This is the archive of the goodrelations dicussion list

Dear Michael, all:

In general, the semantics of additionalType and itemType is the same, and both are equivalent to rdf:type when consumed in an RDF environment. There are the following subtle differences:

1. In a Microdata environment, itemType sets the context for the properties and might thus influence how a local property name (like "description") will be translated into a fully qualified URI, unless the vocabulary assumes global property identifiers, see [1, 2].

additionalType does not influence the context for contextual generation of RDF properties.

This is currently not relevant for schema.org, since schema.org uses the "vocabulary URI generation scheme" [2] and  assumes global property identifiers, i.e. the "description" property maps to the same URI

    http://schema.org/description

and not to a type-dependent URI like

    http://schema.org/Thing&prop=description, if the itemType is http://schema.org/Thing, 

in contrast to

    http://schema.org/Product&prop=description, if the itemType is http://schema.org/Product.

However, if schema.org grows, its sponsors may decide to change the processing to context-based property URIs in order to avoid property name collisions (e.g. if you need a "parent" property for two very different itemTypes, like "person" vs. "folder" - it may become hard to find catchy names that are nice within the context of a certain type but do not conflict with property names defined for other types).

2. One cannot assume that all clients understand all additionalTypes, but a Microdata-compliant client supporting a particular Microdata vocabulary must support all of the types from that vocabulary. In general, a data publisher has to expect that some clients simply ignore the additional type.

3. In practice, the appropriate data cleansing and provenance-based handling of additional types will likely differ from the main itemType, because

a) additionalType markup may contain URIs of information resources (e.g. Wikipedia entries) instead of true type / class identifiers (like http://www.productontology.org class IDs), and

b) markup may contain simple category labels as text. While this is strictly speaking invalid markup, consuming clients should deal with it.

For case a), a consuming client may translate a URI from a namespace known to represent  information resources into a new URI (e.g. by appending "#class" as a suffix).

For case b), a consuming client may translate the category label into a proprietary URI, e.g. by properly URL escaping the string and attaching it to a site-specific base URI (e.g. translating "car audio" into http://example.com/#car%20audio.

Best wishes

Martin

[1] http://www.w3.org/TR/microdata-rdf/#property-uri-generation
[2] http://www.w3.org/TR/microdata-rdf/#generate-predicate-uri

On Jul 19, 2012, at 6:04 PM, Michael F Uschold wrote:

> Good going Martin.
> 
> What is the relationship between the semantics of "itemType", "additionalType" and "rdf:type"?
> Seems like itemType and additionalType might be subproperties of rdf:type?
> 
> Michael 
> 
> Feel free to respond to the list, if appropriate.
> 
> 
> On Wed, Jul 18, 2012 at 12:41 PM, Martin Hepp <martin.hepp at ebusiness-unibw.org> wrote:
> FYI
> 
> Clients consuming schema.org markup in RDF environments should translate the additionalType property to rdf:type, at least if the value of this property is a valid URI that identifies a class (e.g. if it comes from the http://www.productontology.org/id/* namespace).
> 
> Best
> Martin Hepp
> 
> Begin forwarded message:
> 
> > Dear all:
> >
> > I am happy to announce that the sponsors of schema.org, i.e. Google, Yahoo, Bing, and Yandex, have just implemented my proposal (from September 2011) to add a new property additionalType to http://schema.org/Thing and thus also to http://schema.org/Product.
> >
> > This allows using the more than 300,000 precise product type identifiers from our http://www.productontology.org service for telling search engines very, very precise the type of products or services you are selling.
> >
> > For example, a racing bike should now be marked up as follows:
> >
> > <div itemscope itemtype="http://schema.org/Product">
> >    <link itemprop="additionalType" href="http://www.productontology.org/id/Racing_bicycle" />
> >    <!-- other schema.org properties go in here -->
> > </div>
> >
> > You can use any reasonable Wikipedia entry for that.
> >
> > See
> >
> >    http://www.productontology.org/doc/Racing_bicycle#microdata
> >
> > for more details.
> >
> > This is also the first step in the ongoing effort to make the GoodRelations vocabulary fully available from the schema.org namespace.
> >
> > Best wishes
> >
> > Martin
> 
> --------------------------------------------------------
> martin hepp
> e-business & web science research group
> universitaet der bundeswehr muenchen
> 
> e-mail:  hepp at ebusiness-unibw.org
> phone:   +49-(0)89-6004-4217
> fax:     +49-(0)89-6004-4620
> www:     http://www.unibw.de/ebusiness/ (group)
>          http://www.heppnetz.de/ (personal)
> skype:   mfhepp
> twitter: mfhepp
> 
> Check out GoodRelations for E-Commerce on the Web of Linked Data!
> =================================================================
> * Project Main Page: http://purl.org/goodrelations/
> 
> 
> 
> 
> 
> 
> 
> -- 
> Michael Uschold, PhD
>    Senior Ontology Consultant, Semantic Arts
>    http://www.semanticarts.com
>    LinkedIn: http://tr.im/limfu
>    Skype, Twitter: UscholdM
> 
> 
> 
> 

--------------------------------------------------------
martin hepp
e-business & web science research group
universitaet der bundeswehr muenchen

e-mail:  hepp at ebusiness-unibw.org
phone:   +49-(0)89-6004-4217
fax:     +49-(0)89-6004-4620
www:     http://www.unibw.de/ebusiness/ (group)
         http://www.heppnetz.de/ (personal)
skype:   mfhepp 
twitter: mfhepp

Check out GoodRelations for E-Commerce on the Web of Linked Data!
=================================================================
* Project Main Page: http://purl.org/goodrelations/