W3C home > Mailing lists > Public > public-swbp-wg@w3.org > June 2004

Re: Database schema in OWL

From: Aldo Gangemi <a.gangemi@istc.cnr.it>
Date: Mon, 7 Jun 2004 17:34:11 +0200
Message-Id: <p0611041dbcea36aad8cf@[212.34.219.175]>
To: "Daniela F. Brauner" <dani@les.inf.puc-rio.br>, public-swbp-wg@w3.org
Cc: seanb@cs.man.ac.uk, horrocks@cs.man.ac.uk

Thanks for your question, Daniela. In the WonderWeb project 
http://wonderweb.semanticweb.org we have built a use case to show 
good translation practices from db (conceptual or logical) schemas to 
ontologies (also with help of automatic tools). We should put an 
online report asap about that. See below for an example.

At 12:04 -0300 7-06-2004, Daniela F. Brauner wrote:
>Hi all,
>
>Can someone help me with description of a simple database schema to OWL?
>What is the best way to describe it??
>
>Imagine that we have these simple database: (P.S.:assume # to represent PK
>and & to FK)
>
>Students (#id,name,&courseId)
>Course(#id,name,&deptId)
>Department(#id,name,&deptId)

The problem with these logical schemas is that some elements are not 
relevant from an ontological viewpoint, but they concern database 
functioning. Identifiers, specially as fk, are typically so.

A good practice for your schema would be (quickly quickly):

Class(Students partial
  restriction(hasCourse someValuesFrom Course))
Class(Course partial
  restriction(hasDept someValuesFrom Department))
Class(Department)
ObjectProperty(hasCourse
  domain(Students)
  range(Course))
ObjectProperty(hasDept
  domain(Course)
  range(Department))

In other words, identifers that are just there for indexing purpose 
are not really "ontological" (although you can obviously represent 
them in OWL as restrictions on datatype properties), while names can 
be used as class names. Moreover, whenever there is a clear link 
between a "name" and a fk, then you can introduce a property that has 
appropriate domain and range restrictions, and use it to create one 
or more someValuesFrom restriction(s).

There are other practices concerning the way properties and classes 
should introduced as subclasses or subproperties of existing, 
reference ontologies, but this goes beyond what you're asking for.

Good luck
Aldo

>Thanks a lot!
>
>
>---
>Outgoing mail is certified Virus Free.
>Checked by AVG anti-virus system (http://www.grisoft.com).
>Version: 6.0.699 / Virus Database: 456 - Release Date: 6/4/2004


-- 
Aldo Gangemi
Research Scientist
Laboratory for Applied Ontology
Institute for Cognitive Sciences and Technology
National Research Council (ISTC-CNR)
Via Nomentana 56, 00161, Roma, Italy
Tel: +390644161535
Fax: +3906824737
a.gangemi@istc.cnr.it
Received on Monday, 7 June 2004 11:34:15 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:09:39 UTC