliffe: Addressing disconnects between users, standards, and busioess
... Dachis Group, frame a discussion of benefits, issues practitioners struggle with
... How many from industry: a few
... How many from technical background: many
... Getting the right problem definition and requirements: What problem(s) are we trying to solve
... RSS was one of the key early social standards; people were asking for it
... created the early social web, pull from end-users
... Where is the demand going to come from for social standards?
... How can we ensure solutions are a good fit for the problem?
... Successful standards are hard to create.

<bret> Anyone object to closing the doors?

<AnnBassetti> might get warm ... still probs hearing?

Dion_Hinchcliffe: W3C has been fortunate to get high-level adoption of the core HTML
... Key to standards adoption: Simplicity (RSS fits on 2 pages)

<bret> No, I can hear fine, its cold :p

<AnnBassetti> move forward?

Dion_Hinchcliffe: Universal: you can’t propose something be universal and keep control of it.
... Numerous standards have come but mostly gone.

<James_M_Snell> Simplicity is key to success... Imperfections in the specs can be easily tolerated so long as the spec is accessible by everyone

<Kelvin_Lawrence> Will we be able to get a copy of these slides?

<Kelvin_Lawrence> Thanks

Dion_Hinchcliffe: We expect results while deploying social mechanisms "off to the side"
... instead of isolating it, integrate it.
... Bring social together with your systems of record.
... Connect social media with the flow of work.
... How could we create a crystal-clear conception of what we want from social business architecture
... Make it as easy to use as the USB2 port on the side of your macbooks.
... Make it modular, able to talk with all the other apps and data we have.
... 1) federated and interoperable. Easy for anyone to plug in and extend
... We're surrounded by social but none of it works together
... That's our challenge
... What's missing in social standards?

<James_M_Snell> Idea to consider: a "standard" social web app API.. window.social.whoAmI(callback)

[slide 8]

<ryanjbaxter> @James but i have multiple identities, would it be different across the app?

<James_M_Snell> allow apps to fill in the details behind the scenes

<ryanjbaxter> what if I don't want the app to know who I am, there needs to be ways of granting access to that information

Dion_Hinchcliffe: Social Web doesn't look much like WWW; it's not universal

<ryanjbaxter> I like the concept of the APIs being "native" though

Dion_Hinchcliffe: Silos of conversation impede progress

<James_M_Snell> window.social.whoAmI returns the apps view of who I am... so it's limited to what data that app has access to

Dion_Hinchcliffe: Early social media: blogs, RSS, highly decentralized and interoperable
... Next era, large commercial socnets, with their own tech stacks
... if you're on socnet A, you may not be able to talk to people on socnet ZB
... network effects.
... Who's our target audience for socbiz standards?

[slide 10]

<AnnBassetti> what was that stat? 'over 100 social networks, that can't talk to each other'? many with 100Ks of users....??

Dion_Hinchcliffe: consumer socnets, enterprise soc media vendors, developers, IT managers & decision makers, end-users

<hhalpin> we'll post slides on website ASAP

<hhalpin> but feel free to track Dion down for that quote :)

<AnnBassetti> I will ... wasn't on his slide I don't think

Dion_Hinchcliffe: Big issues: fragmentation.
... "Hundreds of social networks with over a million users"
... walled gardens
... business models, often in direct opposition to opnness
... Lack of compelling critical mass of socbiz standars
... No "killer app" standard. How do we get there?

<James_M_Snell> Ann: keyboard jumping, eh? Telling it to sit, stay and lay down isn't helping? You try giving it a treat? Playful keyboards are always a problem

Dion_Hinchcliffe: It would be great if we had a fully federated social Web.
... need to get all networks, consumer and enterprise talking together
... data of record and conversations in the same place
... easy access to conversations, wherever you and they are
... portable information
... be able to take your social capital where you like. you created it, it's your info.
... Pathways: think big or think incremental
... Incremental: build on existing standards
... Revolutionary: We understand use cases, build to fit
... Integrative: synthesize, extend, embrace
... Social Media has been the biggest communications revolution ever, since the WWW

<James_M_Snell> The most significant hurdles to a "fully federated social Web" are not technical, they are political and commercial... it's not apparent that the most significant socnet providers *want* to federate

Ashok_Malhotra: If I'm a huge social network, why would I be interested in opening up?

Dion_Hinchcliff: We need a motivation for them too.

<hhalpin> agreed James. There's not much interest in the consumer-facing space, but we're seeing huge interest in the business space.

<AnnBassetti> Theo Havinis, SymptoLogic

<hhalpin> So first large ecosystem that takes advantage of standardization could get a first-mover effect in the emerging market of integrating social with business processes.

Theo_Havinis: How can we control spam? Facebook is a closed group, that helps keep the spam out.
... There's value to closed networks, e.g. in health care.

<James_M_Snell> then we're not talking about a "fully federated social web"... we're talking about building trusted connections between small private socnets

Dion_Hinchcliff: A key component of social has been that you don't need to pre-define users
... in some business processes, you do need to define the users.

Rajav: What activities are people doing? Communicaitons, product management? What do we want them to do?

Eric_Bates: UCSF, health care. We want to get health care researchers collaborating online.

<James_M_Snell> Social in business == my identity at work extends out to everyone else I interact with in that capacity, including my connections, my activities, my contact details

Eric_Bates: we need open standards because academics from different disciplines won't all join the same networks.

<James_M_Snell> or, my identity in [whatever community] extends out to everyone else I interact with as a member of that community

Ed_Krebs: Ford, we've got several different problems to solve
... we have three sets of scenarios. How do we manage reputatkion on the outside; inside collaboration; "inside and outside"
... How do we work with outside collaborators?

AnnBassetti: Boeing, We have our own internally built social tool
... Adam will speak about it. 85k people on it, inside the company. None of the commercial products worked for us, so we built our own.
... We need open connections in all directions
... how do we build those connections while looking at security, privacy, IP, etc.

Ellen_Feaney: Appfusions, we're a systems integrator. Atlassian, departmentally distributed
... Some large organizations are top-down

Mark_Weitzel: OpenSocial. How do we bring social context to where we're working.

<James_M_Snell> We need a clear definition of what Social *means* in business

<Li> another IRC channel logging tweets, http://indiewebcamp.com/irc/2013-08-07

Mark_Weitzel: contextualized interaction.

<James_M_Snell> contextualized identity

Howard_Winegrad: Tibco. Simplicity is key. there's a mess of different standards now, it's horribly complicated.

