Re: modeling document classes with OWL

Dan Brickley wrote:

>* Jan Algermissen <jalgermissen@topicmapping.com> [2005-04-03 21:14+0200]
>  
>
>>Hi,
>>
>>I want to use OWL class descriptions to define classes of documents but 
>>I can't find out, how to describe (for example) that an 'Article' has a 
>>sequence of sections.
>>
>>Here is an example:
>>
>>First I describe that 'Articles have exactly one dc:title and one 
>>dc:creator'
>>
>><owl:Class rdf:ID="Article">
>> <owl:equivalentClass>
>>   <owl:Class>
>>     <owl:intersectionOf rdf:parseType="Collection">
>>       <owl:Restriction> 
>>         <owl:onProperty rdf:resource="&dc;title" />
>>         <owl:cardinality
>>           rdf:datatype="&xsd;nonNegativeInteger">1</owl:cardinality>
>>       </owl:Restriction>
>>       <owl:Restriction>
>>         <owl:onProperty rdf:resource="&dc;creator" />
>>         <owl:cardinality  
>>           rdf:datatype="&xsd;nonNegativeInteger">1</owl:cardinality>
>>       </owl:Restriction>
>>     </owl:intersectionOf>  
>>   </owl:Class>
>> </owl:equivalentClass>
>></owl:Class>
>>
>>(That is the (a) way to do that I hope).
>>
>>
>>I'd add other required or optional properties (such as teaser, copyright 
>>information, etc.) in the same way. 
>>I (think) I can say that an Article has 0...N sections with 
>>owl:minCardinality and owl:maxCardinality, but I want them to be ordered 
>>(of course). 
>>An instance of the Article class I have in mind would need to 'have' an 
>>rdf:Seq containing the sections, I think (correct?). Could anyone help me 
>>with the question of how I say that in OWL?
>>    
>>
> 
>A brief comment. I think the core issue here is two subtle senses of
>'have'. The thing in the world, ie. the article itself, might have 
>certain things true of it. And then some specific RDF instance data
>might carry RDF statements describing some or all of those things. OWL 
>talks in terms of constraints regarding things in the world, rather than 
>constraints on the content of RDF documents about those things. 
>
Yes, I see what you mean (I think ;-) I have thought about that, too. 
Maybe I can clarify my position by saying that my sections are 
'first-class objects'; that they have identity. Therefore I understand 
the relationship between the article and a section in the same way as a 
relationship between e.g. an employee and a department. IOW, I use the 
same understanding of 'data modeling' for concepts and their 
relationships as for documents (also just a concept) and their content 
elements (also just concepts).


Jan

>XML by 
>contrast is all about types of document (and in RDF/XML, about types 
>of description). So I think you need to know whether your rules are 
>to be rules about the properties of 'Articles' in the abstract, or about
>the characteristics of specific RDF/XML descriptions of them. 
>
>Hmm that's maybe a bit confusing. I'll have a think about how to
>rephrase...
>
>hope it might help anyway,
>
>Dan
>
>
>  
>
>>Note:
>>
>>- I know that XML would be my friend here, but I do want to use RDF
>>- I also do not want to use XML semantics (like the infoset RDF namespace) 
>>to express the desired content model
>>
>>Many! thanks in advance for any clarification of this issue.
>>
>>
>>Jan
>>
>>
>>
>>-- 
>>Jan Algermissen
>>Consultant & Programmer
>>http://jalgermissen.com
>>
>>
>>    
>>
>
>  
>


-- 
Jan Algermissen
Consultant & Programmer
http://jalgermissen.com

Received on Sunday, 3 April 2005 19:47:42 UTC