<manu> trackbot, start meeting

<trackbot> Date: 22 April 2010

<Steven> ack 035

<dongmei> Hello, mit528 is me

<Knud> that's me

<Knud> I didn't see I was muted, kept on talking...

<manu> scribenick: RobW

<manu> Scribe: Robert

manu: skipping action items today, will only cover when needed rather than every meeting.

<Steven> Yay!

<Knud> whohoo

<manu> Facebook's Open Graph Protocol: http://developers.facebook.com/docs/opengraph

<ShaneM> It was actually covered on US National Public Radio this morning

<manu> http://opengraphprotocol.org/

<tinkster> Do they use a proper RDFa parser, I wonder. e.g. does <span property="og:foobar">...</span> work identically to <meta>?

manu: "Great News" -- Facebook seems to have adopted RDFa in their Open Graph Protocol, published yesterday, http://opengraphprotocol.org/, calls out RDFa by name. Several partners and content providers.

<Zakim> ShaneM, you wanted to ask a question about og

<manu> ACTION: Manu to talk to Thomas Roessler about Facebook announcement [recorded in http://www.w3.org/2010/04/22-rdfa-minutes.html#action01]

<trackbot> Created ACTION-25 - Talk to Thomas Roessler about Facebook announcement [on Manu Sporny - due 2010-04-29].

<manu> ACTION: Manu to start discussion about RDFa with Facebook [recorded in http://www.w3.org/2010/04/22-rdfa-minutes.html#action02]

<trackbot> Created ACTION-26 - Start discussion about RDFa with Facebook [on Manu Sporny - due 2010-04-29].

<Zakim> manu, you wanted to talk about their parser

<Zakim> manu, you wanted to move on

Raw triples via DOM API

<Benjamin> http://www.w3.org/2010/02/rdfa/drafts/2010/ED-rdfa-dom-api-20100421/js_docs_out/index.html

Benjamin: Report on some changes in last week

<Benjamin> this is the jsdoc Document

<Benjamin> http://www.w3.org/2010/02/rdfa/drafts/2010/ED-rdfa-dom-api-20100421/rdfa_dom_api.js

Benjamin: Reduced hiearchy of resources, removed inner abstract nodes, leaving leaf notes which can be instantied via API

<Benjamin> rdfa.foaf.name

Benajmin: Add vocabulary mapping for convenient access to URIs, e.g., shortcut rdfa.foaf.name for FOAF vocbulary

Benjamin: Added RDFa object type, based on idea of Mark, JSON-like to access properties

<Benjamin> rdfaObject[rdfa.foaf.name] result in "foo"

<manu> var person = rdfa.extractObject("http://example.org/foo#me");

<manu> person[rdfa.foaf.name] == "Jane Doe";

manu: Agrees (for the most part) to Benjamin's changes. More disagreements with Mark, which we should discuss.

<Benjamin> Current version of the spec is: http://www.w3.org/2010/02/rdfa/sources/rdfa-dom-api/

manu: Trying to make API as "native" to JS and other scripting languages as possible.
... One disagreement is that Mark does not want to add raw triples. Mark wants object projection. Raw triples gives more fundemental access for building on.

<Knud> can't we have both raw triples and projection?

<Benjamin> that's what we did with RDFaObject

<tinkster> ivan++

<tinkster> If you can get at the triples, you can build objects that match your own data model from it; the inverse is not necessarily true.

ivan: Wants possibility to get raw triples. We need to remmeber that RDFa is only one serialization of RDF. There may be other approaches in addition, but no comment on these.

manu: agrees with tinkster, that raw triples at lower level enables building other models on top of it

<ShaneM> +1 to both

<Knud> +1 to both

Benjamin: current RDFa object, allows both triples and project

<manu> PROPOSAL: RDFa DOM API should provide both a raw triples interface and a projectionion-based interface.

<tinkster> +1

<manu> +1

<Knud> +1

+1

<Steven> +0

<Benjamin> +1

<ivan> +1

<ShaneM> 1+1

Steven: Can live with resolution
... not opposed

<manu> RESOLVED: RDFa DOM API should provide both a raw triples interface and a projection-based interface.

Native language objects via DOM API

manu: One issue with RDFaObject in the projection of data onto JS object. But what about when object has multiple objects for same predicate?

<manu> person[foaf.nick] === ["joe", "joeb", "joey"];

<tinkster> RDFaObject could be renamed "Resource" (like RDFS) or "Thing" (like OWL).

<manu> if(person[foaf.nick] instanceof string) ...

<manu> person[foaf.nick] instanceof DOMString

<manu> person[foaf.nick] instanceof Array

manu: The problem is that we don't know if the container, e.g., person[foaf].nick is a DOMString or Array, multiple or single value.

