- From: Ian Horrocks <horrocks@cs.man.ac.uk>
- Date: Thu, 15 Aug 2002 21:48:49 +0100
- To: "R.V.Guha" <guha@guha.com>
- Cc: www-rdf-logic@w3.org
On August 15, R.V.Guha writes: > > <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> > <html> > <head> > <title></title> > </head> > <body> > <br> > Ian,<br> > <br> > I'd like a register a protest of your use of the term "standard logical > language". There have been a very large number of symbolic systems, both implemented > and gedanken, starting from Advice Taker to CycL. Description Logics form > a small subset of such systems. Many of these symbolic systems either lacked > the concept of subClassOf, etc. or included them in the domain of discourse.<br> > <br> > So, please, don't say "standard logical language". Say "standard Description > Logic based language". <br> Why pick on description logics? The standard logical languages to which I refer include modal logics, temporal logics, dynamic logics, standard predicate logics etc. etc. Ian > <br> > guha<br> > <br> > <br> > Ian Horrocks wrote:<br> > <blockquote type="cite" > cite="mid15707.42355.464023.150665@merlin.oaklands.net"> > <pre wrap="">On August 14, R.V.Guha writes: > </pre> > <blockquote type="cite"> > <pre wrap="">Ian, > > Could you explain how RDF can talk about its own syntax? Which syntax > are you referring to, the XML syntax or n-triples? > > Or are you referring to the RDF allowing statements to be made about > classes and predicates ... (that isn't really a syntax issue ...) > > I am very confused. > </pre> > </blockquote> > <pre wrap=""><!----> > RDF (or at least RDF schema, which I believe is now to be considered > part of RDF) gives special meanings to some classes and properties, > e.g., Class, subClassOf, type. In any standard logical language, these > would be considered to be part of the syntax of the language, and > would not be part of the domain of discourse of the language itself - > in predicate logic I can't say anything about the meaning of AND, OR > and NOT. In contrast, RDF allows me to extend the meaning of such > syntax, e.g., by asserting that subClassOf is a subPropertyOf type. > > In RDF itself this can be seen as relatively harmless in the sense > that it simply constrains the set of possible models, even if the > intended meaning of such statements is difficult to understand. When > RDF is extended to a more powerful logic (by giving additional meaning > to some sets of triples), e.g., one including negation, this feature > becomes much more problematical and makes it impossible to support > standard logical entailments (trying to do so makes the language > inherently paradoxical). > > Ian > > </pre> > <blockquote type="cite"> > <pre wrap="">guha > > > > Ian Horrocks wrote: > > </pre> > <blockquote type="cite"> > <pre wrap="">Unfortunately, I don't see how such studies can resolve the practical > problems associated with layering OWL on top of RDF, e.g., those > caused by RDF's ability to talk about its own syntax, and the > inheritance of this ability by wore expressive languages layered on > top of it. > > Regards, Ian > > > > > </pre> > <blockquote type="cite"> > <pre wrap="">thank you > > Guha > > > </pre> > </blockquote> > </blockquote> > </blockquote> > </blockquote> > <br> > </body> > </html>
Received on Thursday, 15 August 2002 16:51:22 UTC