Re: More about Facebook Linked Data

Really cool Jesse!

Now imagine linking this data to the ISWC 2011 metadata for the Linked
Data-a-thon :)
http://iswc2011.semanticweb.org/calls/linked-data-a-thon/

Don't forget, deadline is Oct 15.. and anybody can participate!!!

Juan Sequeda
+1-575-SEQ-UEDA
www.juansequeda.com


On Tue, Oct 11, 2011 at 4:35 PM, Jesse Weaver <weavej3@rpi.edu> wrote:

> APOLOGIES FOR CROSS-POSTING
>
> In this email, I would like to reveal a bit more about the Linked Data
> provided through Facebook's Graph API.
>
>
> AUTHENTICATION
>
> Before I get started, I have been asked about how to get an access token.
>  Information regarding authentication can be found at
> https://developers.facebook.**com/docs/authentication/<https://developers.facebook.com/docs/authentication/>.  If you just want a quick way to obtain a security token, just go to
> https://developers.facebook.**com/docs/reference/api/<https://developers.facebook.com/docs/reference/api/>and click on one of the example "connections" URLs and it will automatically
> generate a temporary access token (in the URL) for you (if you're logged in
> to your Facebook account).
>
>
> MORE THAN JUST USERS
>
> The main thing I want to point out is that the Graph API is much more than
> just basic information about users.  See https://developers.facebook.**
> com/docs/reference/api/<https://developers.facebook.com/docs/reference/api/>for details.  The Facebook graph includes things like pages, checkins, and
> links; it also includes "connections" (edges between things in the graph).
>  Anything you can get from the Graph API can be obtained in Turtle via HTTP
> content negotiation.  Simply include text/turtle in the Accept header (with
> a higher q preference than more typical mime types if necessary).  One of
> the most interesting features of the Graph API is the search mechanism:
>
> https://graph.facebook.com/**search?until=yesterday&q=**
> semantic+web&access_token=<https://graph.facebook.com/search?until=yesterday&q=semantic+web&access_token=>
> <**ACCESS_TOKEN>
>
> This will include information about a variety of things such as statuses,
> links, photos, and videos.  (At this point, I feel like I should point out
> the obvious, that is, that the information you see is only whatever is
> accessible given the access token provided.)
>
> Another cool feature is the ability to ask for information about any number
> of IDs, where an ID could even be the URL of something included in the Open
> Graph.  For example:
>
> http://graph.facebook.com/?**ids=http://www.cs.rpi.edu/~**
> weavej3/index.xhtml,**jesserweaver<http://graph.facebook.com/?ids=http://www.cs.rpi.edu/~weavej3/index.xhtml,jesserweaver>
>
> This includes information about my RPI website (which is in the Open Graph)
> and myself (jesserweaver).
>
>
> LINKED DATA URIs
>
> All of the HTTP(S) URIs in the RDF are dereferenceable.  Most of the URIs
> are hash-URIs (that is, strip off the fragment and fetch the document,
> returning 200 OK), but there are some slash-URIs (for example,
> http://graph.facebook.com/**schema/~/data<http://graph.facebook.com/schema/~/data>) which 303 redirect.  There are also information resources.  It's quite
> varied.
>
>
> SCHEMA
>
> This leads me to discuss the schema (or ontology, or whatever you want to
> call it).  Even the URIs representing properties and classes are
> dereferenceable.  For example, http://graph.facebook.com/**
> schema/link#type <http://graph.facebook.com/schema/link#type> is the URI
> identifying the link class, and http://graph.facebook.com/**schema/link<http://graph.facebook.com/schema/link>will return information (with 200 OK) about the link class.  Information
> includes relevant properties, possibly specifying whether a relevant
> property is a owl:DatatypeProperty, owl:ObjectProperty, and/or
> owl:InverseFunctionalProperty.  There may also be rdfs:domain and/or
> rdfs:range specified where appropriate.
>
> There are also properties that are not associated with a specific class,
