Re: Questions/Concerns about the Data Cube Vocabulary

On 11 July 2013 00:32, Sarven Capadisli <sarven.capadisli@deri.org> wrote:
> On 07/10/2013 02:26 PM, Mark Wilkinson wrote:
>>
>> Dear Data Cube group,
>>
>>
>> The Data Cube Vocabulary came to my attention as part of my mentoring of
>> a Google Summer of Code student, who is using the vocab to model data
>> for e.g. Genome Wide Association Studies in biomedicine.  My personal
>> interest in the vocabulary is more from a Semantic Web perspective, and
>> therefore I am looking at it with an eye to using it with OWL reasoners.
>>
>> It is in this latter respect that I have some concerns about the
>> vocabulary, and I'm hoping you will engage me in this conversation, to
>> either confirm or correct my interpretations and concerns :-)
>>
>> I'll start with my primary concern, which is that some of your terms
>> seem to be used both as predicates and as classes.  For example:
>>
>> Class: qb:DimensionProperty Sub class of: qb:ComponentProperty,
>> qb:CodedProperty
>>
>>   eg:refArea a rdf:Property, qb:DimensionProperty;
>>   rdfs:label "reference area"@en;
>>   rdfs:subPropertyOf sdmx-dimension:refArea;
>>   rdfs:range admingeo:UnitaryAuthority;
>>   qb:concept sdmx-concept:refArea .
>>
>>
>> In OWL, properties and classes are disjoint, so... it looks like the set
>> of statements above violate this rule in that DimensionProperty is
>> declared as a Class (which itself is at least grammatically odd) and
>> then refArea is declared as both an rdf:Property *and* a
>> DimensionProperty(Class)...  This situation would make reasoners very
>> unhappy!  There are numerous examples of this in the data cube vocabulary.
>>
>> Are my concerns justified, or am I mis-interpreting the vocabulary?  I
>> realize that there is nothing *wrong* with the structure, I just have
>> concerns about using it as a sem-web vocabulary; it would be a shame if
>> we weren't able to do reasoning over data cubes!
>>
>> I'd be very grateful if someone from the working group would engage me
>> in this discussion :-)
>>
>> Best wishes!
>>
>> Mark
>
>
> I don't see a problem:
>
> qb:DimensionProperty rdfs:subPropertyOf+ rdf:Property .
> rdf:Property rdfs:subPropertyOf rdfs:Class .
>
> -Sarven

Dear Sarven,

rdf:Property is not the property top, it is the class of all properties [1].
So, in fact:
qb:DimensionProperty rdfs:subClassOf qb:ComponentProperty
rdfs:subClassOf rdf:Property .


Dear Mark,

qb:DimensionProperty is the class of properties which represent the
dimensions of the cube. It is a class, a subclass of rdf:Property. The
statement:

eg:refArea a rdf:Property, qb:DimensionProperty .

makes eg:refArea an *instance* of this class of properties, ie., a
property, not a class itself.


Dear GLD group,

although only minor, I find the statement [2]:
qb:DimensionProperty Sub class of: qb:ComponentProperty, qb:CodedProperty
gratuitous. Since:
qb:CodedProperty Sub class of: qb:ComponentProperty
it would suffice that:
qb:DimensionProperty Sub class of: qb:CodedProperty

Best,
s


[1] http://www.w3.org/TR/rdf-schema/#ch_property
[2] http://www.w3.org/TR/vocab-data-cube/#dfn-qb-dimensionproperty

Received on Thursday, 11 July 2013 03:47:26 UTC