W3C home > Mailing lists > Public > semantic-web@w3.org > April 2005

Re: RDF, RDFS, and OWL language constructs

From: Frank Manola <fmanola@acm.org>
Date: Mon, 18 Apr 2005 10:40:32 -0400
Message-ID: <4263C6E0.3090007@acm.org>
To: Charles McCathieNevile <charles@sidar.org>
CC: Hans Teijgeler <hans.teijgeler@quicknet.nl>, semantic-web@w3.org

Hi guys--

Section 1.1 of the OWL Guide, and Section 8 and Appendix E of the OWL 
Reference, are fairly uncomplicated introductions to the relationship 
between RDF and OWL, and the special characteristics of the OWL 
dialects.  The main issues are the guarantees of decidability and 
computational completeness that you have with OWL DL (and OWL Lite, 
which is a subset of OWL DL), versus the expressibility that you have 
with OWL Full.  Part of the choice depends on the efficiency and 
completeness of the reasoning that you require, and part depends on what 
you need to be able to express in your ontology (your ontologies may 
vary!).

--Frank


Charles McCathieNevile wrote:
> 
> On Sun, 17 Apr 2005 18:47:31 +0200, Hans Teijgeler  
> <hans.teijgeler@quicknet.nl> wrote:
> 
>> Hi!
>> For most of you back to the ABC of OWL, but new to me.
>> In the OWL Web Ontology Language Overview under "1.2 Why OWL?" I read:
> 
> ...
> 
>> The impression is given that OWL is the cumulation of RDF, RDFS, and
>> OWL-specific "language constructs".
>> Then I read "2.1 OWL Lite Synopsis" and "2.2 OWL DL and Full 
>> Synopsis",  and
>> I understand that the language constructs listed there are all there 
>> is  in OWL, so with the exclusion of all-but-one of the RDF constructs 
>> and  seven of the 15 RDFS constructs.
>> Does this mean that these excluded (or non-listed) RDF and RDFS  
>> constructs
>> may not be used in an OWL-compliant document, or that it is commonly not
>> used but valid? What is the rationale?
> 
> 
> Hi Hans,
> 
> There are some RDF and RDFS things that are not in OWL Lite or OWL DL  
> (which are just defined as subsets of OWL which don't include those 
> things  :-)
> 
> Full OWL is OWL that doesn't have any restriction. As I understand it 
> the  restriction is because if you have something that conforms to OWL 
> DL there  is more predictability about the logical implications that you 
> can encode,  and in OWL lite there is even more predictability. I 
> believe that the big  rationale is that you can be sure a tool that 
> processes OWL lite won't get  caught in some logic traps like circular 
> definitions and undecidable  propositions, or something like that, 
> whereas OWL full is capable of  encoding these kinds of statements, so 
> processors need to be able to  handle them.
> 
> This is a pretty rough understanding - someone with more time and  
> expertise might be able to give you a more detailed explanation if you  
> want one.
> 
> cheers
> 
> Chaals
> 
Received on Monday, 18 April 2005 14:30:51 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:44:53 UTC