RE: DAML Notations

> From: Sandro Hawke [mailto:sandro@w3.org]
> Has anyone done work on making a more pleasant ontology syntax for
> DAML+OIL [1] ?   
> 
> Obviously there are lots of personal preferences.  I'm imagining
> something vaguely Java-like:
> 
>    Class Animal extends LivingThing, has {
>       Animal mother;
>       Animal father;
>    }
> 
> with some clever syntax for cardinality and most of the extra
> expressiveness of DAML+OIL in some boolean expression constraint
> syntax.

I'd be concerned about a Java/C/C++/C#-ish syntax; not sure how you would
handle anonymous, arbitrarily-nested structures.  There are plenty of
LISPish syntaxes around in the DL community that are far more pleasant.  I'm
not a great fan of LISP, but its facilities for nesting are much more
appropriate for the kinds of structures likely to be created in DAML.

If you don't like LISP, the OIL syntax at http://www.ontoknowledge.org/oil
works pretty well.  We're looking at using that as part of a compact query
language.

Finally, there are non-DAML XML encodings of knowledge bases such as those
used by FaCT (http://www.cs.man.ac.uk/~horrocks/FaCT) and Cerebra
(http://www.networkinference.com/).  Raw XML syntax isn't 'pleasant' in my
opinion, but is certainly easier to read than DAML+OIL encoded in RDF
encoded in XML.

		- Peter
--
Peter Crowther, VP Development, Network Inference Limited

Received on Friday, 31 August 2001 11:52:54 UTC