<bret> http://xkcd.com/927/

Howard_Winegrad: we need simple standards. JSON is simple, email is simple.
... need to talk specific use cases

Social Business Architectures

Mark_Crawford: SAP, how we approach Social Business from solution provider

<scribe> scribenick: hhalpin

markcrawford: SAP
... Rethinking the Enterprise Social Business
... SAP Jam
... the SAP offering in the social biz space, from SuccessFactors
... built on enterprise social network
... social colloboration tools
... the social interations are necessary with customers/partners, not just sharing pictures
... transforming the use of social from a personal relationship to a *business* relationship
... its important as we're good with transactional business processes

<James_M_Snell> Use case: I am an employee of Foo. In that capacity, I need to access an external community of partners and customers. When I do so, I need my identity as an employee of Foo to extend out to that community, including my reputation

markcrawford: this only a small portion of busines
... most of the day-to-day interactions are non business transactions
... ordering printer, airplane ticket, etc.
... how can we capture that except in email?
... can we do that throughout our entire supply chain
... purpuse built social media analytics
... we need to tie social interactions
... into business records
... thus Jam ties into analytics, cloud apps, business platform
... what are SAP customers doing?
... but we need to support hetereogenous environment
... with 3rd party apps
... lots of use-cases

<James_M_Snell> Another use case: I am an employee of Foo, my company has subscribed to a hosted service that provides various functions... my organization connections (management chain, peers, partners, etc) need to be accessible via that third party service

markcrawford: informal learning, internal communities,HR management, supply chain, social onboarding new employees
... we need to integrate
... Facebook may not always be dominant player
... my kids say its passe
... kids don't email, they tweet or text

<Li> nobody can be the Dominant player, neither facebook, nor google. consumers avoid put eggs in one basket

markcrawford: we need to bring business products into the way a new generation of employees operate

<Kelvin_Lawrence> Mark Crawford quote "80% of employees don't use their company's social networks"

<AnnBassetti> Mark Crawford: need web-based semantic standards for social content

<wseltzer> Ed_Krebs: Join us in developing scenarios in the W3C Social Business CG

<benwerd> (Microformats!)

<Bill_Christian> is there a link to the W3C group Ed spoke of on scenarios?

<wseltzer> [The community group: http://www.w3.org/community/socbizcg/ ]

<wseltzer> scribenick: wseltzer

Ed_Krebs: How do we bundle the standards in the social space?
... architecture: http://www.w3.org/community/socbizcg/wiki/File:Social_ppt-JeffJ_intro_for_AC_v3-1.pptx
... give integration points for our architects, standards discussion, modules.
... At the end of the next two days, I hope you'll know whom else to talk to.
... I insist that my next gen platforms are based on standards.
... Reference architecture I'm presenting inside of Ford.

[slide: Reference Architecture]

<bret> Sorry its already been linked, but are the slides online?

scribe: highlights in papers on CG site: http://www.w3.org/community/socbizcg/

<AnnBassetti> background slides of the previous 'block architecture': Social_ppt-JeffJ_intro_for_AC_v3.pdf‎

<AnnBassetti> pages 9-19

scribe: diagram, with social network platform in the center. that might be build or buy, start with standards around the edges
... We collaborate in a lot of different tools. We don't design cars in Sharepoint.

<AnnBassetti> whoops: http://Social_ppt-JeffJ_intro_for_AC_v3.pdf

<harry> scribenick: harry

EdKrebs: Ed Krebs Ford
... we want to purchase a enterprise social networking platform

<hhalpin_> scribenick: hhalpin_

edkrebs: how can we get proprietary work with other people's work.
... "things on the side", I store files in 3 places
... I don't want to store files in another place

<AnnBassetti> Ed Krebs: need to get at the 'wisdom nuggets'

edkrebs: want wisdom nuggets
... pulled from the network

<wseltzer> scribenick: wseltzer

<AnnBassetti> scribenick: AnnBassetti

Jeff Calusinski, Don Buddenbaum ... IBM

people don't 'get' how to hook systems together

Community Grp identifying use cases, ...

business people want to solve problems

JeffC: ++ people use Facebook, but it's an island itself

in enterprise, 'social' has to be embedded with where people do their work

tech standards will help with 'plumbing', but what is business context

moving to service-based model (esp with 'cloud') ...

brings new challenges

need metrics for business scenarios, help enterprises understand how 'social' is helping

better alignment of business architecture and technical architecture

need both loose coupling and tight integration

can we enrich interactions with social feedback?

DonB: assertions... <see slides>

<btw, these minutes should also reference the Twitter stream on #osfw3c https://twitter.com/search?q=%23osfw3c&src=hash >

<wseltzer> next: Lloyd Fassett (Azteria)

sure

<scribe> scribenick: AnnBassetti

Lloyd Fasset, Azteria ... from staffing realm

esp. healthcare staffing

data and conversation aren't adequately linked yet

I'm representing a different 'process' , rather than a company

quote from Sangeet Paul Choudary, "we have 300 years of how to run businesses to create value"

scribe: <something about new mode being between individuals, and aggregating external data>

summarized as "pipes to platforms"

when we can crowdsource from ++ diff sources, is a whole new opportunity

scribe: but planes and cars won't be built on that info

<I add, though, that Boeing and Ford can glean ideas and feedback from the crowd>

----- ,end of panel ---

<wseltzer> [Q&A, then coffee break]

<wseltzer> Beth Lavender, MITRE

Beth , Mitre

Ed Krebs, Ford -- business has to figure out what we can share (due to intellectual prop, privacy, engineering etc)

most important is if we can connect people with new people, that they need to know

Jeff C: join formal business processes with ad hoc collaboration

metrics really help

Vassil Mladjov, Gartner: why do you call this 'social'?

big diff between collaboration and social

Ed Krebs: we define collaboration as teams working together

<wseltzer> EdKrebs: Collaboration is teams; social is working with connections outside, that you might not even have known existed before

scribe: and 'social' is making connections you didn't have before

Mark Crawford: and, you want your systems to make those connections for you

scribe: to create the interactions automatically

Lloyd Fassett: asks MarkC and EdK: how to you track processes, to minimize repetition?

EdK: bigger challenge is 'how to keep collection of comments about a paper (e.g.,) together?"

<Kelvin_Lawrence> This is a good discussion of the difference between "collaboration" and "social". I have seen this question asked before and people do seem genuinely unsure of where the line is.

EdK: notes that Twitter implemented threaded view; was too hard to track

<wseltzer> Kelvin_Lawrence, agreed. And remember, we can't force those outside the company into the company's preferred platform, so we need standards to meet them where they are.

Jive has concept of 'talk' to 'action'

LloydF: how to 'bake' that into a standard?

Weitzel: start with a limited set of nouns or verbs

Theo Havinis: diff layers (e.g., telephones, email, Facebook) .. each has it's own mode of communication and social

Jason R___ , OpenSocial: who cares if this is "social" or not?

