[Top] [All Lists]

Re: [ontology-summit] [Applications] Launching the conversation about La

To: ontology-summit@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Mon, 23 Jan 2012 09:06:50 -0500
Message-id: <4F1D697A.6070206@xxxxxxxxxxx>
Michael,    (01)

In 1998, I was very strongly supportive of the goals of the
Semantic Web.  But decisions were made that replaced the box
labeled logic in the original layer cake with one very limited
version of logic.  The net result is that anybody who implements
an application must use a totally unrestricted programming
language for the processing -- with no guidance from logic.    (02)

As for OWL, I have no objection to DLs, to the idea of using DLs
as a restricted logic for defining the T-box (terminology), or
even to an XML-based option for representing logic.    (03)

But OWL is the only reasoning component left in the layer cake, and
it's too weak to be used by itself.  You need a much richer A-box
to do anything useful.  SWRL and RuleML were more expressive logics,
but the "decidability thought police" kicked them out of the layer
cake.  RIF is left as a notation, but the only implementation is
limited to what can be expressed in OWL -- and that's not much.    (04)

Even the WHERE clause in SQL supports full first-order logic for
stating stating database constraints and queries.  I don't recommend
SQL notation for FOL, but you need something with the expressive
power of Datalog for DB constraints and queries.  For specifying
software, UML has a variety of notations that taken together support
full FOL.  But nothing left in the layer cake can support SQL and UML.    (05)

> If this is an argument that what the major web companies are doing
> is more appropriate than OWL, then it is important to note that
> major web companies are using something MUCH weaker than OWL.    (06)

No they aren't.  The overwhelming majority of commercial web sites
are built around SQL databases.  And programmers have been using
UML and similar notations for years.  Those tools provide guidelines
for safe and efficient use of expressive power without the dogma
that the logic has to be restricted.    (07)

Please note that Bob MacGregor, who developed LOOM and Power LOOM,
said that users always asked for *more* expressive power.  They
never asked for decidability.  But Bob was drummed out of the DL
community by the "thought police" because he stated the obvious.    (08)

> OWL is syntax neutral, there are many to choose from.  You could
> easily make another one using conceptual graphs.    (09)

OWL was designed to be used with triples.  That is huge syntactic
constraint that made it incompatible with relational DBs.  If OWL
had been designed to support a type system for SQL, UML, and
other widely used tools, it could have become a very successful
component of current best practices.  As it is, mainstream COTS
ignores it, and most programmers have never heard of it.    (010)

> ... you are taking the opportunity to express yet again your
> time-worn  anti-OWL and anti-SemWeb views.,    (011)

No!  See my preamble above and the following paper:    (012)

    http://www.jfsowa.com/pubs/fflogic.pdf    (013)

This paper, by the way, was published in a journal of which Jim Hendler
was the editor.  Jim liked the paper very much, despite the fact that
I cast quite a bit of doubt on Semantic Web dogma.  I don't object
to OWL, but to the dogma that has eliminated every other logic.    (014)

Note that you cannot define a triangle in OWL.  You can say that
it has three sides and three vertices.  But you can't say that each
vertex is connected to the other two because that would create a cycle.
And OWL only permits models that are limited to trees -- you can't
define a benzene ring, a triangle, or a bridge truss with triangles.    (015)

> whether you feel there is a more strongly supported [logic] with
> COTS tools  and other standards that OWL.    (016)

If you're looking for COTS tools, I suggest that you look at SQL
and UML.  The tools built around them use the full expressive power
of FOL, and the software they specify runs the world economy.    (017)

Suggestion:  If we really want the Ontology Summit to address the
use of ontologies for large systems, we should ask a very serious
question:  Why don't large companies like Google, Microsoft (Bing),
Yahoo!, Amazon, and IBM use the Semantic Web tools?    (018)

That could be a very informative track, which would cut through
the layers of confusion promoted by the SW thought police.    (019)

John    (020)

Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012  
Community Portal: http://ontolog.cim3.net/wiki/     (021)
<Prev in Thread] Current Thread [Next in Thread>