W3C home > Mailing lists > Public > www-rdf-interest@w3.org > August 2004

Re: rdfs:Statement for Qualified Predictates

From: Matthew Leingang <leingang@math.harvard.edu>
Date: Wed, 04 Aug 2004 19:15:23 -0400
To: Graham Klyne <GK@ninebynine.org>, <www-rdf-interest@w3.org>
Message-ID: <BD36E64B.12E44%leingang@math.harvard.edu>

Thanks for the reference.  It seems the recommendation there would be to
create (for instance) relation classes like :Purchase and :Gift to be
subclasses of :Acquisition, perhaps, and everything a subclass of "a
relation between a person and some stuff," kind of like I had been modeling
it originally.  But "person" is unnecessarily restrictive and "stuff," too,
so I still wonder what the difference is between that an an rdfs:Statement,
where rather than subclass for every relation, you use a property, namely,
the predicate.

Is the problem that some agents might incorrectly induct from a
qualified rdfs:Statements to an unqualified one?

Luckily, I'm just playing here, but I'm still interested in doing it
"right".

--Matt

On 8/4/04 6:36 PM, "Graham Klyne" <GK@ninebynine.org> wrote:

> A recent draft [2] put out by the Semantic Web Best Practices and
> Deployment Working Group [1] might have some useful ideas for you:
> 
> #g
> --
> 
> [1] http://www.w3.org/2001/sw/BestPractices/
> 
> [2] http://www.w3.org/TR/swbp-n-aryRelations/
> 
> 
> 
> At 16:42 04/08/04 -0400, Matthew Leingang wrote:
> 
>> Hi all,
>> 
>> Long time lurker, first time poster.
>> 
>> I'm thinking about generating RDF from my Books [1] database to publish
>> online (why not? :-). Books allows you store lots of data, not only about
>> the book, but when you got it, who you loaned it to, etc.
>> 
>> How does one best put something like "I have owned this book since August 1,
>> 2004" in RDF?  Without the date I could just say
>> 
>> :me foaf:owns [a book:Book; book:isbn "012345678"] .
>> 
>> (assuming we have a book schema someplace with a Book class and isbn
>> property).  With the date we have a qualification of the foaf:owns property.
>> When implementing this first in python, I started creating
>> "AgentObjectAssociation" classes for want of a better term.  Then about a
>> day later I realized that these AgentObjectAssociations were just subclasses
>> of regular RDF triples.  It seems, then, that you can do this by reifying
>> the statement and adding whatever other properties & types you want:
>> 
>> [a rdfs:Statement;
>>    rdfs:subject :me;
>>    rdfs:predicate foaf:owns;
>>    rdfs:object [a book:Book; book:isbn "012345678"];
>>    a ical:Vevent;
>>    ical:dtstart "2004-08-01".]
>> 
>> I don't know if this is a misuse of ical to talk about *any* event, not just
>> one that one might schedule in a PIM tool.  But another schema for events
>> could be used if necessary.
>> 
>> I recall from the one of the RDF TR's that the set of statements
>> 
>> [a rdfs:Statement; rdfs:subject :a; rdfs:predicate :b; rdfs:object :c]
>> 
>> doesn't necessarily have to imply :a :b :c (which is a good thing:
>> "declaring 'Mission Accomplished' doesn't make it so"!) and that made me shy
>> away from using qualified statements like this.  But it seems useful.
>> Another use would be to give the authority of a statement, so an agent can
>> decide which one to accept:
>> 
>> [a rdfs:Statement;
>>    rdfs:subject    :LeeHarveyOswald;
>>    rdfs:predicate  :killed;
>>    rdfs:object     :JohnFKennedy;
>>    dc:creator      :WarrenCommission
>> ].
>> 
>> [a rdfs:Statement;
>>    rdfs:subject    :Mafia
>>    rdfs:predicate  :killed;
>>    rdfs:object     :JohnFKennedy;
>>    dc:creator      :OliverStone
>> ].
>> 
>> Sorry if my N3 is off, but hopefully you get my drift.  So I guess I'm
>> asking is: is this "in or out of line" (to borrow the question of a previous
>> poster)?  Are there other ways to employ adverbial prepositional phrases
>> like "since" or "according to" in RDF?
>> 
>> Thanks for any help.
>> 
>> --Matthew Leingang
>> 
>> [1] http://books.aetherial.net/
>> 
>> --
>> Matthew Leingang
>> Preceptor in Mathematics
>> Harvard University
>> 
>> URL: http://www.math.harvard.edu/~leingang/
>> vCard: http://www.math.harvard.edu/~leingang/vCard.vcf
> 
> ------------
> Graham Klyne
> For email:
> http://www.ninebynine.org/#Contact
> 

-- 
Matthew Leingang
Preceptor in Mathematics
Harvard University

URL: http://www.math.harvard.edu/~leingang/
vCard: http://www.math.harvard.edu/~leingang/vCard.vcf
Received on Wednesday, 4 August 2004 19:17:06 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:14:57 UTC