<wseltzer> Jason: "Is this house haunted?" "Is this social?" I don't care!

what's cool, is that Ford lets owners name their car

we call our car "Flexi" ....

but Flexi can't interact with the dealer, or the garage..

EdKrebs: we've have ideas about where to take that

<wseltzer> The reason to care about naming, I'd suggest, is to get the right people into the room for the right set of standards coordination.

<wseltzer> The name is just a pointer to help get them involved.

Rawn Shah: actually, we cannot plan what happens with 'social'

not necessarily a standards problem, but rather a business problem that will affect standards

<AnnBassetti_> Rahn Shaw: businesses trust their old processes ... need to trust 'social' more

<James_M_Snell> "Social" is fairly close to being meaningless actually. We ought to be much more precise in our language. "Social" encompasses "Context" (Why am I here, What is my purpose), "Identity" (Who am I right now), "Network" (Who do I collaborate with), "Data" (What information is most relevant right now) and "Activity" (What have I done, What do I need to do)

<AnnBassetti_> EdKrebs: to transform your company who better than your employees, partners, etc?

<AnnBassetti_> example of Rahn's statement: Wikipedia that replaced an earlier way of capturing encyclopedia

<AnnBassetti_> Matt Franklin, Open Social: does not like social vs collaboration distinction

<AnnBassetti_> Larry H .. : are we missing opportunity? users just work with one system..

<AnnBassetti_> Howard W, Tibco: don't need to limit ourselves to one type of collaboration or social

<AnnBassetti_> AnnB: responding to Rahn, big companies will not give up their enteprise processes any time soon

<AnnBassetti_> Beth Lavender, Mitre: what standards might we focus on?

<AnnBassetti_> what record have I just created / uploaded

<AnnBassetti_> finding security

<AnnBassetti_> common identity

<AnnBassetti_> ------------------------

<AnnBassetti_> break /

<wseltzer> [coffee break]

<wseltzer> [return at 11:30 Pacific]

<James_M_Snell> As far as "what standards" are concerned... the critical immediate task is: How do we teach our tools to better recognize and understand Context and Identity. The standards in this area are lacking most. We have OAuth for addressing part of it, but the Web Platform itself does not offer any inherent support for Context and Identity.

<James_M_Snell> which is why things like OpenSocial exist... to add in that notion of "Who am I"

<tantek> James_M_Snell: Are you in the room?

<James_M_Snell> sadly no

<James_M_Snell> couldn't make it

<tantek> Darn. Would've enjoyed some in person chatting.

<James_M_Snell> yes, me too

<James_M_Snell> if there's time later today, we could set up a quick hangout

<tantek_> hoping they can at least set up a hangout for tomorrow's demos

<James_M_Snell> hope so, we shall see

<James_M_Snell> fwiw, I'm hoping to make a run up to SF in September for a couple of days. We ought to organize an informal jam session

<gkellogg> scribe: gkellogg

Use-cases and Requirements

<wseltzer> scribenick: gkellogg

hhalpin: this session about use cases and business examples.

wilkinson: I'm head of engineering at Crushpath, creating a social platform

<LloydFassett> demo post

… at myspace, we integrated with Microsoft Live product, and worked with a number of people working on exchanging social network data.

… Facebook deployed a version based on ???

… We end up switching to JSON

… SocialCast used an ActivityStreams engine, based on a RDMS

… This started collaboration on OpenGraph Protocol. Many customers were using a proprietary tool, but it needed to work across a variety of platforms.

<hhalpin> note that Open Graph Protocol enables the Facebook Like Button

… CrushPath is designed as an advertising tool, but shifted to gaming business.

… Key is a "pitch"

… Example is a business profile, but could be for hiring engineers. Focused on allowing small business to get started with marketing.

… Considerations when building: users are not high-tech users, so they're wary of sharing information, so focus is on client-side integration.

… We wanted people to be able to login with sites they're familiar with, but concerned that it might be a "weird" experience.

… E.G., google will ask people to create a Plus profile, and they might not be familiar with that.

… We use Activity Streams everywhere, as it's a good design point.

… Not a problem that FaceBook uses their own protocol rather than AS.

… We use PubSubHubub, allowed reuse of existing tools, rather than rewriting from scratch

… Standards help people ship faster: there's a lot of collaboration.

… We use Twitter Cards because we need to, even though they don't correspond to standards.

… the OGP vocabulary is closed, but useful.

Meeks: Work on combining linked data with open social

… UCSF profiles show a lot of data about researchers, much like LinkedIn

… Many organizations do something similar.

<James_M_Snell> native browser support for access to social data is critical

… Contains deep researcher data, based on open source code created by Harvard

… UCSF was first group out of Harvard to use the code

… Added OpenSocial to allow for extending the model for our purposes

… Have Youtube videos on website, as researchers find that really useful.

… We've been pushing other institutions to adopt OpenSocial, but there are challenges:

… Wake Forrest needed to get at researchers keywords; we came up with different ways of doing the same thing.

… The other area is on the use of the Linked Data standard; there was some perchption that it competes with OpenSocial, but they don't.

… LOD is supported by a lot of different solutions. Linked Open Data is like a machine readable version of a website; it's about expressing semantic data on the web

<tantek> "t's about expressing semantic data on the web" (in a way that's much harder than microformats)

… VIVO is a competitive product based on RDF and gaining traction.

… We have OpenSocial on one hand and LOD on the other, how to put together?

… There are ways to express this in different RDF formats, but they're not simple.

… First thing we found was an RDF/XML to JSON converter, but it wasn't ideal and wasn't a standard.

… Meanwhile, JSON-LD is being developed that solves exactly these problems, and that's what we support now.

… Added an OSAPI.rdf adapter to allow easy extension of OpenSocial to JSON-LD/RDF

… JSON-LD allows the data model to be de-coupled from the API, this is a big deal.

… JSON is really easy to work with, and not intimidating developers is important.

… (VIVO's an extension of FOAF)

… Because JSON-LD is so easy to work with, we think more people will embrace the RDF data model

Boyet: Social Networking within Boeing

… "Connections in all directions" is our catch phrase, we want connections everywhere.

… We have been working on inSite a long time with about 85K profiles

… That's about 1/2 of the whole company.

… Culture is the biggest obstacle, not technology.

… This allows us to find experts across the company; they can build up trust networks and make relationships.

… Create formal and informal teams, may require VP approval.

… We want people to collaborate in the open rather than behind closed doors.

<aaronpk> Did anybody catch the URL of his PDF on their inSite system from 2009?

… InSite built from scratch in Java and Apache. This allows users to work from use cases and turn them around, allowing Boeing to be responsive.

… It's integrated in with enterprise identity systems, meaning that you don't need to log in to use it.

… Because it's behind firewalls, content can be indexed from other systems.

