W3C home > Mailing lists > Public > public-lod@w3.org > November 2012

Re: Breaking news: GoodRelations now fully integrated with schema.org!

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Mon, 12 Nov 2012 16:05:36 +0100
Message-ID: <CAK4ZFVHMWyp1FuR3vzG3YvK4d=22UpdSGxWSRTiyvLr1BFCSnw@mail.gmail.com>
To: Dan Brickley <danbri@danbri.org>
Cc: Guha <guha@google.com>, Martin Hepp <martin.hepp@ebusiness-unibw.org>, Semantic Web <semantic-web@w3.org>, "public-vocabs@w3.org Vocabs" <public-vocabs@w3.org>, public-lod <public-lod@w3.org>
Dan, Martin, all

This breaking news made me un-earth the couple of questions I already
discussed with you regarding the (more or less declared) "soft" semantics
of schema.org, and how both http://schema.org/docs/schemaorg.owl and
http://schema.rdfs.org are interpreting those semantics a bit harder than
it should, in particular regarding domains and ranges of properties.

I take now for granted from your message that :
- The reference file for schema.org declared semantics is
http://schema.org/docs/schema_org_rdfa.html (rather than the outdated, or
at lesat not clearly dated OWL file at http://schema.org/docs/schemaorg.owl)
- It declares explicitly schema.org types (classes) as instances of
rdfs:Class, and attached properties as instances of rdf:Property.
- It uses rdfs:subClassOf for the type hierarchy. There is no use of
rdfs:subPropertyOf
- It uses specific properties  http://schema.org/domain and
http://schema.org/range
to attach properties to classes.

The latter is the most interesting and innovative feature. It should be
good to document in th file the implied semantics of those properties, of
which semantics is weaker than the ones of rdfs:domain and rdfs:range, as
implicitly (explicitly?) stated in http://schema.org/docs/datamodel.html.
And maybe it would be wise to rename them otherwise, since confusion is
likely to occur (the more so that http://schema.rdfs.org has interpreted
them abusively as rdfs:domain and rdfs:range). Why not call them the same
as in the html pages : "expectedOnType" and "expectedValueType", since it's
really what they mean.

Side question to Martin. Is there any issue in formally mapping the OWL
classes and properties of GoodRelations to their schema.org equivalents,
which do not even rely on RDFS semantics? I'm pretty sure you have thought
about it and I would be happy to have your take on this.

Another point is since you now declare that the RDF expression of
schema.orgis the root of it, why not publish a proper RDF schema that
could be GET
from the http://schema.org/ namespace through content negotiation, as any
other vocabulary conformant to SW publishing best practices? BTW for
example we would be happy to have such a thing in order to integrate
seamlessly schema.org in LOV. So far we use the
http://schema.rdfs.orgsource but this is really suboptimal, we would
like to get rid of this, and
insert the real stuff.

I submitted the page to the W3C vRDFa validator at
http://www.w3.org/2012/pyRdfa/Validator.html it's happy with the file and
produces a very clean n3 file, the kind it would be cool to have in above
said content negotiation.

Best

Bernard




2012/11/9 Dan Brickley <danbri@danbri.org>

>
> This latest build of schema.org uses a different approach to previous
> updates. Earlier versions (apart from health/medicine) were relatively
> small, and could be hand coded. With Good Relations, the approach we
> took was to use an import system that reads schema definitions
> expressed in HTML+RDFa/RDFS and generates the site as an aggregation
> of these 'layers'. In other words, schema.org is built by a system
> that reads a collection of schema definitions expressed using W3C
> standards. The public site is also now more standards-friendly, aiming
> for 'Polyglot' HTML that works as HTML5 and XHTML, and you can find an
> RDFa view of the overall schema at
> http://schema.org/docs/schema_org_rdfa.html
>
>
> I'm really happy to see Good Relations go live, and look forward to
> catching up on the other contributions that are in the queue. The
> approach will be to express each of these in HTML/RDFa/RDFS and make
> some test sites on Appspot that show each proposal 'in place', and in
> combination with other proposals. Since schemas tend to overlap in
> coverage, this is really important for improving the quality and
> integration of schema.org as we grow. While it took us a little while
> to get this mechanism in place, I'm glad we now have this
> standards-based machinery in place that will help us scale up the
> collaboration around schema.org.
>
> Thanks again to all involved,
>
> Dan
>
>


-- 
*Bernard Vatant
*
Vocabularies & Data Engineering
Tel :  + 33 (0)9 71 48 84 59
 Skype : bernard.vatant
Blog : the wheel and the hub <http://blog.hubjects.com/>

--------------------------------------------------------
*Mondeca**          **                   *
3 cité Nollez 75018 Paris, France
www.mondeca.com
Follow us on Twitter : @mondecanews <http://twitter.com/#%21/mondecanews>
Received on Monday, 12 November 2012 15:06:29 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 31 March 2013 14:24:43 UTC