<manu> person[foaf.nick][0]

<manu> if(person[foaf.nick] instanceof Array) {do x} else if(person[foaf.nick] instanceof String) {do y}

<Knud> can we have a new class that behaves nicely in both ways in JavaScript?

manu: Fundemental limitation of language to do this both ways in JS

<manu> person[foaf.nick] == <single value>

<Benjamin> we provide addional methods:

<Benjamin> rdfaObject.getValue(rdfa.rdfs.comment): bar

<Benjamin> rdfaObject.getValues(rdfa.rdfs.comment): foo,bar

<manu> personb[foaf.nick] = persona[foaf.nick];

<manu> personb[foaf.nick] === "joe";

<manu> persona[foaf.nick] === ["joe", "joeb", "joey"];

manu: For example, above assignment would only copy first item in array

<tinkster> if (person.has(rdfa.foaf.nick, "Joe")) { window.alert("It's Joe!"); }

<manu> var nick = personb[foaf.nick].getSingleValue();

manu: is there general agreement that we should provide native language objects?
... any opposition?

<manu> persona.getProperty(foaf.nick)

<tinkster> I think it's definitely a good idea to have these; but not sure if they need to be part of the base DOM API - this could be a library layered on top of the API.

<ivan> tinkster++++

<ivan> keep it simple

ivan: Agrees with tinkster. Keep in Simple. Put the basic stuff out there and add layers as time permits

<manu> global rdfa

manu: two different things -- global rdfa object and then RDFaObject

<manu> RDFaObject is returned from rdfa.list()

<Benjamin> rdfa.extractObject()

<manu> RDFaObject is returned from rdfa.getObject();

<Zakim> manu, you wanted to discuss that this is Projection-based mechanism.

ivan: First step should be triple interface, then global rdfa object. These should be seperate from any additional things we add on top of this API. Concerned with over-engineering

<Benjamin> we provide 8 methods :)

<ShaneM> I think its fine... +1

manu: any objection to putting native language mechanism out in FPWD?

<manu> PROPOSAL: Put Native language object support into DOM API for FPWD.

<tinkster> +1

<ShaneM> +1

<Steven> Fine by me

+

<Benjamin> +1

<ivan> 0

<manu> +1

+1

<manu> RESOLVED: Put Native language object support into DOM API for FPWD.

<Knud> +1

Triple-DOM Element Association

<Knud> :)

manu: associating DOM elements with generated triples, so you know where each part of a triple comes from.
... Any feedback on tying elements to each part of triple (subject/predicate/object)

ivan: would like to hear a use case

manu: E.g., highlight all subjects on page, or all objects on page

<tinkster> No, it's a good idea. In my parsing API, I can feed the calling code a DOM element for each triple, but it's just the "[current element]" in RDFa processing sequence terminology, so not necessarily useful.

<tinkster> So if people can get to the other relevant nodes, that's a good thing.

ShaneM: e.g., trace back to subject or object node from the predicate

manu: more complicated in some programming platforms, with significant implementation costs, e.g., with SAX parser

ivan: Agrees. But might be a price to pay

<tinkster> I don't think it involves *that* much extra work. It should just be a matter of passing around three extra variables in the parsing context.

<manu> Blank Nodes, URIs, TypedLiterals, PlainLiterals would all have .element

<tinkster> getChildElementByID doesn't use "node"

<tinkster> getElementByID rather.

manu: Any closing thoughts? Next week likely resolution to create FPWD on API

<Steven> http://www.w3.org/News/2010.html#entry-8777

<Knud> I'l give it a shot

<manu> ACTION: Shane to review RDFa DOM API FPWD [recorded in http://www.w3.org/2010/04/22-rdfa-minutes.html#action03]

<trackbot> Created ACTION-27 - Review RDFa DOM API FPWD [on Shane McCarron - due 2010-04-29].

<tinkster> k thx bye

<Knud> :)

<manu> ACTION: Knud to review RDFa DOM API FPWD [recorded in http://www.w3.org/2010/04/22-rdfa-minutes.html#action04]

<trackbot> Created ACTION-28 - Review RDFa DOM API FPWD [on Knud Möller - due 2010-04-29].

<Knud> who is going to be at WWW or WebScience? Should we have a little informal f2f?

This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/project/projection/
Succeeded: s/typing/tying/
Succeeded: s/e..g/e.g./
Found ScribeNick: RobW
Found Scribe: Robert
Present: Ivan Knud Manu RobW Shane Steven Toby
Regrets: MarkB JeffS BenA MarkusG
Agenda: http://lists.w3.org/Archives/Public/public-rdfa-wg/2010Apr/0094.html
Found Date: 22 Apr 2010
Guessing minutes URL: http://www.w3.org/2010/04/22-rdfa-minutes.html
People with action items: knud manu shane