… Boeing has large installations of many different collaborative technologies (SharePoint, Exchange, Wiki, ...)

… We're researching social collaboration on the factory floor; they don't have their own devices and share workstations.

… Culture much different than people working at desks.

… External customers want to use InSite for their own purposes.

… Challenges are to make InSite profiles available to other applications; want the user profile to travel with them as they move around the company.

<AnnBassetti> it's not so much that external customers want to use inSite .. rather that we'd like to be able to 'socially' collaborate external using inSite

… Expertise profiles are not just user-declared, but can be asserted by peers.

… How to get out of sharePoint and into other systems?

… We've created a simple web framework to allow anyone to create presence information from anywhere. Through ActivityStreams and a profile widget that you can add anywhere.

… Finally, authentication is a nightmare

… It's a big barrier for products to exchange information.

… Examples of searching for people by expertice

Ding: Memory Connected: Extending Memory on the Web ...

… Started from big data created by humans

… Walled gardens have kept this formation from getting out.

… Our brains can host about 2.5 Peda-bytes, but information is contained in silos with small amounts of information

… Leverage laptop to bridge the silos

… Eventually, leverage a cell phone to extend the brain (through access to information)

… Working from human-centered design. Things like RDF allow user to tell the computer what they want, but we also need to make sure knowledge is readable and sharable between users.

… developers favor data formats which are more readable (Turtle vs RDF/MXL)

… Exposing everything through URLs may be diffcult

… For example, we want to read our emails, headers are typically hidden.

… Just-in-time knowledge scheme.

… Better ways to access data than keyword search

… Tangible rewards and cost reduction

… Make it easy to do things and have an instant reward

… tweets are instant, blogs take much more time to be digested

… SCRUM process to allow issues to be solved in a short amount of time.

… spreadsheets are about the best tool created so far, we need to learn from this interaction method to give immediate rewards.

… Bottom-up growth scales

… E.g., twitter hash-tag is consensus based and extremely effective

… JSON and CSV are well adopted

… Looking at mailing lists and realtime conversation.

… Also look at common knowledge within a team; are we reaching consensus, or is this a fantasy of the project manager?

… Vision is extended memory on a phone or laptop, which is organized to be more useful than it is now.

… Important to exchange this information between people through brokers (e.g., google)

Schutzer: Thoughts on the Requirements for Social Networking

<AnnBassetti> @aaronpk the link for Adam's original talk at TPAC09 is http://www.w3.org/2009/Talks/Boeing-tpac09.pdf

<aaronpk> thank you!

… social standards must meet business needs

… members don't usually build their own internal tools: rely on SharePoint, Twitter, etc.

… Customers are using tablets/mobile much more than laptops

… People often care multiple devices for personal/business needs

… People maintain contacts for themselves, not out of business loyalty

… security and privacy, risk and compliance are really important.

… We don't really have security/privacy controls.

… We like to learn about customers from social media, but need to protect information

… Need to protect customers _from_ social networks

… Email and apps may impersonate an business entity

… Risk and Compliance, necessary for fraud investigations and disclosure requirements

… If employees talk about stocks, this could cause financial/legal problems

… Employee's personal sites could also violate policy

… goals are required so that dominant social networks want to comply

… finding minimal standards we can agree upon really important.

… Privacy controls which are consistently implemented and imported/exported would work well.

… This allows a consistent way to talk about platforms.

Boyet: We investigated OpenSocial, but started before it was availble.

… It was simpler to use our own framework, and make it easy to install and administer.

… Makes it simple to get information out.

… Constantly updating, re-archtecting and open to change.

questioner: monica used terse, open-ended and extensible. These are great characteristics.

… Eric mentioned that having teenager information backed into basic profiles is an impediment to adoption. If you're corse entity is really terse but extensible, and let people develop they're own data models rather than baking them into the core model.

… Developers may have 2- or 4-week release cycles, we need to accommodate such rapid change.

<evanp> James_M_Snell: are you at the event?

<evanp> danbri: also, you?

meeks: the data model thing is a real problem, there is no one-size fits all.

<James_M_Snell> evanp: sadly no

… Getting to the linked data point in the first place isn't triveal

<James_M_Snell> re: data model... we don't need one size that fits all right now

<James_M_Snell> we need to start with: how do we access the data in a standard way, before we build enough common patterns for the data model

questioner2: we're struggling around identifying expertice, two people looking for the sam thing should see different things, because context is different.

… Is there a way to contributed identified information back to social network.

<tantek> process q? how do we get on the Q&A queue? Zakim?

<James_M_Snell> re: "one profile everywhere"... we don't need that either. Our user agents need to learn to understand who we are and to provide that info to our applications

<tantek> well, that didn't work, no Zakim I guess

… need to integrate collaboration tools

<wseltzer> James_M_Snell, interesting thought re context as a role for the browser

<tantek> "we have more lawyers than developers" … " don't quote me on that"

… companies are going to do what they need to do, even if social standards don't address.

<James_M_Snell> my mobile phone has a better sense of who I am than my employer or any of the individual cloud services I use

<tantek> and Lunch interrupt supercedes all other queues ;)

<tantek> <br class="h-x-lunch"/>

<wseltzer> heh

<wseltzer> that's quite a queue

Social Standards Architecture

<wseltzer> Bryan_Sullivan: challenges: security, privacy, data ownership: trust

<wseltzer> ... scalability in a global mobile environment

<Andrew_Mallis> ATT deployed the first mobile PUSH system in the US, based on WAP.

<wseltzer> ... user experience; identity management

<lloydfassett> Social Networking Web SNew

<evanp> What's SNEW?

<lloydfassett> it is based upon things that are already out there like Salmon PSHB, Ostatus

<wseltzer> ... Federation requires capability exchange, under control of user (privacy)

<lloydfassett> Challenges to cover include federation, privacy (and control of the user)

<wseltzer> scribenick: lloydfassett

Activity Streams templates to standardize some workflows

like NFC, Augmented Reality, 2D barcodes

OpenId Foundation has a lot of momentum behind it.

Multi-Factor identificaitons

expects to see more to come from the Open Identity Foundation

Application Resource Optiimizer is available to share connections on mobile phone. Created by AT&T

<evanp> lloydfassett: I can scribe also if needed

Fabio Mondin Telecom Italia presents

Social - Content Revolution. A Vision for the Future Social Oriented Networking.

<AnnBassetti> scribenick: lloydfassett

The FP7 eCOUSIN - content centric and social aware

to improve network efficiencies

4 use cases

Two individuals meet on vacation and want to share information

<tantek> if anyone here (at Tank 18) has their own domain and uses it, see me for a red sharpie (to mod your badge)

They want to share media files from different systems like video

Reduce network usage through social collaboration

3rd make the network location centric and not content centric

4th place content closer to final user in order to reduce network load, videos of a concert can be published and consumed in physically close locations.

