W3C home > Mailing lists > Public > www-rdf-rules@w3.org > November 2003

Re: plain rules, please [was: Semantic Web Rule Language (SWRL) 0.5 released]

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Tue, 25 Nov 2003 10:17:10 -0800
Cc: Mike Dean <mdean@bbn.com>, www-rdf-rules@w3.org
To: Dan Connolly <connolly@w3.org>
Message-Id: <95EE6BCA-1F73-11D8-AFFE-0003939E0B44@isr.umd.edu>

On Tuesday, November 25, 2003, at 09:43  AM, Dan Connolly wrote:

> On Thu, 2003-11-20 at 02:21, Mike Dean wrote:
>> Please direct comments to www-rdf-rules@w3.org.
>> [2] http://www.daml.org/2003/11/swrl/
>
> "This syntax extends the abstract syntax of OWL".
>
> Why?

My impression from the DAML PI meeting was that this proposal intends 
to solve the knotty issue of *layering* rules on OWL.

> I was expecting plain old horn rules. I don't see
> why all the terms and structures from OWL-DL are
> carried into SWRL.
[snip]

OWL Rules extends OWL with rules. There needs to be *some* story on how 
OWL and rules work together. The other two current alternatives, 
AFAICT, are Description Logic Programming (DLP)[1] which identifies a 
fairly inexpressive and somewhat odd to many peoples' taste common 
subset of OWL and Logic Programming like constructs (sometimes called 
OWL Feather :)), and keeping them almost entirely disjoint (you want 
ontologies? use owl, you want rules? use RuleWhatever). Hmm. The main 
possible difference would be in the size of the overlapping subset. 
TRIPLE might be an example of the latter: As long as it can get triples 
out of the DLed model, it's happy (but you don't get rules affecting 
that model).

OWL Rules has some nice logical (it's just a larger subset of FOL) and 
pragmatic (it lets you reuse your hard won OWL expertise) features.

Of course, given the increasing recognition that Deduction is Evil(tm) 
or, at least, that Deduction is Silly and Pointless(r), the former 
properties might not be so nice.

Cheers,
Bijan Parsia.

[1]http://www2003.org/cdrom/papers/refereed/p117/p117-grosof.html
Received on Tuesday, 25 November 2003 13:15:19 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:15 UTC