Social Web and schema.org breakout

29 Oct 2014

See also: IRC log




<jtauber> elf-pavlik: Ann is setting up now

<jasnell> for those accessing remotely, we are starting a breakout session discussing schema.org/Actions and AS 2.0 Actions overlap and coordination

<sandro> elf-pavlik, AnnBassetti is doing that

Sam Goto I work at Google. We've been working on completed and potential Actions for some time

<AnnBassetti> trying to get there elf ..

Sam Goto: past actions are the closed to AS work

<jtauber> http://blog.sgo.to/2014/09/schemaorg-actions-implementations.html

<Ralph> Sam cites https://developers.google.com/+/api/moment-types/listen-activity

Sam Goto: aggregation plus past activiteis

Sam Goto: toward potential action in music industry

<AnnBassetti> http://blog.sgo.to/2014/09/schemaorg-actions-implementations.html

query Google with Lady Gaga, which is in KG

Landing pages have potential actions, in microdata, Google is agnostic, json-LD is good too

from the noun you reach the potential action, which is "Listen"...via going to that URL.

Sam Goto: We're designing this for the mobile wb


Sam Goto: demoing launched products...why don't we embed schema markup in Gmails too?

Sam Goto: build rsvp's in line

Sam Goto: markup shown in presentation

Sam Goto: property in noun 'potential action'

Sam Goto: a couple weeks ago we launched search on on SERP pages on Google

Sam Goto: noun has search action associated with it

Sam Goto: it's a semantic goal, a binding

Peter: why do we care about so many Google products?

jsnell: yes we do, I can demo deep linking in AS

<inserted> Schema.org and Social WG breakout session proposal

<Loqi> wseltzer meant to say: Meeting: Social Web and schema.org breakout

Sam Goto: I think I ran out of examples too

Sam Goto: demo's Yandex implementation

James Snell: presenting

<aaronpk> wseltzer: s/././ will regex based on the last thing you said. but doesn't replace other peoples' text

James Snell:Social WG just started up. I don't care about details of gadgets, but we took the use cases for in context actions

<aaronpk> what is the usual behavior?

<aaronpk> I didn't see anything else in IRC happen

James Snell: what do we basically need? A part of that format lets us represent things through a basic syntax

<AnnBassetti> James Snell: http://tpac.mybluemix.net/as2.html#19

James Snell:and attach actions to it

James Snell:this is very similar to Schema...it's essentially the same model

James Snell: demo's a REST call. We can share static content. Can do a browser view, deep link to a native app in IOS. The models are close at this point

James Snell: When we first started working on this a year ago the models were divergent.

James Snell:when I had beers in san Francisco with Sam Goto a year ago it was more different

James Snell: the protocol bits can be similar, without the vocabulary.

James Snell: we can have multiple actions on an object with JSON-LD with a Link Data vocabulary

James Snell: we can describe the payload coming back

James Snell: the technical differences are the vocabulary

Sam Goto: it looks like the Hydra model. Comparing the two could be constructive

James Snell: Indie Actions, primarily an html based using a custom component

James Snell:when we talked about it yesterday we realized it was similar too. We're getting a lot of convergence on the models.

Harry H: From a data modeling standpoint, Sam can you explain how you came to your process

<Ralph> [James mentions the related work in the -> http://www.w3.org/community/hydra/ Hydra Community Group]

Sam Goto: We asked what is the vocabulary you need to complete an action?

<hadleybeeman> Us there a more recent update on this? http://www.hydra-cg.com/spec/latest/schema.org/

Sam Goto: if you wanted to reserve a hotel room, what do you need to ask?

Sam Goto: Hydra and AS inumerates the input requirements

Sam Goto: it felt awkward for complicated use cases

Sam Goto: how difficult would it be to develop

Sam Goto: Java Abstract Classes are the thing that maps to my skills set. That's the thing Schema.org went went.

Sam Goto: YOu partially instantiate an abstract class. We are not married to this approach. If you have better ideas, we're open to it?

Danbri: How did we do?

Sam Goto: what I like the most is the notiion of attaching an actino to a noun

Sam Goto: You can share without knowing what you're sharing. Start with a noun and attach a verb to it.

(person) is this fire and forget?

<AnnBassetti> Ken Laskey: are these single step fire and forget actions?

Sam Goto: It is not.

<AnnBassetti> ...

Sam Goto: once you know the end point, if they return a potential action, they can build it

<AnnBassetti> Ken: you're taking atomic actions, fire and forget, but you take them and string them together

<AnnBassetti> eg. reservation making

<AnnBassetti> then you get another obje ct with another (potential) action with another question\

<AnnBassetti> sam: as you are making a reservation, perhaps marriot return a reservation that can be cancelled

<AnnBassetti> or confirmed, cancelled, shared etc

KenLaskey: you're taking potential actions, then we can get a response wiht another question. Marriot can return an API could be to cancel, confirm or share a reservation.