Use cases common needs are monitoring, look-up and networking (making network aware of social content and how it will be consumed)

Jason Roygary, Senior Board Member on Open Social and IBM Collaboration

Talking about things around events and roles

IBM use case prototype showed pain points.

Mobile, events, roles and people are hot topics.

Motorola 1973 New York City, worlds first cell phone call. He called Joe Engle at Bell Labs to mock his competitors.

Use case, demonstrated at a conference, might be demonstrated tomorrow

Logistics company..keeps track of trucks and events in an Activity Stream.

Different people with different roles can have embedded events in the stream, in real time.

UX is through a mobile device.

They created a couple buckets, PSHB, LOD, Messaging payloads

they discrovered there are no role based profiles

MQRR was used for PSHB, JSON-LD for Linked Data

wanted to build apps easier, still wanted to deal with roles

thinks Roles is the killer app in Social Business

hasn't heard call for volunteers to figure out Roles. That will be a part of creating kick butt apps.

Presenter: Ashok Malhotra, Oracle, Property Graphs

we should write a standard for Property Graphs

Oracle isn't a social networking company, they create software.

They want to do useful things with the data social networks create

they require a data model standard

Socialn networds are using property graphs and "mine" their networks, some tools are available but

all of them have a different data model

declarative or navigational interface

Node/edge model, edges have properties, can be modelled in RDF

Example from Tinkerpop

Property Graph Model / a set of nodes and a set of edges

details are different between graphs

W3C should start a community to standardize property graphs, the other to standardize the vocabulary for relationships and properties

<wseltzer> Oracle is willing to create a member submission to seed the first CG (standard for property graphs)

Oracle is willing to create a member submission to seed the first CG

Keith from Cisco asked why RDF is not a good fit to Asok

Asok: it doesn't quit fit well because the arc has to have a property, which needs a subject. That doesn't feel natural

Can we modify RDF?

Asok would think about it for a while. Personally you could add stuff to RDF, possibly easier direction.

Kelvin from IBM; interested in graph topic...looked at opensource world and it's vibrant

we can't ignore work that's done before when looking at graphs

challenge is to find the right area to work on. he's interested in hearing more. How do I export profile information? Possibly through graphs.

Open Social has some solutions that could aply

commentor: how we define a property graph standard. hesitant to modify RDF. There's a reason people have developed other standards.

keeping it simple is important.

Asok: we're starting to speak about what we should do? Once it's started we can discuss the options

Commentor: Is there a more elegant way to deal with non-textual data?

Ed Krebs: He's on board with roles. None of us wears a single hat. How do you do context switching? Does the system have some intelligence?

<tantek> I'm a bit surprised anyone here is taking RDF seriously for anything. Actually not that surprised. But seriously?

Jason got Ed to volunteer to contribute to working on it. Jason is starting around individual apps

<tantek> is anyone who is "working on" something actually have something now? (URL?) or is this all hypothetical wishlist stuff?

The bigger model around multiple roles and contexts need to be worked through from a basic model.

<James_M_Snell> tantek: Yes, there are pieces that exist now

<tantek> James_M_Snell - citation needed

<James_M_Snell> http://activitystrea.ms

<James_M_Snell> :-)

Brian: not sure roles are specific..it ties in with identity and trust...it is an inhibitor to adoption

<tantek> I see a spec overview page that hasn't been updated in ages

<tantek> and nothing on http://activitystrea.ms/ actually itself uses ActivityStreams

<James_M_Snell> sigh.. yes, that is true

<tantek> so no - I dispute "exist" except maybe "on paper"

<James_M_Snell> there are bits and pieces that exist today, but no one has a complete picture

commentor (Cisco): roles are a fundamental part of the model. You have to take into account people have multiple roles or the fundamental model will not be applicable.

correction (TIbco)

the products he's involved in treat roles as something users can have more than one of

<tantek> James_M_Snell - as far as the public / open web is concerned - I see little to no actual deployment of ActivityStreams (and *I* myself have an Atom ActivityStream on my site)

Mark Weitzel to Boeing: how do you deal with roles?

<tantek> James_M_Snell: http://indiewebcamp.com/activitystreams#Lack_of_selfdogfooding

<James_M_Snell> tantek: yes, adoption has been light on the public / open web.

Boeing: they don't handle it or manage it. They have properties on the user.

<tantek> if it's not adopted on the public web, it's a dead end technology

They are all internal only roles to Insite (their app)

<James_M_Snell> we have product that uses it, and I know others are using it

<James_M_Snell> but it's mostly closed environments

<James_M_Snell> walled gardens

<tantek> plenty of ActiveX plugins in closed environments too

Jason thought the Boeing process of handling expertise could be an interesting part of a role.

AT&T: from a W3C perspective the way the internet works, people have relationships with advertisers and others. The scalability of managing roles is critical to making things work.

TIBCO: talked about information flow between doctor and patient. The kinds of permissions each person can do in these roles are limited. Same as in enterprise applications.

some information can be shared between doctors, it's different with other kinds of information and purposes.

<James_M_Snell> tantek: very true... and fwiw, I agree with you, there's too much spec writing and not enough real problem solving going on

Rawn Shawn: context can describe limits. We have roles and transactions in the process world. In the social world we know transactions exist, but there is a significant switching of contexts.

<Andrew_Mallis> Roles are being referred to in relation to agency (transactional), but they also are impactful when we speak about identity.

Jason: that is lesser of issues in the contexts we're looking at. We need to the right questions. We blundered into this problem by trying to solve a use case

commentor: how would RDF and OWL fit into this? You can have two IRI's for the same thing.
... provinence was discovered as important later.

Ellen: LDAP is a hidden db. you don't know who's role is what.
... the lack of transparancy in AD and LDAP are the problem.

Jason: I think it's possible that secret IT things may be coming to an end because the richness of profiles is creating significant pull for social.

<James_M_Snell> Role = Identity + Context

Wendy Seltzer: Transparancy, roles, context. What could we be doing to reinforce the users ability to recognize when context has shifted. W3c usually does not promote a UI.

<James_M_Snell> Role = Who am I at a given moment and Why

maybe we could display to users how they are participating.

Jason: I think about the implications of this in Germany. We get into interesting cultural challenges.

AT&T: we have effective aims that focus our attention on a particular application, the browser. Too much focus on that, more on expectations to help people to undersand your graph.

<bret> Why don't people use existing and widely deployed microformats instead of RDF?

<Andrew_Mallis> Role relationships would be important. Teacher to Parent, Student to Student.

eBay Collaboration: we built a collaboration and then it didn't work with BYOD. Why don't we have one browser so this pain will go away?

