Compatibility with Semantic Web standards of W3C

[Anand]

Hi,
How is schema.org vocabulary is associated with standards like RDF/OWL
etc. proposed by W3C. It would be great if it will be made flexible
like triple store structure of RDF graphs. The linked data for example
contains millions of RDF triples. and thousands of datasets waiting to
be harnessed for its full potential.
It would be wonderful if the search leaders value the Semantic Web
stack and linked data efforts

thanks
Anand

[Dan Brickley]

Hi Anand,

The short answer is to say that there is already a close relationship,
and things are getting closer.

In http://schema.org/docs/datamodel.html from the start, it said that
Schema.org's data model is based on RDF Schema (a spec I've co-edited
for W3C with Guha since 1997). There were some areas where it didn't
completely use RDFS, due to a desire to have flexibility with domain
and range constraints. So now (with help from the nice folk behind
http://schema.rdfs.org/) there is also now an OWL version of
Schema.org (linked from the datamodel page),
http://schema.org/docs/schemaorg.owl ... so the standards story is OK
there, and builds on a long history.

On the syntax side, Schema.org initially focussed on consuming
Microdata syntax, which itself was based on W3C's RDFa. The original
announcement of the Microdata design (see
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019681.html
) began a long conversation in the standards community about possible
simplifications to RDFa, and many people have been working towards
convergence and proposals for an "RDFa 1.1 Lite". The recent
Schema.org workshop brought together a lot of this expertise (see
http://blog.schema.org/2011/09/schemaorg-workshop-wrap-up.html ) and
the general feeling of the meeting (as I remember it anyway!) was
towards pluralism, ie. supporting multiple syntaxes. The Semantic Web,
Linked Data and RDF perspective was well represented at the workshop;
and is in fact already well represented within the Schema.org
initiative.

Schema.org's vocabulary is specified in a syntax-neutral way. So far,
the main emphasis has been on the Microdata encoding. However the
vocabulary is simply a collection of terms (classes and properties)
that can be published and consumed in lots of ways. Already there has
been activity in the Linked Data and RDF community to express mappings
with other RDF vocabularies. For example, my friends in the Dublin
Core group have a draft mapping here -
http://wiki.dublincore.org/index.php/Schema.org_Alignment/Mappings -
and there is similar work around DBpedia -
http://sourceforge.net/mailarchive/message.php?msg_id=27769857 ...and
I expect also at Freebase.

Currently the Schema.org collaboration is around a single unified
vocabulary, whose design is sometimes informed explicitly by the
design of existing RDF vocabularies. So see for example the recent
announcement around rNews:
http://www.iptc.org/site/Home/Media_Releases/schema.org_adopts_IPTC's_rNews_for_news_markup

To continue these collaborations: around syntax, and around actual
schemas, there are two new groups being created at W3C. See
http://www.w3.org/QA/2011/09/proposing_two_new_sw_interest.html and
http://blog.schema.org/2011/10/w3c-web-schemas-group-is-our-new-public.html

In particular, the plan is to migrate discussion across from this
Google Group into W3C's new "public...@w3.org" "Web Schemas task
force" mailing list, which is the new home for these discussions.

By making the main feedback forum for Schema.org be a W3C group,
hopefully the intent to work more closely with the standards community
is clear...

To come back to your original question: Schema.org data is already
defined using RDF Schema, and can be consumed (if you choose) with RDF
tools. It's vocabulary is being mapped with, and informed by, other
RDF vocabularies/schemas. You can also (especially in RDFa) mix
Schema.org markup with markup that uses other RDF vocabularies. This
is important as it allows Schema.org to stay manageable, without
attempting to describe all things in full detail.

cheers,

Dan

[Stéphane Corlosquet]

Hi Anand,

On Thu, Oct 6, 2011 at 9:03 AM, Anand <mail.a...@gmail.com> wrote:

Hi,
How is schema.org vocabulary is associated with standards like RDF/OWL
etc. proposed by W3C.

The schema.org vocabulary is published in OWL at http://schema.org/docs/schemaorg.owl. In the past some people reported some discrepancies with the HTML version, but I'm sure that's something that can be fixed, if not fixed already. http://schema.rdfs.org/ also has several RDF oriented tools and documentations.

 

It would be great if it will be made flexible
like triple store structure of RDF graphs. The linked data for example
contains millions of RDF triples. and thousands of datasets waiting to
be harnessed for its full potential.
It would be wonderful if the search leaders value the Semantic Web
stack and linked data efforts

Could you be more precise about what you think is not compatible with RDF and Linked Data?

It's up to publishers to use the schema.org vocabulary when publishing their data as RDF (in whatever syntax they use). As it is, the schema itself is not tied to any syntax, we use it in Drupal for example to publish data in RDFa, index it in a triple store and run SPARQL queries against it.

Steph.

[Kingsley Idehen]

Anand,

I've written a post with actual examples. Basically, just make and
publish your HTML resources with Schema.org based structured data
islands and the Web of Linked Data will simply do its thing i.e., make
your contribution much more discoverable :-)

The Web is the database, and you contributions are just new items in
this global graph model database :-)

Links:

1. http://goo.gl/uuDyM -- basically shows the existing Linked Open Data
cloud and Schema.org are already working together

--

Regards,

Kingsley Idehen
President& CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen