Re: MINUES: Teleconference 30 Jan 2003

From: "Smith, Michael K" <michael.smith@eds.com>
Subject: MINUES: Teleconference 30 Jan 2003
Date: Thu, 30 Jan 2003 13:31:56 -0600

[...]

> ACTION: Peter Patel-Schneider to study annotations bug and come up
> with proposal.

[...]

Here is the situation with annotations:

Summary:

Annotations are only a part of the abstract syntax.  There they have
no semantic impact.  So that OWL Lite and OWL DL can have
annotation-like constructs in the triple syntax, annotations are
translated into triples, which do have semantic impact.

In the presence of annotations, entailments do not line up between the
direct semantics and the RDF-style semantics.  However, there is no
non-monotonicity, contrary to my thinking in yesterday's telecon.


Solutions:

1/ Do nothing.

2/ Change the semantics document a whole lot.  This may take a while.

   Several components to handle annotations would have to be added to
   the direct semantics interpretations, and incorporated into the
   semantic rules.  The correspondence proof would have to be
   overhauled as well.  I think that this can be done, but it is
   significant work.

3/ Remove annotations from the triple syntax for OWL DL and OWL Lite.
   This would be easy. 

peter

Received on Friday, 31 January 2003 08:39:45 UTC