AT&T: Open Web Testing platform is where should look. W3C is building a comprehensive system to validate how apps work in browsers.

Jason: you just put fuel in your car. You don't wonder if it will work. We need standard testing.

ItalyTelecom: even one browser will not address multiple mobile devices.

<wseltzer> [W3C's standards testing efforts aim to make sure the technologies work interoperably on all browsers and devices. See http://testthewebforward.org/ ]

Mark thought the gas analogy was great, until everyone starts driving electric

Ann B: leaned on eBay to join W3C. We need participation.

Harry H: eBay is a member. Jeff Hodges from Paypal is managing it.

<James_M_Snell> We don't necessarily need participation for writing yet another spec... we need participation to build usable, useful common tools

<James_M_Snell> APIs

<James_M_Snell> browser-support

<James_M_Snell> common data models

<tantek> "There are companies that don't use HTML" [like the folks presenting PDF or Powerpoint today]

Harry: We need to ensure interop before we endorse a standard. There will be a test sweet (mentioned above).

<AnnBassetti> haha, Tantek .. on the money

<tantek> James_M_Snell - re: too many specs, not enough implementations. For this reason I don't believe in any more format efforts that do things with anything other than HTML.

<tantek> JSON is syntactic sugar for devs

<tantek> formats should be defined in HTML first, with a mapping to canonical JSON.

<tantek> defining formats in JSON first is putting the dev cart before the publisher horse

<bret> very much so

<James_M_Snell> tantek: I deal with a lot of third party integration that does not operate at the HTML/UI level

<James_M_Snell> at that level, an HTML based data format doesn't make sense

<tantek> more HTML publishers than all such third parties put together

<James_M_Snell> but I ought to be able to take the same basic data model and apply it in both JSON or HTML

<tantek> by far

<tantek> any non-HTML effort is fighting a losing battle

TIBCO: we use origin for security context within the browser. We shouldn't focus on language specific API's. The core is more around the payload, produced by Ruby and consumed by JS doesn't matter.

<tantek> JSON should be side-effect of well structured HTML

<bret> can someone put the irc log link into the topic?

<James_M_Snell> tantek: so what? that doesn't make the non-html cases any less valid

<wseltzer> Howard: Is personal/role context similar to origin-restriction? are we trying to help people protect themselves from context-phishing?

ASOK: I agree with you but some people would prefer their own hybrid interfaces

<tantek> it makes designing for them first less valid

<wseltzer> [break]

<James_M_Snell> Standard browser support for social ... +1

<James_M_Snell> window.social.whoAmI(callback)

<James_M_Snell> let the user-agent tell the application who I am and what role I'm currently playing

<James_M_Snell> window.social.whoIKnow('@friends',callback)

<James_M_Snell> let the user-agent tell the application who my network is based on what it knows about my identity and my role... all with my permission of course

<bret> James_M_Snell: have you seem mozilla's social API?

<James_M_Snell> bret: yes, bits and pieces of it at least. I need a refresher

<James_M_Snell> bret: have their been recent updates?

<bret> its the closes thing to what you are asking for that actually works right now

<bret> yes, it started as F1 (an extension) and now is in firefox as a supported API. Facebook supports it, if you want to test it out

<James_M_Snell> bret: that kind of functionality ought to be browser standard, with the data dependent on the users Identity, current context and role..

<James_M_Snell> so if I'm on facebook or a fb connected site, window.social.whoAmI returns my FB identity... but if I'm on IBM's Social business infrastructure and I'm using a third party integrated service, window.social.whoAmI returns my IBM social identity

<bret> James_M_Snell: the API is open to anyone to use. Im sitting next to someone who is using it to post onto their own website

<bret> In terms of getting it onto other browsers, its yet to be seen

<bret> but its the closest to anything *like* a social api standard

<bret> for browsers

<benwerd> (I'm that person, and it really should be a cross-browser standard. Just Firefox for now though.)

<James_M_Snell> bret: I'm more than willing to help do whatever I can to help promote and move that forward...

<James_M_Snell> bret: btw, I have a working prototype using HTML5 custom elements and shadowdom to enable a window.social API in chrome

<James_M_Snell> Using polymer polyfill

<James_M_Snell> import the custom element, drop <ibm-social /> onto the page, and the window.social API is enabled.

<bret> James_M_Snell: the developer behind the Social API from mozilla is here

<James_M_Snell> bret: sadly I'm only "here" virtually.. participating via IRC and twitter only

<bret> *doh*

<bret> I don't think he is on his computer atm.

<James_M_Snell> We need to take some time to reconcile things like the mozsocial manifest, schema.org/Action handlers, web intents, web finger etc... so user-agents can learn to better understand their users and the services they use

<James_M_Snell> then how to make that info available to apps through a standard browser api

Federating the Social Web

<wseltzer> [returning from break]

<bret> what lets my access that info? I missed that

<James_M_Snell> bret: right now, when an application wants to know your location, it asks the browser, the browser asks you, "Do you want to share?"

<bret> right, social api is a similar mechanism

<bret> a browser feature that website can access

<bret> or ask to

<James_M_Snell> browser then works with the user to determine identity, context, etc

<James_M_Snell> right now, if an apps have to do all the identity stuff themselves

<James_M_Snell> why not build OAuth support into the user-agent

<James_M_Snell> Google Services as an example...

<aaronpk_> James_M_Snell: I will be talking about that tomorrow actually

<James_M_Snell> aaronpk: +1

<wseltzer> Gregg Kellogg: LD: when you name things, name them using URIs

<James_M_Snell> bret & aaronpk: when I log into a site, and the useragent says, "Hey, do you want me to remember your identity?" ... the user-agent and the identity provider can interact to establish the right identity... more than just saving a user name and password

<hhalpin> e/me let me remind EvanP it's his turn to scribe

<bret> How is the json-ld data typically stored? Is it a mirror of html content, like rss and atom typically are?

<tantek> bret - who knows

<bret> Any examples in the wild that I can go look at?

<tantek> bret - good question

<bret> json obviously, is all over the place

<bret> but this -ld extension

<hhalpin> As fror the storage, I'd hope it would be stored in some graphstore like Mongo but not sure how messy this is. Seems like starting a keyword with (at) sign makes life a bit tricky.

<tantek> hhalpin - any real world examples of such? or is this is all hopeful handwaving?

<evanpro> Zakim: scribenick evanpro

<evanpro> I think

<James_M_Snell> don't get hung up on the -ld extension... focus on the json part

<James_M_Snell> any json can be json-ld depending on how you process it

<aaronpk> why is this different from returning HTML info and using <a> tags to link HTML documents together?

<bret> i see that -ld is an effort to standardize vocabulary, but for what

<James_M_Snell> json-ld to me means making use of a common data model

<evanpro> Graph definition allows specifying graph of JSON data

<bret> but how is it stored/accessed?

