Hi All --
I was surprised to learn that RDF reasoning appears to be very
restricted, as in:
"...a more complex fact is expressed in RDF using a
conjunction (logical-AND) of simple binary relationships. RDF does not
provide means to express negation (NOT) or disjunction (OR). The
expressive power of RDF corresponds to the existential-conjunctive (EC)
subset of first order logic"
[http://www.w3.org/TR/rdf-concepts/]
Surely, for RDF to function as useful service-finding glue in the
Semantic Web, it should at least support discovery of resources by being
able to walk arbitrary depth first- and second-order
hierarchies/ontologies ?
Ah, I hear you say, but the inferencing needed can be undecidable, or at
best inefficient.
Not so ! It's been known for a number of years that a form of
datalog augmented with negation-as-failure has efficient, terminating,
sound and complete inference methods.
There's a rather formal paper about this: Backchain
Iteration: Towards a Practical Inference Method that is Simple Enough to
be Proved Terminating, Sound and Complete. Journal of Automated
Reasoning, 11:1-22, 1993. (Yes, these results have been around for
a while !)
On the less formal side, here's a little example showing the kind
of thing one would expect for really useful Semantic Web
inferencing. The notation is that of the Internet Business
Logic system, and the example can be run by pointing Netscape 7 to
www.reengineeringllc.com
and selecting DataModelling1 .
Example:
1) Table heading:
this-relationship
is a specialization of this-higher-relationship
=================================================================
This is an "X IS-A relation Y" before inserting any
tuples.
2) Populate the table
this-relationship is a specialization of this-higher-relationship
=================================================================
sibling
family
marriage
family
family
the class of all relationships
3) Say that IS-A is transitive
some-relationship1 is a specialization of some-relationship2
that-relationship2 is a specialization of
some-relationship3
------------------------------------------------------------
that-relationship1 is a specialization of
that-relationship3
4) Say that if two people are related by R1, and R1
IS-A R2, then they are related by R2
some-person is related through some-relationship to
some-other-person
that-relationship is a specialization of
some-higher-relationship
---------------------------------------------------------------------------
that-person is related through
that-higher-relationship to that-other-person
5) Provide some facts
this-person is related through this-relationship to
this-other-person
=====================================================================
Fred
marriage
Jane
Fred
sibling
John
6) Ask who is related to whom and by what
relationship, and get the answer table
this-person is related through this-relationship to
this-other-person
=====================================================================
Fred
family
Jane
Fred
family
John
Fred
marriage
Jane
Fred
sibling
John
Fred
the class of all relationships Jane
Fred
the class of all relationships John
7) Ask for an explanation of the last line of the
answer, and get this:
Fred is related through sibling to John
sibling is a specialization of the class of all
relationships
----------------------------------------------------------------
Fred is related through the class of all
relationships to John
sibling is a specialization of family
family is a specialization of the class of all
relationships
--------------------------------------------------------------
sibling is a specialization of the class of all
relationships
The alert reader will have noted that the example has sneaked
in some second-order reasoning, while staying inside datalog.
One of the interesting things one can explore using this notation,
is how to write the English phrasing and logic to avoid traps like John
Sowa's famous example "Clyde is and elephant, and
an elephant is a species, therefore Clyde is
a species" . (You can see that the error crept
in through the sloppy phrase in red.)
So, this also begins to address the capture of English meanings in a form
suitable for computation.
Hope this helps. All comments gratefully
received. -- Adrian
PS: If your screen does not show monospace fonts, the
above tables may be misaligned -- apologies.
Adrian Walker
www.reengineeringllc.com
PO Box 1412
Bristol
CT 06011-1412 USA
adrianw@snet.net
Phone: USA 860 583 9677
Cell: USA 860 830 2085
Fax: USA 860 314 1029