W3C home > Mailing lists > Public > public-vocabs@w3.org > October 2011

Re: Cardinality of schema.org properties (ISSUE-5)

From: Martin Hepp <martin.hepp@ebusiness-unibw.org>
Date: Tue, 25 Oct 2011 18:09:28 +0200
Cc: Dan Brickley <danbri@danbri.org>, Philip Jägenstedt <philipj@opera.com>, "jpanzer@google.com" <jpanzer@google.com>, "public-vocabs@w3.org" <public-vocabs@w3.org>
Message-Id: <1A97B49B-A12F-4993-9A08-84B4E4F5E563@ebusiness-unibw.org>
To: Gregg Kellogg <gregg@kellogg-assoc.com>
Hi Gregg:

In most scenarios, I would say that the modeling of cardinalities via OWL constructs is no viable option:

1. The exact semantics does typically not match what people with a database background would expect, see
OWL DL vs. OWL flight: conceptual modeling and reasoning for the semantic Web
PDF from http://dl.acm.org/citation.cfm?id=1060745.1060836

2. You will need a pretty comprehensive OWL DL reasoner just to find out that there is a contradicting class membership resulting from
a cardinality violation.

3. It is not possible to state that e.f. the same property can be applied only once for the same natural language 
(e.g. one description per language per product only).

So while this is theoretically possible, I think it is no real option.

One could model such constraints in SPIN, but I suggest to add them to the label of the element; you can then easily write a few SPARQL rules that fetch the upper and lower bounds from the label and list entities that violate them.

Best

Martin


On Oct 21, 2011, at 7:30 PM, Gregg Kellogg wrote:

> On Oct 21, 2011, at 4:56 AM, Dan Brickley wrote:
> 
>> On 20 October 2011 10:19, Philip Jägenstedt <philipj@opera.com> wrote:
>>> On Thu, 20 Oct 2011 01:22:02 +0200, John Panzer <jpanzer@google.com> wrote:
>>> 
>>>> I'm trying to determine how to know what the intended cardinality of any
>>>> given schema.org defined property.
>>> 
>>> I asked about this and got this answer: [1]
>>> 
>>> On Thu, Jun 16, 2011 at 06:56, Guha <guha@google.com> wrote:
>>> 
>>>> Right now, it is always allowed to have multiple values.
>>>> In the future, we could/should introduce a property of properties that
>>>> specifies when a property may have only a single value.
>>> 
>>> Not exactly a satisfactory situation, but there it is.
> 
> Given that schema.org references an OWL vocabulary description [2], it would seem logical to use something like the following:
> 
> schema:CreativeWork a owl:Class;
>   rdfs:label "CreativeWork"@en;
>   rdfs:comment "The most generic kind of creative work, including books, movies, photographs, software programs, etc."@en;
>   rdfs:subClassOf
>     schema:Thing,
>     [ a owl:Restriction; owl:maxCardinality 1; owl:onProperty schema:awards ].
> 
>>> [1] http://groups.google.com/group/schemaorg-discussion/browse_thread/thread/95dfc4aa1ab89dbf?pli=1
> [2] http://schema.org/docs/schemaorg.owl
>> 
>> I've recorded this in an issue at
>> http://www.w3.org/2011/webschema/track/issues/5
>> 
>> Dan
> 
> Gregg
> 
> 
Received on Tuesday, 25 October 2011 16:09:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:48:56 GMT