James SNell: the potential actions travel with the data. we can continue the work flow by passing the nouns around

<Loqi> Ralph meant to say: [James mentions the related work in the -> http://www.w3.org/community/hydra/ Hydra Community Group]

<AnnBassetti> hadleybeeman, wave your arms at harry

Peter: What does this have to do with Schema? James Snell: with actions

Tim Berners: Some ontologies come with with the protocol implicit.

Peter: THis is non-discoverabel

Sam Goto: WE could do a better job of documentation

Guha: We lookign at creating stable snapshots
... we push out a new version twice a month

Danbri: I disagree. In the last year we've tweaked the text.

<pfps> My issue here is how does the meaning of actions interact with schema.org. I don't see any way of getting from schema.org to the meaning of actions as things that can be invoked.

James: Right now the Schema.org actions is declaring the action. Can you work with the WG on that so we have a consistent model?

<lehawes_> Change nick to lehawes

Danbri: The protocol.

Guha: There is 750K websites using schema markup. We have to justify changes.

James Snell: we agree it's better to have one model than two

Sam Goto: we work with Markus Lanthaler from Hydra to incorporate his comments.

James Snell: at least on the Actions, I encourage you to come work with the WG.

<Loqi> sandro meant to say: q+

Danbri: Can we publish a Note with W3C?

<sandro> Loqi, you're messed up

<timbl> http://schema.org/docs/actions.html

Handley: please publish stable snapshots. I advise the UK government and we need a stable thing to develop to.

Hadley: That's all of a plea for standards work.

<Loqi> wseltzer meant to say: aaronpk, on other channels, s/A/B/ works on the prior expression anywhere in the minutes

Hadley: sequence of actions, is there a benefit of stringing actions together, or have a unique id in the datamodel

Sam Goto: the changllenge with API's is that you have to read documentation to understand them. The work we're doing is to make it that a machine can get a JSON payload and understand what can be done with that payload.

<elf-pavlik> affordances can also change depending on who makes the request!

Sam Goto: you don't ask a user to read a manual to browse the web.

Hadley: that's helpful, but Tim's point about machine readable would be helpful

Sandro: The main difference is how the data shapes are handled.

<Lloyd_Fassett_> Sam Goto: We have a website that describes the languages we looked at Sparkl, resource shapes, Hydra html5 forms. Our process has been documented.

<hadleybeeman> s/sparkl/SPARQL

<Loqi> hadleybeeman meant to say: ack me

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

<aaronpk> heh k i can disable that here

<Lloyd_Fassett_> Sam Goto: the intruction of semantic roles is different with Schema and AS

<Lloyd_Fassett_> JAmes SNell: AS jsut focuses on actions without going into the roles

<Lloyd_Fassett_> James Snell: they are different but compatible

<Lloyd_Fassett_> Sam Goto: Because of the tree you can just understand parts of the tree, you don't have to understand all the actions. That's important programatically.

<Lloyd_Fassett_> Sam Goto: you can be as specific as your users want it to be.

<Lloyd_Fassett_> Peter: You have to incorprate a lot of Schema when you incorporate Actions.

<Lloyd_Fassett_> James Snell: There is an inheritence, Tantek points out that in Schema volcanoe's can have fax machines

<Lloyd_Fassett_> Danbri: countries can have opening hours too

<Lloyd_Fassett_> James Snell: from a purely rdf point of view you get some weir artifacts

<Zakim> pfps, you wanted to note that schema.org actions bring in much more of the schema.org vocabulary

<pfps> if you use schema.org actions you get that the agent of an action is either a person *or* an organization, this brings in a large chunk of schema.org vocabulary plus disjunctive ranges

<AdamB> just a heads up, there is about 15 minutes left

<goto> hey all, this is the documention i mentioned to you right now with regards to the options we looked at for IDLs. check it out: http://blog.sgo.to/2014/03/rows-and-idls.html Feedback welcomed.

<elf-pavlik> hi goto, thanks for joining! :)

<Lloyd_Fassett_> Tim BL: in the past Schema and rdf...it would be wonderful to have a schema that maps to rdf. The reason schema has funny artifacts compared to rdf is that they work differently. There is s lots of useful information inside schema that rdf could use.

<goto> here is the list of examples i showed earlier http://blog.sgo.to/2014/09/schemaorg-actions-implementations.html

<KevinMarks> hm, OK I'lll try firefox

<Lloyd_Fassett_> Tim BL: when I use rdf I can dereference and pick up very useful information. In the RDF world we have one version of a thing.

<Lloyd_Fassett_> Danbri: I don't know if everyone is following

<Lloyd_Fassett_> Danbri: We keep changing the schema

<KevinMarks> got sound in firefox

<Lloyd_Fassett_> Tim B'l: Nobody is askign you to make claims you don't want to make.

