Re: modelling issue?

I hesitate to get involved, but here goes...

On Sep 27, 2009, at 6:07 AM, Paola Di Maio wrote:

>
> During Vocamp Glasgow, I tried to confront my difficulties in  
> identifying some domain range of few vocabularies that I started  
> rdfizing as practice, and from explosing my questions to a whole  
> range of RDF doctors (thanks Norm, Keith, Serge) two things emerged,  
> that i did not know before
>
> 1) an entity (class, object, subject) does not necessarily have  
> domain /range
>
> Is that so, and what's the rule/ and possibly exceptions/ that can  
> be inferred and applied?

In RDF, domains and ranges belong to properties (the things in the  
middle of the triples). The idea is that if the domain of P is D and  
the range of P is R, then when you write a triple

a P b .

then you are implying that a is in D and b is in R. Obviously D and R  
have to be classes. For example, the domain and range of motherOf  
might be respectively Woman and Human, so that if I write

Betty motherOf Pat .

you can infer that Pat is human and Betty is a woman.

That is the full story on domains and ranges in RDFS (and in OWL, for  
that matter). So I'm not sure what you mean by the domain/range of an  
*entity*.

BTW, there is no RDF requirement that domains and ranges be specified.  
You can just say nothing about them unless you want to. And you can  
give something two (or more) domains or ranges, and then the above  
constraint applies to both of them, eg a has to be in D1 and in D2 if  
they are both domains of P.

>
> that did not emerge at Vocamp
>
>
> 2) Apparently a triple can be of two kinds

Well, no. There is only one kind of triple in RDF. However, the third  
item in it can be a URI or a literal, indeed. This only really  
seriously matters in OWL, which gets very anal about the literal/non- 
literal distinction.

> :
>
> class:relation:class
>
>  but also
> class:attribute:value
>
> Of this i would like some confirmtion (is this right?

No. Or at any rate, not if I am following you. First, the first item  
(subject) of the triple isn't necessarily a class. Second, the terms  
'relation' and 'attribute' are not used in RDF, though they are both  
used more widely to mean what RDF calls a property.

> ),
> Finally,  finally, wouldnt' this ambiguity be confusing?

There is no ambiguity. Relation = attribute (= property), so those are  
the same; and the only RDF-meaningful distinction for the third item  
is between a thing denoted by a URI, or a thing denoted by a literal.  
The latter is often called a value, but a value can be denoted by a  
URI as well, so thats not a distinction in kind so much as a  
distinction between two ways to refer to something.

>
> i dont have a case study for this yet, but if this is true I suspect  
> it could cause some possible  logical conflict/ambiguity
> in semantic data model and its implementation

 From your message, I cannot see what distinction you are wanting to  
address. It is true that RDF is a very weak language and cannot  
express or represent all kinds of subtleties, but this does not make  
it ambiguous or conflicted. The RDF semantics is quite clear and  
unambiguous. Can you give some examples of the kind of contrast you  
have in mind here?

Pat Hayes

> am I the only one thinking so?

>
>
> Are the above points addressed in some RDF tutorial
>
> please enlighten!
> thanks a lot
>
> PDM
>
>

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes

Received on Monday, 28 September 2009 17:11:41 UTC