Re: DAML+RDFS: potentials for simplifications?

Hi,

thank you a lot for your answer!

yes, the simplification I proposed looks not really amazing. however, I
think from an application developer's view it could mean a lot: it enables
direct parsing of structured knowledge instead of parsing huge amount of
triples and then figuring out the structure...

however my problem is that i do not know if i "need RDF".
When I hear RDF, 2 positive associations come to my mind:
- exhaustive use of the simple but effective URI concept, thus enabling a
truly interconnected web (of ontologies)
- some correspondence to First order predicate logic, which, in turn,
suggest capabilities for automatic reasoning and proofing and such.

The first of the positive assets I associate to RDF can easily be replaced
or emulated by an arbitrary XML attribute or element, which, by convention
contains an URI (e.g. about="..", resource="..", ID="..")

The second positive asset is much more difficult for me, because i have no
background in logics, and therefore i do not know if I risk losing power
when dropping RDF.


thanks,
joachim




                                                                                                                    
                    "Peter F.                                                                                       
                    Patel-Schneider"           An:     Joachim.Peer@unisg.ch                                        
                    <pfps@research.bell        Kopie:  www-rdf-logic@w3.org                                         
                    -labs.com>                 Thema:  Re: DAML+RDFS: potentials for simplifications?               
                    Gesendet von:                                                                                   
                    www-rdf-logic-reque                                                                             
                    st@w3.org                                                                                       
                                                                                                                    
                                                                                                                    
                    28.11.2001 17:21                                                                                
                                                                                                                    
                                                                                                                    




First of all, I think that you are on a reasonable track.  The basic
conclusions that you make are correct.  However, there are some detail
problems, and you may not have gone far enough.

In particular, your simplified syntax is little more than removing rdfs:
from some of the tags.  If you don't need to encode DAML+OIL in RDF
triples, then there are lots better syntaxes for languages like DAML+OIL.
For example, OIL (http://www.ontoknowledge.org/oil/) has an XML syntax
that has lots to recommend it.

DAML+OIL was designed to have an RDF syntax.  Its model theory is based on
RDF triples but is not really very compatible with the new RDF model
theory, largely because triples are inadequate for carrying both structure
and meaning.  I would prefer a logic that kept closer to the RDF model
theory, but that had a syntax that was an extension of the RDF XML syntax.
Some of my concerns are voiced in a recent message I sent out to the
DAML+OIL joint committee
(http://www.daml.org/listarchive/joint-committee/0934.html).


Peter F. Patel-Schneider
Bell Labs Research

Received on Wednesday, 28 November 2001 12:40:49 UTC