[Top] [All Lists]

Re: [ontolog-forum] Ontology modules and namespaces

To: ian@xxxxxxxxxxxxxxxx
Cc: "'[ontolog-forum]'" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Wed, 28 Oct 2009 10:24:46 -0500
Message-id: <4AE8623E.7070303@xxxxxxxxxxx>
Dear Ian, Matthew, Randall, and Chris,    (01)

I don't disagree with Ian about the choice of UML for
IDEAS, but I also agree with the other responses to Ian.    (02)

IB> I think CL is very difficult to use, and it doesn't do
 > anything I can't already do with a well thought-out diagram
 > notation. I'm sure you disagree...    (03)

MW> CL is an abstract syntax. The concrete syntaxes actually
 > include a graphical notation - conceptual graphs. If you chose,
 > you could probably map the ideas notation to the CL abstract
 > syntax, or a subset of it.    (04)

I violently agree with both of you.  I strongly endorse the use
of diagrams such as UML and others, and I certainly agree that
most versions of logic are horribly difficult to read and use
(and that includes all the SemWeb notations).    (05)

As Matthew said, the CGIF dialect has a one-to-one mapping to
conceptual graphs.  I suspect that the CGIF dialect would be
easier to map to UML diagrams than CLIF.  But either mapping
would define exactly the same semantics.    (06)

IB> I think some of the stuff I can do simply in my diagrams
 > would be a nightmare in CL.    (07)

I agree, and it would also be a nightmare in RDF and OWL.
In fact, one of my major complaints about the Semantic Web
is that they did *not* adopt and extend UML as the primary
interface for the people who had to read and write that stuff.    (08)

IB> [IDEAS] founding categories are individual (something with
 > spatio-temporal extent), type and tuple. I can see how CL handles
 > type and tuple (albeit through some arcane notion, probably
 > starting with "for all..."), but it doesn't seem to have a
 > fundamental category of something with spatio-temporal extent.    (09)

MW> It does not have to. It just has not made that commitment, and
 > should not either, CL is a language not an ontology. It does not
 > prevent you from declaring that there are individuals that are
 > spatio-temporal extents.    (010)

Matthew answered for me.  I just want to emphasize that CL has
*no preferred notation* of any kind.  The semantics is defined
by an abstract syntax, and all notations, graphic or linear,
that support that semantics are on an equal footing.    (011)

If somebody in the IDEAS group defines a mapping to the CL
semantics, the IDEAS diagrams would automatically become
CL dialects.    (012)

IB> Maybe we just built the right wheel for what we're doing,
 > instead of re-inventing one.    (013)

I think you have the right wheel for your purpose.  My suggestion
is that you attach that wheel to the CL carriage.  The advantage of
doing so is that you get a precise definition and interoperability
with any and every notation that is definable in Common Logic --
and that includes almost every declarative notation in use in
computer science and related fields.  (There is a slight extension
to CL, called IKL, which covers most of the remainder.)    (014)

RRS> Diagrams support only human comprehension. Logical formalisms
 > can be diagrammed, but diagrams don't support inference or any
 > other form of computation.    (015)

Formally defined diagrams *are* a logical formalism.  They support
automated inference in exactly the same way as linear notations --
they get translated to some internal form with bits and pointers.    (016)

MW>> Well if it's a graphical notation, chances are it's just boxes
 >> and lines again.    (017)

CM> That's all I've found, but perhaps there is a rigorous semantics
 > for it lurking somewhere.    (018)

Their notation is UML, which has a very long history behind it.
The "three amigos" who merged their notations to form UML, did not
have a formal definition.  However, there has been a long drawn-out
battle between the pro- and anti-logic crowds.    (019)

The result of all that in-fighting is that there are many, many
publications by computer scientists who have defined various UML
diagrams in various notations for logic.  But those publications
are incompatible with one another, and they are ignored by the
people who actually use UML.  If you go to Google Scholar and type
"UML logical foundations", you get 20,000 different foundations.    (020)

As I said in a previous note to Ian, I believe there is a very
strong *requirement* for defining UML diagrams in terms of CL.    (021)

John    (022)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (023)

<Prev in Thread] Current Thread [Next in Thread>