> and these (generally) use the http://graph.facebook.com/**schema/~/<http://graph.facebook.com/schema/~/>namespace URI.  Properties using this namespace have very vague semantics.
>  For example, the property identified by URI http://graph.facebook.com/**
> schema/~/data <http://graph.facebook.com/schema/~/data> (which 303
> redirects to http://graph.facebook.com/**schema?tag=data<http://graph.facebook.com/schema?tag=data>) has the following description:
>
> "A tag having no semantics beyond the conventional semantics of the JSON
> key \"data\" as used in the Facebook Graph API."
>
> So if never_used_as_a_key is never used as a key in the Graph API JSON,
> then the property http://graph.facebook.com/**schema/~/never_used_as_a_key<http://graph.facebook.com/schema/~/never_used_as_a_key>is essentially meaningless.  There are two special cases, however.  One is
> http://graph.facebook.com/**schema/~/id<http://graph.facebook.com/schema/~/id>which is a owl:InverseFunctionalProperty with rdfs:range xsd:string.  The
> other exception is any URI beginning with http://graph.facebook.com/**
> schema/~/ <http://graph.facebook.com/schema/~/> followed by an optional
> underscore and then a non-negative integer.  These properties are an attempt
> at more practical container membership properties.  As an example, check out
>
> http://graph.facebook.com/**schema/~/_14<http://graph.facebook.com/schema/~/_14>
>
> http://graph.facebook.com/**schema/~/_14<http://graph.facebook.com/schema/~/_14>is a rdfs:subPropertyOf tag:
> graph.facebook.com,2011:/**has (the invented, generic, container
> membership property) and has an explicitly defined integer index as the
> value of the tag:graph.facebook.com,2011:/**index property (something
> missing from the RDF container membership properties).
>
> @prefix rdf: <http://www.w3.org/1999/02/22-**rdf-syntax-ns#<http://www.w3.org/1999/02/22-rdf-syntax-ns#>>
> .
> @prefix rdfs: <http://www.w3.org/2000/01/**rdf-schema#<http://www.w3.org/2000/01/rdf-schema#>>
> .
> @prefix owl: <http://www.w3.org/2002/07/**owl#<http://www.w3.org/2002/07/owl#>>
> .
> @prefix xsd: <http://www.w3.org/2001/**XMLSchema#<http://www.w3.org/2001/XMLSchema#>>
> .
> @prefix api: <tag:graph.facebook.com,2011:/**> .
> @prefix og: <http://ogp.me/ns#> .
> @prefix fb: <http://ogp.me/ns/fb#> .
> @prefix : <http://graph.facebook.com/**schema/~/<http://graph.facebook.com/schema/~/>>
> .
> :_14 a rdf:Property ;
>        rdfs:label "_14" ;
>        rdfs:comment "A tag having no semantics beyond the conventional
> semantics of the JSON key \"14\" as used in the Facebook Graph API." ;
>        rdfs:subPropertyOf api:has ;
>        api:index 14 .
>
>
> OPEN GRAPH
>
> URIs used as properties in the Open Graph (except for those in the
> http://ogp.me/ns# and http://ogp.me/ns/fb# namespaces; those URIs go
> elsewhere) also redirect to the Graph API schema feature.  For example,
> http://ogp.me/ns/book#author is a property, and http://ogp.me/ns/book302-redirects to
> http://graph.facebook.com/**schema/og/book<http://graph.facebook.com/schema/og/book>(302 instead of 303 because they are hash URIs).  For information about Open
> Graph, see http://ogp.me and https://developers.facebook.**
> com/docs/beta/opengraph/<https://developers.facebook.com/docs/beta/opengraph/>.  Please understand that the Graph API and the Open Graph are NOT the same
> thing.
>
>
> STATUS OF FB LINKED DATA
>
> These are just some of the features of the Facebook Graph API Linked Data.
>  I have been told that Facebook Linked Data is considered experimental and
> that continued support depends on the degree of use by developers.
>
>
> I hope this was helpful, and I am happy to attempt to answer any questions.
>  Please understand, though, that I am not a Facebook employee; I only
> recently finished an internship there.
>
>
> Regards,
>
> Jesse Weaver
> Ph.D. Student, Patroon Fellow
> Tetherless World Constellation
> Rensselaer Polytechnic Institute
> http://www.cs.rpi.edu/~**weavej3/index.xhtml<http://www.cs.rpi.edu/~weavej3/index.xhtml>
>
>
>