<danbri1> searchj for volcanos with schema.org: http://danbri.org/2014/cse/volcano.html

<danbri1> it seems to confirm an absence of fax numbers

<Lloyd_Fassett_> Harry: we have ten minutes left. We had previously had an concern with IPR. Thanks to Scott Peterson, we now just point to W3C's policy

<Lloyd_Fassett_> Harry: We have really good proof that schema.org is responsive and we should be happy that that work happened

<KevinMarks> http://schema.org/Volcano has faxNumber as the 5th property listed, danbri1

<Lloyd_Fassett_> GUha: we have to figure out what we call a snapshot

<Lloyd_Fassett_> Guha: we'll figure out what we can do on that

<Lloyd_Fassett_> Danbri: the point Hadley brought up, we need to let people point to the version they are using.

<Lloyd_Fassett_> Hadley: thank you

<Lloyd_Fassett_> Hadley: We use standards for interoperability that hold us back so we can swap data...it's a messy situation that we deal with in our world

<Lloyd_Fassett_> Guha; schema is a live ecosystem that we get feedback on

<Lloyd_Fassett_> James Snell: there are model issues people have with Schema. Snapshots will make it significantly easier. If we can get that, great.

<Lloyd_Fassett_> Taisuke Fukuno is schema the world's hottest vocabulary?

<Lloyd_Fassett_> Danbri: there are a lot of microformats out there.

<Lloyd_Fassett_> Harry: there is more and more microformats on the web.

<Lloyd_Fassett_> Danbri: it's not a competition.

<Lloyd_Fassett_> rtroncy is there a plan for multiple language

<KevinMarks> it is annoying for publishers to have to add multiple types of markup to get previews on different silos

<elf-pavlik> microformats as in http://microformats.org/ ?

<Lloyd_Fassett_> Guah: the plan is to be in ASC

<Lloyd_Fassett_> Guha: you can translate it.

<Lloyd_Fassett_> Tim BL: When you dereference it does it come in english?

<Lloyd_Fassett_> Sam Goto: Elf has been contributing a lot

<elf-pavlik> once i can use WebID instead of state ids :)

<Lloyd_Fassett_> Tim bL: These actions are a part of a work flow, when you go through a process, when there is a problem do you have a plan to make a protocol around these situations?

<yusuke> on real i10n word-to-word translation may not work as expected -- atomic concept is different by language and culture especially

<yusuke> ... especially for 'actions'

<Lloyd_Fassett_> Sam Goto: The RESTful Hydra community has been thinking about that.

<Lloyd_Fassett_> Danbri: We've made some progress.

<KevinMarks> hadleybeeman: harry was talking about wanting domains that aren't rented for IDs - is that something interesting at the govt level

<goto> fyi

<AnnBassetti> OK .. bye elf and Kevin

<goto> here is another piece i wrote on hypermedia apis

<goto> http://blog.sgo.to/2014/04/what-your-api-would-look-like-as-webpage.html

<elf-pavlik> thank you AnnBassetti enabling remote participation :)

<wseltzer> [adjourned]

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2014/10/29 17:17:56 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.138  of Date: 2013-04-25 13:59:11  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/Sam Goto: Peter:/Peter:/
Succeeded: s/(person)/KenLaskey:/
Succeeded: s/Markus FAlker/Markus Lanthaler/
Succeeded: s/enouraged/advise/
FAILED: s/sparkl/SPARQL/
Succeeded: s/reseach shapes/resource shapes/
Succeeded: i|Meeting: Social Web|-> https://www.w3.org/wiki/TPAC2014/SessionIdeas#Schema.org_and_Social_WG Schema.org and Social WG breakout session proposal
Succeeded: s/preop/interoperability/
Succeeded: s/(person)/Taisuke Fukuno/
Succeeded: s/(person)/rtroncy/
No ScribeNick specified.  Guessing ScribeNick: Lloyd_Fassett
Inferring Scribes: Lloyd_Fassett

WARNING: No "Topic:" lines found.

WARNING: No "Present: ... " found!
Possibly Present: AdamB AnnBassetti Arnaud Danbri Guah Guha Hadley Handley Harry JakeHart James Ken KenLaskey KevinMarks Lloyd_Fassett_ Loqi MarkCrawford Norm PaulJeong Peter Ralph aaronpk azaroth bblfish bryan chaals chaals1 danbri1 dret elf-pavlik ericstephan fabien-gandon fjh goto hadleybeeman inserted jasnell jasnell_ joined jsnell jtandy jtauber lehawes lehawes_ nicolagreco pfps raphael rayd rhiaro_ sam sandro scottp social tcole timbl timbl_ wseltzer yusuke yusuke1
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Got date from IRC log name: 29 Oct 2014
Guessing minutes URL: http://www.w3.org/2014/10/29-social-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report

[End of scribe.perl diagnostic output]