<James_M_Snell> well... for anything really.. using json-ld as a bridge, for instance, I can model schema.org data model in activity streams json

<evanpro> JSON-LD includes syntax for rich data types

<bret> and is it a replica of a web page? database?

<evanpro> Possible to express named graphs

<evanpro> Basic system is RDF

<evanpro> Example: Activity Streams 2.0 which is JSON-LD

<hhalpin> tantek - JSON-LD is still pretty new - thus, I think it falls in "handwaving" category. But it is already getting traction from Google etc.,

<evanpro> There are standard transformations between JSON and XML versions

<James_M_Snell> concrete example... {"@context": {"objectType":"@type", "person": "http://schema.org/Person", "displayName": "http://schema.org/name"}, "objectType": "person", "displayName": "Joe"}

<evanpro> json-ld.org has more information

<tantek> James_M_Snell - yes that's an example. nothing concrete about it

<tantek> anything you can find anyone publishing at actual URLs on the open web?

<evanpro> Ed Krebs "The need for a PubSubHub and how social plays into this need"

<bret> what is google doing with json-ld

<tantek> bret - I'm calling BS unless someone provides a URL. [citation needed]

<evanpro> Ed is using the term PubSubHub to mean a different thing from PubSubHubbub

<evanpro> Call it Ed's middleware

<tantek> "are you trying to say PubSubHubbub?" "No, I'm saying PubSubHub"

<tantek> (nervous laughter)

<evanpro> hhalpin: I don't know if I set myself up as scribe correctly

<bret> :]

<hhalpin> Given that JSON-LD is just JSON, Mongo/Couchbase that can take native JSON can also take JSON-LD.

<hhalpin> That is how I would store it at least.

<James_M_Snell> tantek: I've got this mental image of you as the grumpy old troll, lol ;-) ...

<evanpro> Example of PubSubHub: "Flip Style mobile reader"

<bret> What part of fliboard? the fold in the middle part or the rss reader?

<tantek> James_M_Snell - asking for citations is trolling?

<James_M_Snell> tantek: I'm being silly

<hhalpin> https://developers.google.com/gmail/schemas/reference/formats/json-ld

<evanpro> Uses information from business systems, which are sent to an enterprise social system

<bret> thanks hhalpin

<tantek> thx hhalpin

<evanpro> ... which in turn drives a mobile reader

<hhalpin> I'm not sure if there is *actually* momentum for JSON-LD. Obviously there is for JSON. And obviously graph-store dataformat need to be standardized. Nonetheless, the RDF data-model may or may not be perfect fit.

<hhalpin> I'd like to hear various opinions on this about how it is.

<tantek> hhalpin - yet to see an actual use case that needs JSON-LD above and beyond "just JSON"

<hhalpin> for end-developers re storing/using JSON-LD vs. just sticking to JSON vs. using whatever else.

<bret> The beauty of JSON was that it was already there before the name was applied

<tantek> and while you're doing "just JSON" - you can publish that trivially with HTML+microformats2

<hhalpin> I think the idea of JSON-LD is just sticking URIs into JSON

<evanpro> They push information about car part availability into their activity stream

<tantek> hhalpin - funny, sticking URLs into HTML seems to be much more well understood/adopted.

<hhalpin> The trick is via the @context, which I guess can just be defaulted in some way (perhaps in spec).

<evanpro> And potentially other application actions

<evanpro> Within the mobile reader they want to provide social actions

<tantek> hhalpin - sounds like the classic, don't understand the problem? add a layer of abstraction (@context).

<aaronpk> my HTML has a JSON representation: http://aaronparecki.com/notes/2013/08/07/3/osfw3c.json

<evanpro> e.g. approve or deny a purchase order

<hhalpin> JSON-LD is just adding URIs+links to JSON, which sounds sensible - so ideally, JSON can then be a data-version of the HTML. However, this does hit the maintenance issue that tantek correctly brings up and fixes re microfoamts.

<hhalpin> the question is whether or not it makes sense to push stripped down HTML with markup that is easily made into JSON (i.e. microformats) or just ship the data around direclty (JSON, JSON-LD)

<evanpro> Ed's middleware ("PubSubHub") provides identity, and can connect to the enterprise social platform

<tantek> hhalpin - exactly. much easier to maintain your data *once* in the HTML+microformats2 which automatically generates JSON, than maintain it *twice* in HTML and JSONLD.

<hhalpin> I'm OK with either to be honest, but would like to hear more pro and against each approach.

<hhalpin> Yes, I have long-agreed with your opinion on that tantek.

<evanpro> Includes some roles and access control

<bret> aaronpk: can you add an object with a key pair that describes how the json was generated (IE "generated with mf2php from my html")

<tantek> hhalpin - stripped down HTML with markup that is easily made into JSON (i.e. microformats) works and easier to debug in browsers *today*.

<aaronpk> bret: look at the bottom :)

<James_M_Snell> json-ld let's us very easily map the same microdata data in html and json without jumping through a bunch of hoops

<evanpro> "You can share to here but not to here"

<tantek> hhalpin - why is it so hard to communicate the maintenance problem?

<bret> derp

<tantek> my guess is that folks advocating duplicate data have very little experience with actual maintenance of real world content/systmes.

<tantek> *systems

<bret> My atom feed constantly breaks as I develop my site

<evanpro> PubSubHub system is extensible such that new business systems can feed data into the middleware without changes to all the rest of the servers

<bret> Im down, lets use pubsubhubbub

<evanpro> Sam Goto talking about Action schema in schema.org

<evanpro> Information on http://schema.org/Action

<evanpro> "a taxonomy of verbs"

<evanpro> The requirements for actions: model past, present, future activities

<evanpro> model discovery and execution of activities

<bret> are these based off of Activity streams?

<evanpro> bret: no. yes. maybe.

<tantek> sounds like a fork

<bret> o.0

<tantek> wow mentioning web intents (which Google killed months ago)

<evanpro> similar to web intents or web actions or soap

<evanpro> <sorry for the editorializing>

<evanpro> It would be good for Google to know that there are actions allowed with Web services

<evanpro> e.g. Netflix is for watching movies

<bret> oh right, gmail json stuff from IO this year

<evanpro> use case: rich experience in Gmail

<evanpro> use case: Google+ interactable activities when sharing a site

<bret> what is more human readable, :p G+ redesign or the JSON representation (There is one right?)? Only teasing

<James_M_Snell> schema.org/Action was going to be a fork at first

<James_M_Snell> then there was push back

<James_M_Snell> And collaboration

<James_M_Snell> Cooperation

<James_M_Snell> etc

<James_M_Snell> Beer

<bret> thats good

<James_M_Snell> you know, important stuff

<evanpro> use case: search

<bret> :)

<bryan_> there are two "action" entries under schema.org, e.g. http://schema.org/action is about muscles... confusing

<evanpro> use case: google now can include action information

<James_M_Snell> OpenSocial calls this "Embedded Experiences" .. has their own model for it

<James_M_Snell> that needs to be reconciled

<evanpro> Goals of Action schema:

<evanpro> Rich set of nouns

<evanpro> Disambiguation

<tantek> schema is non-starter due to lack of RF patent for consuming schema

<evanpro> Structure for synonyms

<evanpro> Contextual arguments

<evanpro> Execution and handling

<evanpro> Structured composition

<evanpro> Principles:

<evanpro> verbs are types on actions

<evanpro> They have arguments - semantic roles for the verb

<James_M_Snell> tantek: yeah, that's definitely a problem

<evanpro> They involved linguists in the process

<tantek> you have rights to publish only

<tantek> read the TOS

<evanpro> And developed a big hierarchy of verbs

<tantek> http://schema.org/docs/terms.html

<bryan_> that's a lot easier than just saying "i bought outliers on amazon"

<James_M_Snell> tantek: that's a large part of why trying to rely on a one-size-fits-all data model isn't going to work yet

<tantek> not sure what you mean by that James

<evanpro> OPENi is for innovation in mobile applications

<evanpro> improves interop

<evanpro> improves the trust model

<evanpro> and brings new open source app framework

<evanpro> MyLife app is for consumer social data

<evanpro> OPENi also works on advertising framework

<evanpro> And personalized in-store shopping experience

<evanpro> Learned from their analysis: too many APIs

<James_M_Snell> tantek: I mean, some folks wanna use schema.org, others want to use {insert data model name here}.

<evanpro> >140 APIs reviewed

<tantek> James_M_Snell - SEO spammers use schema

<tantek> there's no ecosystem there

<tantek> because no one else is allowed to consume

<James_M_Snell> for some, schema.org is a non-starter, for others, {insert data model name here} is a non-starter

<tantek> it

<evanpro> They're providing a single API

<evanpro> And include privacy by design

<tantek> James - except others have open licenses

<tantek> whether JSONLD, RDFa, microformats etc.

<tantek> for publishing/consuming/parsing etc.

<tantek> so this is a very different problem

<evanpro> OPENi framework expects consumers to hold own data in "cloudlets"

<evanpro> developers can use data based on consumer opt-in

<evanpro> And put them all in a single structure

<evanpro> OPENi classified multiple APIs

<James_M_Snell> tantek: yes, my point is that whatever standardized social apis we have, or whatever "standardized tools" are provided, they need to be independent of the data model, and independent of any possible licensing issues those models are tied to

<tantek> James_M_Snell, anyone using or advocating for schema-org examples/markup/vocabulary is supporting a Google patents-reserved position. Not open.

<tantek> Has no business in any W3C discussions.

<evanpro> They use Graph API for identity, and Activity Streams to address them

<evanpro> JSON format

<tantek> nor examples

<tantek> nor specs

<tantek> search w3.org/TR for "schema.org" and find the violators.

<evanpro> It includes context API

<evanpro> App can define a context

<evanpro> Questions

<evanpro> Questions for PubSub: XMPP considered?

<evanpro> Ed: XMPP is one candidate for what he's doing

<evanpro> XMPP is a good example for federation: technical and business issues come up early

<evanpro> IM is a good reference point

<bret> I lost all faith in xmpp when google dropped it

<bret> but I hear it is a mess when it comes to the actual spec vs all of the different implementations

<bret> Pubsubhubub is Pubsub for the web

<evanpro> q: why should vendors get on board for federation?

<evanpro> Different kinds of commercial service differentiate on some features

<bret> I want to know that too evanpro.

<evanpro> Doing a poor job of it I think

<evanpro> bret: I'm trying to scribe

<evanpro> Q: how important is context to the audience?

<evanp> Zakim: scribenick evanp

<evanp> I hope that's the right way to do that

<tantek> really don't understand all these "context" discussions

<evanp> tantek: is it something technical that I don't understand?

<evanp> ISTR there is some security term

<evanp> Federation user experience: is it possible to share between servers

<evanp> Some sharing of data between systems is possible, and you can fake it

<tantek> "most important thing is that it's shipped" [about federation]

<tantek> "user-centric models"

<tantek> "so we can look at the important stuff, models of interaction, without having to worry too much about specs on the front-end"

<evanp> tantek asks that federation standards have open IP standards

<tantek> evanp - unlike schema

<tantek> per: http://tantek.com/2013/219/t19/osfw3c-contrast-microformats2-cc0-owfa-ogp-schema

<evanp> audience: federation should allow distributed control

<AnnBassetti_> question: who supports / likes schema.org

<AnnBassetti_> should've been a "don't know" vote taken ...

<James_M_Snell> Supports / Likes any given {thing} is not particularly useful or helpful

<evanp> my question: relationship between Action schema and activity streams

<James_M_Snell> Evan: There is no direct relationship between Action schema and Activity Streams

<evanp> sam answers based on some discussions

<James_M_Snell> Activity Streams 2.0 fixes some things and aligns with JSON-LD

<evanp> probably some alignment

<James_M_Snell> via JSON-LD, you can use the AS 2.0 syntax to represent Schema.org defined concepts

<evanp> (hard when I'm asking questions and scribing at the same time)

<James_M_Snell> evanp: understood

<James_M_Snell> but Activity Streams 2.0 != schema.org/Action

<James_M_Snell> you can use AS 2.0 without *any* reference to JSON-LD or Schema.org

<evanp> opening later is poor

<James_M_Snell> and you can use Schema.org/Action without any reference to Activity Streams 2.0

<evanp> tantek points out that patent license for schema.org

<evanp> tantek says that discussion should happen on the public system earlier

<evanp> does not grant license for consumers

<AnnBassetti_> +1 to Tantek pinging Google on openness

<James_M_Snell> Ann: +1

<evanp> monica asks: does schema.org want to work on activity streams?

<James_M_Snell> the ActivityStrea.ms group is open to all

<evanp> question of activity streams uptake

<James_M_Snell> anyone can join and participate

<James_M_Snell> anyone can implement

<James_M_Snell> it ought to be noted that schema.org/Action proposal started out very different

<evanp> James_M_Snell: good point

<evanp> We also have explicit patent license for work done

<James_M_Snell> **IF** AS 2.0 does get published as an RFC, that wouldn't change

<evanp> harry points out that some work can be done to allow further use

<James_M_Snell> I don't believe anyone in the Activity Streams community has any desire to *tie* AS to schema.org or any other particular model

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2013/08/08 00:09:49 $