W3C home > Mailing lists > Public > www-rdf-logic@w3.org > January 2005

Re: same-syntax extensions to RDF

From: R.V.Guha <guha@guha.com>
Date: Wed, 05 Jan 2005 08:52:51 -0800
Message-ID: <41DC1B63.9090503@guha.com>
To: Sandro Hawke <sandro@w3.org>
CC: www-rdf-logic@w3.org

Given all the response, I'd like to clarify what I meant.

The simple triple model is a good, uniform data model for 
expressing lots of simple machine readable data (such as 
specifying the capital of countries,  the author of a book as 
well as structures such as taxonomy trees and bookmarks). RDFS 
adopts a promiscous notion of an ontology (ontology as in the 
Quine sense of the word, i.e., the objects in the domain of 
discourse) as advocated first by McCarthy and later used by 
Cyc and other systems. Indeed, if you look at most of the RDF 
on the web today (ODP, RSS, Mozilla, MusicBrainz), you will
find that it is used in this spirit.

The problems with RDF arise when

a) it is "combined" with logical formalisms such as 
description logics which have a very different philosophy
b) when the RDF syntax and model are used to do syntax.

I am not sure there is a clean solution for (a). I don't think 
OWL should be made to build on top of RDFS. They are just too 
different. Those who like description logics would probably 
have preferred to build it directly on top of XML. And those 
who have no interest in it would have added the vocabulary in 
OWL-Lite to RDFS.

As to (b), we should just avoid it. RDF's syntax is baroque at 
best. Which is why even W3C folks use N3, etc.


Sandro Hawke wrote:

>>This infatuation with reducing everything to RDF is at best 
> Really?
> Don't you find it useful that so much can be stated in RDF?  I find
> RDF data very nice to work with. 
>>[1] http://www.w3.org/TR/NOTE-MCF-XML/
>     -- sandro
Received on Wednesday, 5 January 2005 16:55:34 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:43 UTC